Kernel Judge

Evaluation 1

Status
passed
Score
0.72x
Submission
submission 1
Challenge
lm_head
Run
starter
Judge version
1
Node
mlsys-b200.ucsd.edu
GPUs
GPU-f1451c55-b65f-6b4e-8d94-87473005fb46, GPU-dc6665d0-263b-9b9c-1197-68357d15ba3a, GPU-1561f1bf-2371-a785-bf45-41ecca0d4fbf, GPU-dfaf1799-2999-ec9e-0786-39ea8fda3590
Driver
595.71.05
Started
2026-11-06 10:00 UTC
Finished
2026-11-06 10:01 UTC

Shapes

ShapeBaselineSubmissionSpeedupConfidence intervalFloorChecks
batch 1 122.2 us 207.8 us 0.59x 0.59 to 0.60 41.60 us passed
batch 8 197.6 us 279.6 us 0.71x 0.70 to 0.71 43.50 us passed
batch 32 353.1 us 440.2 us 0.80x 0.80 to 0.80 53.00 us passed
batch 64 531.3 us 668.6 us 0.79x 0.79 to 0.79 105.9 us passed

Tests

TestOutcomeMessage
batch 1, shuffle true passed hidden: 0.000e+00 (worst row 0.000e+00, bound 0.02); logits rank 0: 1.542e-04 (worst row 2.290e-04, bound 0.02); logits rank 1: 1.542e-04 (worst row 2.290e-04, bound 0.02); logits rank 2: 1.542e-04 (worst row 2.290e-04, bound 0.02); logits rank 3: 1.542e-04 (worst row 2.290e-04, bound 0.02)
batch 8, shuffle true passed hidden: 0.000e+00 (worst row 0.000e+00, bound 0.02); logits rank 0: 1.375e-04 (worst row 3.434e-04, bound 0.02); logits rank 1: 1.375e-04 (worst row 3.434e-04, bound 0.02); logits rank 2: 1.375e-04 (worst row 3.434e-04, bound 0.02); logits rank 3: 1.375e-04 (worst row 3.434e-04, bound 0.02)
batch 32, shuffle true passed hidden: 0.000e+00 (worst row 0.000e+00, bound 0.02); logits rank 0: 4.545e-04 (worst row 1.684e-03, bound 0.02); logits rank 1: 4.545e-04 (worst row 1.684e-03, bound 0.02); logits rank 2: 4.545e-04 (worst row 1.684e-03, bound 0.02); logits rank 3: 4.545e-04 (worst row 1.684e-03, bound 0.02)
batch 1 passed
batch 8 passed
batch 32 passed
batch 64 passed

Files

Attempts

AttemptStateNodeClaimedFinishedError
1 expired mlsys-b200.ucsd.edu 2026-11-06 09:55 UTC 2026-11-06 10:00 UTC lease expired
2 done mlsys-b200.ucsd.edu 2026-11-06 10:00 UTC 2026-11-06 10:01 UTC