[Feature] Multi-GPU support - #55

Open

yoshikisd wants to merge 19 commits into

Commits on Oct 19, 2025

Created the distributed module
yoshikisd
committed
0268f5b

Commits on Oct 20, 2025

Added additional functions to distributed
yoshikisd
committed
fded573

Commits on Nov 8, 2025

Commits on Nov 9, 2025

Added example scripts for running multi-GPU jobs with spawn or torchrun
yoshikisd
committed
6f15c36
Added example script for running a multi-GPU speed test
yoshikisd
committed
842bb49
Modified runtime error message in multigpu.setup
yoshikisd
committed
1a3d0cd
Fixed linting issues on multigpu
yoshikisd
committed
cad3d9c
multigpu.setup now recognizes MASTER_ADDR and MASTER_PORT when setting up spawn-based jobs
yoshikisd
committed
cd584fc
Simplified the logic in MASTER_ADDR and MASTER_PORT definition in multigpu.setup
yoshikisd
committed
a8d2bfa
Added multi-GPU pytests
yoshikisd
committed
32e9256
Merge branch 'master' into multigpu
yoshikisd
committed
e5996a4

Commits on Dec 12, 2025

Commits on Dec 18, 2025

Make the timeout default 30 seconds and enforce timeout in units of seconds
yoshikisd
committed
e059c18

Commits on Feb 22, 2026

Merge remote-tracking branch 'origin/master' into multigpu
yoshikisd
committed
9795f80