temp / kernel-design-agents /ablation-plots /coverage_summary.csv
Dongyun Zou
Render latency progress as continuous best-so-far curves
9ebee69 verified
Raw
History Blame Contribute Delete
6.83 kB
kernel,model,workflow,benchmark_csv,valid_full_runs,token_mapped_runs,token_source,tokens_at_first_mapped_run,tokens_at_last_mapped_run,first_valid_utc,last_valid_utc,observed_hours,best_commit,best_git_dirty,best_mean_kernel_ms,best_mean_baseline_ms,best_arithmetic_mean_speedup
gdn-prefill,Opus 4.8,goal,claude-opus-4.8/goal/gdn-prefill/benchmark.csv,1,1,Claude session JSONL (main + subagents),65028497,65028497,2026-07-06T14:00:38.881154+00:00,2026-07-06T14:00:38.881154+00:00,0.0000,21b180195ae51210c747d9a33180d87fb8f3888c,False,5.77528336,0.227514026,0.382483417
gdn-prefill,Opus 4.8,kda-humanize,claude-opus-4.8/kda-humanize/gdn-prefill/benchmark.csv,4,4,OMH artifact usage (22 transcripts),30396077,75938257,2026-07-06T18:32:43.468243+00:00,2026-07-07T02:32:32.555402+00:00,7.9970,61aad40f08c2f9e3b5e810e388d115104560b73a,False,0.592703162,0.227289866,1.78860515
gdn-prefill,Fable 5,goal,fable-5/goal/gdn-prefill/benchmark.csv,2,2,Claude session JSONL (main + subagents),55597399,92019233,2026-07-06T13:49:20.639118+00:00,2026-07-06T15:44:10.756930+00:00,1.9139,cdc89dda8a6c5d96204d5517d59a1dd2d878568a,False,0.14419803,0.227744266,2.67217695
gdn-prefill,Fable 5,kda-humanize,fable-5/kda-humanize/gdn-prefill/benchmark.csv,2,2,OMH artifact usage (11 transcripts),3038128,12504817,2026-07-06T15:29:52.088168+00:00,2026-07-06T17:54:53.565569+00:00,2.4171,5dc9728ba7bb893a2eee82441c1bd67f1a945921,False,0.256144033,0.227761119,2.35759475
gdn-prefill,GPT-5.5,goal-arar,gpt-5.5/goal-arar/gdn-prefill/benchmark.csv,2,2,Codex token_count events (4 sessions),131672977,676780886,2026-07-06T14:17:40.543429+00:00,2026-07-07T00:19:57.421397+00:00,10.0380,2aab298f50b62bf285d3febf0be2320213ea41e6,False,0.853272695,0.225297919,1.42390791
gdn-prefill,GPT-5.5,kda-humanize,gpt-5.5/kda-humanize/gdn-prefill/benchmark.csv,4,4,OMH artifact usage (38 transcripts),24602414,45558574,2026-07-06T18:43:21.407325+00:00,2026-07-06T23:03:28.828113+00:00,4.3354,3301df9c2411f3fde7ee0bf4af894ca33f0d3040,False,0.599637279,0.225598293,1.45446555
gdn-prefill,GPT-5.6-sol,goal-arar,gpt-5.6-sol/goal-arar/gdn-prefill/benchmark.csv,5,5,Codex token_count events (20 sessions),89823015,786403380,2026-07-06T12:30:02.303945+00:00,2026-07-06T15:33:17.269459+00:00,3.0542,ea028bdafc41d08dcf38380ddcb09f56ba9a58ad,False,0.502797066,0.226680479,1.46495053
gdn-prefill,GPT-5.6-sol,kda-humanize,gpt-5.6-sol/kda-humanize/gdn-prefill/benchmark.csv,8,8,OMH artifact usage (71 transcripts),7833934,75297959,2026-07-06T17:01:53.504293+00:00,2026-07-06T23:34:23.702943+00:00,6.5417,f9bffb882d5587fe4dcfe2b2321b7e86c94ca6c7,False,0.786903039,0.226488426,0.620070777
fp8-moe,Opus 4.8,goal,claude-opus-4.8/goal/fp8-moe/benchmark.csv,11,11,Claude session JSONL (main + subagents),144996171,303535066,2026-07-06T15:26:39.589024+00:00,2026-07-06T19:05:47.907189+00:00,3.6523,5c671823569a29d2cf6aa028ac52811c6fa7b0f4,False,0.372966062,0.368777684,0.909576529
fp8-moe,Opus 4.8,kda-humanize,claude-opus-4.8/kda-humanize/fp8-moe/benchmark.csv,5,5,OMH artifact usage (30 transcripts),32556431,98937654,2026-07-06T19:05:01.297381+00:00,2026-07-07T01:44:33.774599+00:00,6.6590,5e67008774d4d6f28f8da849d8bd66a14c0f790d,False,61.5741685,0.369880421,0.0397346807
fp8-moe,Fable 5,goal,fable-5/goal/fp8-moe/benchmark.csv,2,2,Claude session JSONL (main + subagents),175263184,216414729,2026-07-06T16:12:38.097629+00:00,2026-07-06T16:57:03.329737+00:00,0.7403,adfe8fda6cf6537b1e74e4fc8a6c6f7baf1e652d,False,0.270075789,0.368493054,1.20165542
fp8-moe,Fable 5,kda-humanize,fable-5/kda-humanize/fp8-moe/benchmark.csv,10,10,OMH artifact usage (36 transcripts),4623411,116464997,2026-07-06T16:17:57.354762+00:00,2026-07-06T23:21:05.618637+00:00,7.0523,694a84bd4608923354f4b4ae5b1cfe572ddd2e9f,False,0.319949726,0.373259368,1.03953605
fp8-moe,GPT-5.5,goal-arar,gpt-5.5/goal-arar/fp8-moe/benchmark.csv,17,17,Codex token_count events (3 sessions),36454232,215611168,2026-07-06T15:37:26.277254+00:00,2026-07-07T02:34:01.624292+00:00,10.9432,2f16cd95a6ae3dadc8b8ca4b08c2646160c8a266,False,0.218006232,0.375112839,1.58618403
fp8-moe,GPT-5.5,kda-humanize,missing,0,0,missing benchmark.csv,,,,,,,,,,
fp8-moe,GPT-5.6-sol,goal-arar,gpt-5.6-sol/goal-arar/fp8-moe/benchmark.csv,26,26,Codex token_count events (41 sessions),389715605,7982751358,2026-07-06T14:51:13.921596+00:00,2026-07-07T00:41:53.242072+00:00,9.8443,cf12fa1e7b6d7539e6ad2f4604893311ffdc6585,False,0.2192032,0.369058525,1.5040291
fp8-moe,GPT-5.6-sol,kda-humanize,gpt-5.6-sol/kda-humanize/fp8-moe/benchmark.csv,6,6,OMH artifact usage (62 transcripts),23433699,137194406,2026-07-06T18:35:50.015451+00:00,2026-07-07T00:20:38.788601+00:00,5.7469,ebf12746f42318a3c22d8445149d139e4571e993,False,0.2814208,0.36801768,1.14459011
dsa,Opus 4.8,goal,claude-opus-4.8/goal/dsa/benchmark.csv,8,8,Claude session JSONL (main + subagents),141961355,190012558,2026-07-06T15:01:20.929114+00:00,2026-07-06T16:43:15.489743+00:00,1.6985,673cf877e2ce4f07428631ef095780e6f73b693c,False,0.122307884,0.0167501445,0.147988389
dsa,Opus 4.8,kda-humanize,claude-opus-4.8/kda-humanize/dsa/benchmark.csv,11,11,OMH artifact usage (35 transcripts),3730197,102808552,2026-07-06T14:40:29.600365+00:00,2026-07-07T01:11:52.129468+00:00,10.5229,0c9df0b7d42db55ffa570be91773c63112b15ea9,False,0.0541799415,0.0167828402,0.388416466
dsa,Fable 5,goal,fable-5/goal/dsa/benchmark.csv,13,12,Claude session JSONL (main + subagents),32782068,206717175,2026-07-06T13:15:55.457927+00:00,2026-07-06T23:06:38.399363+00:00,9.8453,5c6c0d4d41633a13b4bc30d8a0fc117468c008c0,False,0.0104256449,0.0167146664,1.66470205
dsa,Fable 5,kda-humanize,fable-5/kda-humanize/dsa/benchmark.csv,9,7,OMH artifact usage (42 transcripts),8254326,75502626,2026-07-06T13:54:17.591404+00:00,2026-07-06T21:38:38.549732+00:00,7.7392,1c7bc212dea454567d6bd6f8d2eb3e5a146f95ba,False,0.0608677609,0.0161713622,0.353639151
dsa,GPT-5.5,goal-arar,gpt-5.5/goal-arar/dsa/benchmark.csv,0,0,Codex token_count events (2 sessions),,,,,,,,,,
dsa,GPT-5.5,kda-humanize,gpt-5.5/kda-humanize/dsa/benchmark.csv,1,1,OMH artifact usage (50 transcripts),51015214,51015214,2026-07-07T00:21:13.160225+00:00,2026-07-07T00:21:13.160225+00:00,0.0000,a7cabad9884d655a0fb3597d11e51654f588fd9d,False,0.0399033044,0.0161657975,0.486637124
dsa,GPT-5.6-sol,goal-arar,gpt-5.6-sol/goal-arar/dsa/benchmark.csv,37,37,Codex token_count events (44 sessions),81845850,13146120285,2026-07-06T12:27:49.000743+00:00,2026-07-06T23:48:24.252996+00:00,11.3431,3f5e8a911908d815538ed33707bd349a0bad83f4,False,0.0203512754,0.0160340874,1.03661013
dsa,GPT-5.6-sol,kda-humanize,gpt-5.6-sol/kda-humanize/dsa/benchmark.csv,14,14,OMH artifact usage (80 transcripts),12874775,190682034,2026-07-06T16:34:25.720657+00:00,2026-07-07T01:40:35.802645+00:00,9.1028,c20bc5aa1df02bc1c7044bbe42bd0617cd2a0f15,False,0.10661658,0.0160361743,0.52692928