Skip to content
GitLab
Explore
Sign in
和我作队
qwen-vllm-xiandao-2026
Repository
Branches
Overview
Active
Stale
All
gemm-silu-threshold-256-0709
a41c24f7
·
test gemm silu threshold 256
·
Jul 09, 2026
gemm-silu-threshold-128-0709
b7c700e3
·
test gemm silu threshold 128
·
Jul 09, 2026
gemm-silu-bf16-mid-0710
c3cadd7b
·
test gemm silu bf16 intermediate
·
Jul 10, 2026
gemm-silu-bf16-mid-static-0710
0a187a7c
·
test gemm silu static bf16 intermediate
·
Jul 10, 2026
codex/chunk-h-autotune-warps8-bv32-0710
52abb172
·
expand GDN chunk autotune for DTK
·
Jul 10, 2026
best-68.0961-0712
5793a94d
·
test BF16 up branch in fused GEMM SiLU
·
Jul 12, 2026
archive-main-before-clean-0712
ff3bb4ec
·
Promote verified 68.0961 baseline to main
·
Jul 12, 2026
submission-68.0961-clean-0712
67cd3408
·
document submission compliance review
·
Jul 12, 2026
best-71.6201-0713
2f8800eb
·
preserve ROCm small-batch attention threshold
·
Jul 13, 2026
best-compliant-71.5705-0713
ff2b0d7b
·
limit ROCm prefix optimization to pure prefill
·
Jul 13, 2026
experiment-gpu095-af948b7-0713
af948b73
·
increase gpu_memory_utilization to 0.95
·
Jul 13, 2026
best-compliant-76.2214-0714
aab9e932
·
optimize ROCm page784 prefix query tile
·
Jul 13, 2026
rocm-page784-bm64-0713
aab9e932
·
optimize ROCm page784 prefix query tile
·
Jul 13, 2026
best-compliant-77.8698-0714
d42b8143
·
reduce ROCm page784 prefix cache unroll
·
Jul 14, 2026
rocm-page784-bm64-unroll2-0714
d42b8143
·
reduce ROCm page784 prefix cache unroll
·
Jul 14, 2026
best-compliant-80.0198-0714
ab6f410c
·
test minimal ROCm page784 prefix unroll
·
Jul 14, 2026
rocm-page784-bm64-unroll1-0714
ab6f410c
·
test minimal ROCm page784 prefix unroll
·
Jul 14, 2026
experiment-long-warmup-02393bd-0714
02393bd5
·
add long context warmup for attention kernel
·
Jul 14, 2026
rocm-page784-address-fastpath-0714
c46b8643
·
record BW1000 address fastpath results
·
Jul 14, 2026
best-compliant-80.2281-0714
1f70f910
·
record page784 launch matrix results
·
Jul 14, 2026
Prev
1
2
Next