issues
search
ROCm
/
triton
Development repository for the Triton language and compiler
MIT License
80
stars
23
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
[GEMM][Tutorial] Refine test_correctness
#463
zhanglx13
closed
6 months ago
0
support configure multiple waves in flash-attention
#462
scxiao
closed
5 months ago
0
Flash Attention Triton for Mi50s
#461
ThePerfectComputer
closed
5 months ago
3
Merage aot features
#460
groenenboomj
closed
5 months ago
2
Add autotuning for FA
#459
vgokhale
closed
6 months ago
0
Dockerfile and test
#458
micmelesse
closed
6 months ago
0
[Tuning] Gemm tuning v3
#457
zhanglx13
closed
6 months ago
0
[TUTORIAL] Enable all types in gemm tutorial
#456
zhanglx13
closed
6 months ago
3
New CI
#455
micmelesse
closed
6 months ago
0
[Triton] [PyTorch UT] `tl.reshape` cherry-pick support
#454
jataylo
closed
5 months ago
2
enable layout conversion from mfma to dot_op for mfma16.
#453
scxiao
closed
6 months ago
0
[BACKEND] Add support for reshape op (#2676)
#452
zhanglx13
closed
6 months ago
0
[HotFix] Fix dot op for RDNA3 architecture
#451
joviliast
closed
6 months ago
0
[tool] Added a script to print occupancy info
#450
zhanglx13
closed
6 months ago
0
[GEMM] [Tuning] Skip BLOCK_SIZE that is too large compare to M/N
#449
zhanglx13
closed
6 months ago
0
WMMA instructions are not supported for GEMM
#448
joviliast
closed
1 month ago
7
Bug for the scripts/tune_gemm.py
#447
ybai62868
closed
6 months ago
29
[BUG] Discrepancy Between Triton JIT Computed Sum and Torch Sum
#446
Ldpe2G
closed
3 months ago
4
Use full-vectorized load instructions for load vectorization
#445
htyu
closed
6 months ago
20
Merge changes from upstream FA bwd kernel
#444
vgokhale
closed
6 months ago
0
[PYTORCH UT] Assertion `!NodePtr->isKnownSentinel()' failed.
#443
jataylo
closed
1 month ago
4
Add support for MFMA layout to view_slice instruction
#442
oplavsic
closed
6 months ago
0
[Backend] Refactor mfma selection
#441
zhanglx13
closed
6 months ago
1
Dot slicing pass
#440
oplavsic
closed
6 months ago
5
[Backend] Refactor sharedToDotOperandMFMA lowering
#439
zhanglx13
closed
6 months ago
1
optimize splitK FA for attention decode
#438
scxiao
closed
5 months ago
1
FA splitK for decode
#437
scxiao
closed
6 months ago
0
Support matmul semanthics for WMMA dot operation
#436
joviliast
closed
4 months ago
3
Extend encoding attributes for WMMA layout
#435
joviliast
closed
4 months ago
1
Minor edits to HBM bandwidth measurement kernel
#434
vgokhale
closed
7 months ago
0
Ifu 12 15 2023
#433
micmelesse
closed
6 months ago
1
[MFMA] Support 64x4 and 4x64 tile size
#432
binarman
closed
6 months ago
2
Add kernel to check HBM BW
#431
vgokhale
closed
7 months ago
0
Add a kernel to measure HBM bandwidth vs WGs
#430
vgokhale
closed
7 months ago
0
[MFMA] Remove CTA related code from layout
#429
binarman
closed
6 months ago
1
Issue with test_reduce_layouts
#428
micmelesse
closed
3 months ago
0
Add view_slice ttgir instruction
#427
oplavsic
closed
6 months ago
0
[ROCM] drop GIL for launch, and set value=false upon pointer error
#426
jayfurmanek
closed
7 months ago
0
Generated Backend
#425
micmelesse
closed
6 months ago
2
[MFMA] Reenable removed CDNA3 int and fp8 support
#424
binarman
closed
7 months ago
1
[WMMA][Dot] Support WMMA layout in TritonAMDGPUAccelerateMatmulPass
#423
joviliast
closed
7 months ago
1
RMS Norm achieving poor memory bandwidth on MI300
#422
anupambhatnagar
closed
1 month ago
5
Lds vec size
#421
zhanglx13
closed
6 months ago
0
[GEMM] [Tuning] Make tuning script more verbose
#420
binarman
closed
7 months ago
1
[GEMM] Add script to run one tuning config
#419
binarman
closed
7 months ago
0
Replace inline assembly in commonShflSync with intrinsics
#418
binarman
closed
7 months ago
0
Add support for ALiBi-style attention bias
#417
vgokhale
closed
7 months ago
2
Split k mm fix
#416
xiaohuguo2023
opened
7 months ago
4
Optimized stream-k kernel for AMD GPUs
#415
zhanglx13
closed
4 months ago
4
support type conversion between fp8 formats and bf16/fp32 with HW instructions on MI300
#414
scxiao
closed
6 months ago
1
Previous
Next