• Joined on 2025-12-08
hexu pushed to wu-blackwell-4lane-tc at wu-arch/kernels 2026-07-12 23:33:43 +08:00
f1aa1303d2 feat: add flash pipeline kernel cases
d6fbd447c3 Add Wu TMEM FlashAttention validation cases
Compare 2 commits »
hexu pushed to feature/scalar-tmem-softmax-vector at wu-arch/kernels 2026-07-12 23:33:43 +08:00
ec437cce6e fix: update fp16 softmax case25 score expectation
hexu pushed to asplos-ae at wu-arch/chipyard 2026-07-12 23:33:32 +08:00
hexu created branch asplos-ae in wu-arch/chipyard 2026-07-12 23:33:32 +08:00
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/chipyard 2026-07-12 23:33:32 +08:00
fc32e7f59c chore: update radiance pipelined softmax
hexu pushed to wu-blackwell-fp8-e4m3 at wu-arch/chipyard 2026-07-12 23:33:32 +08:00
8fe292015c chore: update radiance pipelined softmax
hexu pushed to wu-blackwell-4lane-tc at wu-arch/chipyard 2026-07-12 23:33:32 +08:00
b0a79c26ee Update radiance pre-TMEM-bank baseline
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/radiance 2026-07-12 23:32:59 +08:00
cdbf07ab9d feat: pipeline scalar softmax and blackwell mma issue
hexu pushed to wu-blackwell-fp8-e4m3 at wu-arch/radiance 2026-07-12 23:32:59 +08:00
4ab18ab6b1 feat: pipeline fp8 scalar softmax and blackwell mma issue
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/vortex 2026-07-12 23:32:32 +08:00
eb1b2ea330 feat: add pipelined scalar tmem softmax unit
hexu pushed to wu-blackwell-fp8-e4m3 at wu-arch/vortex 2026-07-12 23:32:32 +08:00
331c4decaa feat: add pipelined fp8 scalar tmem softmax unit
hexu pushed to wu-blackwell-4lane-tc at wu-arch/vortex 2026-07-12 23:32:32 +08:00
9251ba0a24 feat: add scalar fexp support
97a1eff701 Add scalar TMEM load-store path
Compare 2 commits »
hexu pushed to wu-blackwell-fp8-e4m3 at wu-arch/radiance 2026-07-10 21:20:52 +08:00
524e4c016d feat: split blackwell tmem a c arrays
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/chipyard 2026-07-10 21:20:39 +08:00
71c84f63a5 chore: update radiance fp16 scalar tmem softmax
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/radiance 2026-07-10 21:20:20 +08:00
3e8976490d feat: add fp16 scalar tmem softmax and split tmem
hexu pushed to wu-tmem-bank-scalar-softmax at wu-arch/vortex 2026-07-10 21:20:07 +08:00
2bfc6c4bde fix: pack scalar tmem softmax as fp16
hexu created branch wu-tmem-bank-scalar-softmax in wu-arch/vortex 2026-07-10 21:20:07 +08:00
hexu pushed to feature/scalar-tmem-softmax-vector at wu-arch/chipyard 2026-07-10 16:29:42 +08:00
234bfbc3a4 chore: update radiance scalar tmem softmax
hexu created branch feature/scalar-tmem-softmax-vector in wu-arch/chipyard 2026-07-10 16:29:41 +08:00
hexu pushed to feature/scalar-tmem-softmax-vector at wu-arch/kernels 2026-07-10 16:29:30 +08:00
ca42426e5c feat: add scalar tmem softmax vector cases