pytorch/xla

Pull Requests Commits

test fix applied for autocast

ManfeiBai committed 2 years ago

Verified f6933775

Update pjrt_computation_client.cc

ManfeiBai committed 2 years ago

Verified ad2a7a7b

Update computation_client.h

ManfeiBai committed 2 years ago

Verified 06c871d3

Update setup.py

ManfeiBai committed 2 years ago

Verified a9c1d76c

Delete openxla_patches/gpu_topk_rewriter.diff

ManfeiBai committed 2 years ago

Verified d66d3e28

Update constexpr_return.diff

ManfeiBai committed 2 years ago

Verified eeb3663e

Update gpu_race_condition.diff

ManfeiBai committed 2 years ago

Verified 0ed40933

test openxla pin update nov24

ManfeiBai committed 2 years ago

Verified e37ee80c

Distribute Literal->Tensor copies across thread pool (#5825)

jonb377 committed 2 years ago

Verified ec54fd44

Add stack frame id metadata and correct filename and line number for custom op_name (#5838)

mrnikwaws committed 2 years ago

Verified 5fa89223

Update troubleshooting doc with few common env var combination (#5929)

JackCaoG committed 2 years ago

Verified 6c5b7b8e

Add graph hash and num input/output to PT_XLA_DEBUG (#5947)

JackCaoG committed 2 years ago

Verified f3b75ba1

Refactor ExecuteReplicated to operate on sharded data directly (#5737)

will-cromar committed 2 years ago

Verified 7b92a949

Error when changing `PJRT_DEVICE` after runtime initialized (#5948)

will-cromar committed 2 years ago

Verified 346ef7c1

Add benchmark noise reducing info. (#5924)

golechwierowicz committed 2 years ago

Verified c8ce0620

Add link to work log issue in FIX_LOWERING_FOR_CORE_ATEN_OPS.md (#5944)

wonjoo-wj committed 2 years ago

Verified 09f2a14c

Parallelize d2h transfers (#5824)

jonb377 committed 2 years ago

Verified b9475d9e

Update pjrt.md (#5941)

Liyang90 committed 2 years ago

Verified eb728a89

Truncate python stack when outputting frame that cause the graph executation (#5933)

JackCaoG committed 2 years ago

Verified a5e07389

Don't fallback for pow (#5919)

qihqi committed 2 years ago

Verified 402166be

Lower quant/dequant torch op to StableHLO (#5763)

lsy323 committed 2 years ago

Verified a3b0c6ec

Fix `as_strided` for inputs smaller than the arguments specification. (#5914)

ysiraichi committed 2 years ago

Verified 4ac255ae

Collect CUDA/CPU profiling info into result sheets. (#5921)

golechwierowicz committed 2 years ago

Verified c03afb14

Fix get_gpu_name returning empty name. (#5925)

golechwierowicz committed 2 years ago

Verified 16a40cc6

Revert "Configure PATH and LD_LIBRARY_PATH to include cuda (#5761)" (#5928)

vanbasten23 committed 2 years ago

Verified cfb36175

Copy input tensors before async transfer (#5830)

will-cromar committed 2 years ago

Verified b69fb93d

Refactor Dynamo (custom op) integration code (#5805)

yeounoh committed 2 years ago

Verified 3385bd64

update result analyzer script (#5920)

zpcore committed 2 years ago

Verified 037bbd66

Fix tier import failure for relative import (#5918)

frgossen committed 2 years ago

Verified fe7b0075

Merge benchmark scripts into the main branch (#5913)

frgossen committed 2 years ago

Verified 94ca4597

Older