Skip to content

Pull requests: ROCm/triton

New pull request New
44 Open 536 Closed
44 Open 536 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

add blocked version to address performance issue of when N is large
#672 opened Dec 4, 2024 by xiaohuguo2023 Loading…
6
persistent kernel version of the flash attention forward + FLOP calculation fix when seqlen_q != seqlen_k enhancement New feature or request
#670 opened Nov 29, 2024 by juuso-oskari Loading…
5 tasks done
@juuso-oskari
1
Tianxing/fa int8
#669 opened Nov 29, 2024 by Chi-Chu319 Loading…
5 of 7 tasks
@Chi-Chu319
Add gemm tuning configs for weekly tuning CI
#662 opened Nov 13, 2024 by AlexAUT Loading…
1
Perf Kernels benchmark workflow
#651 opened Oct 29, 2024 by NISHIY-EKSDEE • Draft
Use mask during load for Softmax
#645 opened Sep 24, 2024 by rahulbatra85 Loading…
Add performance reference for important matmul kernels
#642 opened Sep 16, 2024 by zhanglx13 Loading…
5
RMSNorm Blocked Implementation
#638 opened Sep 12, 2024 by rahulbatra85 Loading…
11
Added a sketch of the instr. sched. group barriers
#622 opened Aug 2, 2024 by ravil-mobile • Draft
Add INT4 quant/de-quant kernels
#620 opened Jul 29, 2024 by rahulbatra85 Loading…
2
Peel off the last iteration of matmul loop
#617 opened Jul 27, 2024 by zhanglx13 • Draft
Add bench_gemm to benchmark matmul kernel with Triton and hipblaslt
#616 opened Jul 24, 2024 by zhanglx13 • Draft
[CODE SHARING] Ravil/sched inst
#611 opened Jul 10, 2024 by ravil-mobile • Draft
[CODE SHARING] Insertions of custom LLVM IR and AMDGCN codes to triton
#610 opened Jul 10, 2024 by ravil-mobile • Draft
Add more unit tests to FA fwd kernels.
#609 opened Jun 28, 2024 by xinyazhang Loading…
2
Add a script for tuning flash attention kernels
#605 opened Jun 25, 2024 by yiqian1 Loading…
Fixed streamk kernel bug
#602 opened Jun 20, 2024 by ravil-mobile Loading…
4
Groenenboomj/fixes causal
#575 opened May 8, 2024 by groenenboomj Loading…
2
[release/internal/2.2.x] Only include HIP headers for triton
#561 opened Apr 17, 2024 by jithunnair-amd • Draft
support bypassing data layout conversion for atomic operator
#556 opened Apr 9, 2024 by xiaohuguo2023 Loading…
[MFMA] Implement MFMA 4x64 v3
#550 opened Apr 1, 2024 by binarman • Draft
Aot change merge
#549 opened Mar 28, 2024 by groenenboomj • Draft
8
Add llvm flag
#547 opened Mar 28, 2024 by zhanglx13 Loading…
3
fix: replace if/else statement with tl.where
#545 opened Mar 22, 2024 by Sara-KS Loading…
Solve official flash-attention.py fails: typeError:'function' object is not subscriptable
#541 opened Mar 20, 2024 by zhangxiao-stack Loading…
ProTip! Mix and match filters to narrow down what you’re looking for.