Morph MORPH
SIGN IN SIGN UP
dhruvhead / xla UNCLAIMED
0 0 1 C++

COMMITS

/ benchmarks/benchmark_experiment.py
test
October 2, 2024
Y
[benchmarks] Default to functionalization disabled. (#8093)
Yukio Siraichi committed
b6bcf03
September 13, 2024
Y
[benchmarks] Disallow `XRT` for XLA:CUDA. (#8006)
Yukio Siraichi committed
afbd7bd
August 7, 2024
P
remove openxla_eval in test (#7811)
Pei Zhang committed
c7d9755
July 26, 2024
Y
[benchmarks] Turn CUDAGraphs on for inductor. (#7749)
Yukio Siraichi committed
0b99085
July 24, 2024
Y
Re-implement and refactor the verifier. (#7724)
Yukio Siraichi committed
c69742c
June 20, 2024
P
Type annotation for benchmarks/ (#7289)
Pei Zhang committed
98dd99e
June 10, 2024
P
torchbench CI (#7162)
Pei Zhang committed
291764f
June 6, 2024
I
Add options in benchmark script to keep model on CUDA (#7193)
iefgnoix committed
f95454a
June 3, 2024
P
isolate the PJRT dependency between torch_xla and torch_xla2 (#7144)
Pei Zhang committed
0f9bff3
May 16, 2024
P
Support torch_xla2 benchmarking using torchbench (#7013)
Pei Zhang committed
9e18935
February 23, 2024
M
Allow Inductor benchmarks for CPU backends. (#6555)
Mason Chang committed
263da4e
January 31, 2024
Y
[benchmarks] Fix CUDA device check. (#6419)
Yukio Siraichi committed
be44bc6
January 9, 2024
P
Revert "remove model training using openxla backend (#6262)" (#6276)
Pei Zhang committed
551c0df
January 8, 2024
P
remove model training using openxla backend (#6262)
Pei Zhang committed
a89ebef
December 20, 2023
E
benchmark/experiment_runner: save accelerator_model on error (#6218)
Emilio Cota committed
86636ba
December 19, 2023
F
Add option to dump HLOs during benchmark runs (#6200)
frgossen committed
61d42eb
December 15, 2023
F
Fix benchmarking for non-dynamo path (#6179)
frgossen committed
68e48db
F
Forward subprocess stdout and stderr in all cases (#6154)
frgossen committed
5c00aad
December 14, 2023
F
Reapply with fix "Add comments to _is_available and raise exceptions … (#6143)
frgossen committed
f9dc824
F
Add comments in benchmark_experiment and generalize workaround for too long strings in `filename_str` (#6117)
frgossen committed
29b183b
December 13, 2023
F
Revert "Add comments to _is_available and raise exceptions in unimple… (#6141)
frgossen committed
ddc2e34
F
Add comments to _is_available and raise exceptions in unimplemented cases (#6118)
frgossen committed
491cbed
F
Re-land experiment runner tests (#6113)
frgossen committed
3bbafc2
F
Enforce successful imports and cause failure if not set up correctly (#6119)
frgossen committed
ad02e79
F
Remove unused experiment name, which was always `run_all` (#6135)
frgossen committed
a29ff4e
December 9, 2023
Y
Reduce benchmark experiment filename string. (#6069)
Yukio Siraichi committed
0857f2a
December 7, 2023
G
Fix CUDA profiling dump. (#6030)
Grzegorz Olechwierowicz committed
65176ff
December 2, 2023
Y
Allow benchmark runner config: `openxla` for inference runs. (#5939)
Yukio Siraichi committed
dcdd66e
November 22, 2023
F
Merge benchmark scripts into the main branch (#5913)
frgossen committed
94ca459