TensorRT-LLMs/cpp/tensorrt_llm/kernels/decoderMaskedMultiheadAttention/cubin
Perkz Zheng 4d711be8f4
Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564)
* move cubins to LFS

Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>

* update cubins

Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>

* add sliding-window-attention generation-phase kernels on Blackwell

Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>

* address comments

Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>

---------

Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>
2025-05-26 09:06:33 +08:00
..
xqa_kernel_cubin.h Update TensorRT-LLM (#2755) 2025-02-11 03:01:00 +00:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_bf16_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_64_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_64_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_128_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_128_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_bf16_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_bf16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_64_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_0_m_32_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_128_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_e4m3_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_fp16_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_16_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_32_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_64_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_8_m_8_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_1_kvt_int8_pagedKV_128_nqpkv_16_m_16_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_e4m3_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_64_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_64_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_128_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_128_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_fp16_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_16_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_32_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_64_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_80.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_86.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_89.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_1_sm_90.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00
xqa_kernel_dt_fp16_d_256_beam_4_kvt_int8_pagedKV_128_nqpkv_1_m_4_sm_120.cubin.cpp Feat: add sliding-window-attention generation-phase kernels on Blackwell (#4564) 2025-05-26 09:06:33 +08:00