pytorch/xla

Pull Requests Commits

Use libtpu from 20240118

yeounoh committed 1 year ago

1cee69b1

* Use xla::KeyValueStoreInterface in GpuClientOptions.

yeounoh committed 1 year ago

843f7ddb

Update openxla_patches/gpu_race_condition.diff

yeounoh committed 1 year ago

1e7c235b

Remove openxla_patches/gpu_compile_options.diff

yeounoh committed 1 year ago

04f7dec5

Remove openxla_patches/cpu_compile_options.diff

yeounoh committed 1 year ago

5f4204f7

Remove openxla_patches/cuda_driver.diff

yeounoh committed 1 year ago

00fa5917

Remove openxla_patches/constexpr_return.diff

yeounoh committed 1 year ago

140390d0

Move OpenXLA pin to 9708d9059a62a908474302c00f31dcac0e0fcb3b

yeounoh committed 1 year ago

1cd6d032

Use spawn as the fork method for the profiler test. (#6302)

ysiraichi committed 1 year ago

Verified 52ef8efc

Add doc images for auto-sharding (#6323)

yeounoh committed 1 year ago

Verified e632897b

Re-land: Fix model initialization. (#6296)

ysiraichi committed 1 year ago

Verified a8b27eb1

[HOT FIX] Disable GPU CI Coverage in master (#6317)

yeounoh committed 1 year ago

Verified ec2db46c

Handle dynamo counter change after fake tensor cache (#6308)

JackCaoG committed 1 year ago

Verified d6dc1a01

Address GPU CI coverage timeout (#6306)

yeounoh committed 1 year ago

Verified 7fae2ab1

update release build to use rc8 tag (#6310)

zpcore committed 1 year ago

Verified 1cfd0cf5

Support mixed precision for torch.where (#6303)

qihqi committed 1 year ago

Verified d49fc70d

Fix global_runtime_device_count() for spmd case (#6266)

vanbasten23 committed 1 year ago

Verified 27ae6dc9

fix gpu spmd core dump (#6275)

mars1248 committed 1 year ago

Verified de27258c

Fix some more aten core ops (#6301)

wonjoo-wj committed 1 year ago

Verified 5d0de94b

Correct order of constant vs args (#6300)

qihqi committed 1 year ago

Verified 55177798

Enable Tensor index check in both dimension (#6299)

ManfeiBai committed 1 year ago

Verified 28502884

Track mean relative error across models. (#6250)

golechwierowicz committed 1 year ago

Verified 30c6b01c

Support of implicit broadcasting with unbounded dynamism (#6219)

sdasgup3 committed 1 year ago

Verified 896de17e

[Core Aten] Add and enable tests for aten index_select and logical_and (#6293)

chunnienc committed 1 year ago

Verified 04e1238a

Ignore non-XLA nodes and their direct dependents. (#6170)

ysiraichi committed 1 year ago

Verified 81410785

Set XLA_USE_SPMD for spmd cpp tests. (#6273)

vanbasten23 committed 1 year ago

Verified a728afe8

Fix output type of custom calls while lowering quant/dequant torch op to HLO (#6283)

sdasgup3 committed 1 year ago

Verified 68f47508

Add single graph option as a runnable script. (#6277)

golechwierowicz committed 1 year ago

Verified ddf4be26

Fix some more core aten opset tests (#6284)

wonjoo-wj committed 1 year ago

Verified 9f1afbd3

Update configuration.yaml to include XLA_DISABLE_FUNCTIONALIZATION (#6290)

wonjoo-wj committed 1 year ago

Verified 4b1eaeef

Older