Best speedup
4.62x
accepted edits move this up
Live GPU optimization
Protean turns kernel edits into measurable speedup.
Track every candidate, correctness verdict, latency shift, reward, GPU utilization sample, and HUD trace from Spark or Modal while the optimizer keeps improving kernels.
Best latency
0.0310 ms
lower is better
Accepted / trials
17/24
Rejected rows stay in the audit log
Latest reward
4.620
Rewards greater than 1 are preserved
Speedup up, latency down
Run detailPreparing chart...
Trial reward curve
accepted edits are chart annotationsPreparing chart...
Latest candidates
All runs| Trial | Op | Candidate | Verdict | Reward | Speedup | Latency | Cost |
|---|---|---|---|---|---|---|---|
| #24 | sum_reduction | two_stage_reduce | accepted | 4.620 | 4.62x | 0.0310 ms | $0.0043 |
| #23 | prefix_scan | warp_scan_v2 | accepted | 4.310 | 4.31x | 0.0360 ms | $0.0041 |
| #22 | prefix_scan | bad_scan_order incorrect prefix order | incorrect | 0.000 | n/a | n/a | $0.0040 |
| #21 | embedding_lookup | coalesced_cache_v1 | accepted | 4.080 | 4.08x | 0.0390 ms | $0.0038 |
| #20 | quantize_dequant | vectorized_pack_v1 | accepted | 3.760 | 3.76x | 0.0440 ms | $0.0036 |
| #19 | fused_mlp | persistent_blocks_v3 | accepted | 3.420 | 3.42x | 0.0520 ms | $0.0034 |
| #18 | fused_mlp | persistent_blocks_v2 correct but not best | rejected | 2.440 | 2.44x | 0.2130 ms | $0.0032 |
| #17 | fused_mlp | fuse_gelu_matmul | accepted | 2.910 | 2.91x | 0.1790 ms | $0.0031 |