Skip to content

Update moe github test cases. #3077

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Victarry merged 4 commits into from
Feb 13, 2026
Merged

Update moe github test cases. #3077

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
15c3ce3
Update moe github test cases.
Feb 6, 2026
737d090
Merge branch 'main' into denliu/update_moe_github_tests_main
Phlip79 Feb 12, 2026
1288f01
Fix.
Feb 13, 2026
9945aa9
Remove iteration time check.
Feb 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions megatron/training/training.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2245,7 +2245,8 @@ def save_checkpoint_and_time(

# Stop timer to get accurate train interval time and exclude checkpointing duration
timers('interval-time').stop()
energy_monitor.pause()
if args.log_energy:
energy_monitor.pause()

# Extra barrier is added to make sure all ranks report the max time.
timer_key = 'save-checkpoint-non-persistent' if non_persistent_ckpt else 'save-checkpoint'
Expand Down Expand Up @@ -2298,7 +2299,8 @@ def save_checkpoint_and_time(
)

# Recover timing
energy_monitor.resume()
if args.log_energy:
energy_monitor.resume()
timers('interval-time', log_level=0).start(barrier=True)


Expand Down
1 change: 0 additions & 1 deletion ...st_cases/moe/gpt3_moe_mcore_te_tp2_pp2_ep4_etp1_fine_grained_offloading/model_config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,7 +133,6 @@ MODEL_ARGS:
--overlap-moe-expert-parallel-comm: true
TEST_TYPE: regular # Usually ckpt-resume, but as a WAR to #513 set to regular
METRICS:
- "iteration-time"
- "lm loss"
- "mem-allocated-bytes"
- "mem-max-allocated-bytes"
Expand Down
9 changes: 2 additions & 7 deletions tests/test_utils/recipes/h100/moe.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -176,7 +176,7 @@ products:
- test_case: [gpt3_moe_mcore_te_tp2_pp2_ep4_etp1_no_mtp_no_a2a_ovlp_fine_grained_offloading]
products:
- environment: [dev]
scope: [mr]
scope: [mr, mr-github]
platforms: [dgx_h100]
- test_case: [gpt3_mcore_te_tp2_pp1_te_8experts2parallel_ddp_average_in_collective]
products:
Expand Down Expand Up @@ -205,15 +205,10 @@ products:
- environment: [dev]
scope: [mr, mr-github, mr-slim]
platforms: [dgx_h100]
- test_case: [gpt3_moe_mcore_te_tp2_pp2_ep4_etp1_no_mtp_no_a2a_ovlp_fine_grained_offloading]
products:
- environment: [dev]
scope: [mr-broken]
platforms: [dgx_h100]
- test_case: [gpt3_moe_mcore_te_tp2_pp2_ep4_etp1_fine_grained_offloading]
products:
- environment: [dev]
scope: [mr-broken]
scope: [mr, mr-github]
platforms: [dgx_h100]
#######################################################################
# Super important mr, mr-github tests that run for both DEV and LTS per mr, mr-github #
Expand Down
Loading