Entry #3
Completed- Hotkey
- 5ELubE…xWae9X
- Sent
- 10 Nov 2026, 18:06 UTC
- Source
- bfc02bbd753a
Progress
- Queued
- Proof
- Planned
- Timing
- Done
Done means every board has a decision. It does not mean the entry took a board over.
On each board
| Board | Runs | Time | Decision |
|---|---|---|---|
| Overallvs #8 | Done2 of 2 runs | 2.400×Generated 2.390×, real models 2.409× | Above the barGain lower bound -165.0%Guard passed |
| MLPvs #1 | Done2 of 2 runs | 0.994×Generated 1.011×, real models 0.977× | Above the barGain lower bound -8.3%Guard passed |
| ResNetvs #2 | Done2 of 2 runs | 0.524×Generated 0.311×, real models 0.884× | QualifiesGain lower bound 35.4%Guard passed |
| MobileNetvs the shipped reference | Done2 of 2 runs | 0.899×Generated 0.625×, real models 1.295× | Not significantGain lower bound -11.1%Guard passed |
| Normalized MLPvs #6 | Done2 of 2 runs | 41.102×Generated 54.439×, real models 31.032× | Above the barGain lower bound -5176.7%Guard passed |
| Mixervs #7 | Done2 of 2 runs | 105.119×Generated 162.815×, real models 67.868× | Above the barGain lower bound -13769.3%Guard passed |
| Recurrentvs the shipped reference | Done2 of 2 runs | 0.998×Generated 0.995×, real models 1.001× | Above the barGain lower bound -0.4%Guard passed |
Every run
Each board times the entry on every benchmark machine, over fresh seeds, on generated workloads and on trained models with real data. A run's time combines the two suites by their published weights; the bars below split it by model, batch and input, and by real model.
Overall2 runs
RunsGenerated 2.390× · real models 2.409×
MachineTime
- prod-amd64Seed 5486203562.358×Done1 attempt0.337× of the baseline kernels' time
- prod-arm64Seed 5486203562.442×Done1 attempt0.364× of the baseline kernels' time
By model and inputGuard at 1.100×
MLP
- Batch 1, normal1.020×
- Batch 1, sparse1.039×
- Batch 1, wide1.251×
Mixer
- Batch 1, normal142.830×
- Batch 1, sparse152.053×
- Batch 1, wide183.872×
ResNet
- Batch 1, normal0.314×
- Batch 1, sparse0.324×
- Batch 1, wide0.310×
MobileNet
- Batch 1, normal0.525×
- Batch 1, sparse0.531×
- Batch 1, wide0.523×
Recurrent
- Batch 1, normal0.114×
- Batch 1, sparse0.113×
- Batch 1, wide0.117×
Normalized MLP
- Batch 1, normal58.392×
- Batch 1, sparse58.843×
- Batch 1, wide61.076×
Real models
- appliance-energy-mixer on Appliance Energy Prediction210.197×
- compact-mixer on digits48.822×
- compact-mixer on fashion24.140×
- compact-mixer on mnist22.859×
- compact-mlp on digits1.403×
- compact-mlp on fashion1.055×
- compact-mlp on mnist0.871×
- compact-mobilenet on digits0.747×
- compact-mobilenet on fashion1.438×
- compact-mobilenet on mnist2.693×
- compact-normalized_mlp on digits79.349×
- compact-normalized_mlp on fashion30.409×
- compact-normalized_mlp on mnist19.824×
- compact-resnet on digits0.765×
- compact-resnet on fashion0.911×
- compact-resnet on mnist2.475×
- appliance-gru on appliance-energy0.119×
- mitbih-gru on mitbih0.102×
- mnist-rows-lstm on mnist-rows0.104×
- speech-commands-gru on speech-commands0.146×
- speech-commands-lstm-gru on speech-commands0.127×
- spoken-arabic-digit-lstm on spoken-arabic-digit0.094×
- tiny-shakespeare-lstm on tiny-shakespeare0.125×
- uci-har-gru on uci-har0.114×
- uci-har-lstm on uci-har0.092×
- uci-har-mixer on UCI-HAR126.975×
- uci-har-mlp on UCI-HAR0.985×
- uci-har-mobilenet on UCI-HAR0.465×
- uci-har-normalized_mlp on UCI-HAR18.763×
- uci-har-resnet on UCI-HAR0.286×
MLP2 runs
RunsGenerated 1.011× · real models 0.977×
MachineTime
- prod-amd64Seed 5486203560.992×Done1 attempt0.017× of the baseline kernels' time
- prod-arm64Seed 5486203560.996×Done1 attempt0.018× of the baseline kernels' time
By model and inputGuard at 1.100×
MLP
- Batch 1, normal1.025×
- Batch 1, sparse1.013×
- Batch 1, wide1.020×
Real models
- compact-mlp on digits0.963×
- compact-mlp on fashion0.994×
- compact-mlp on mnist1.011×
- uci-har-mlp on UCI-HAR0.941×
ResNet2 runs
RunsGenerated 0.311× · real models 0.884×
MachineTime
- prod-amd64Seed 5486203560.514×Done1 attempt0.292× of the baseline kernels' time
- prod-arm64Seed 5486203560.534×Done1 attempt0.319× of the baseline kernels' time
By model and inputGuard at 1.100×
ResNet
- Batch 1, normal0.315×
- Batch 1, sparse0.321×
- Batch 1, wide0.308×
Real models
- compact-resnet on digits0.759×
- compact-resnet on fashion0.907×
- compact-resnet on mnist2.455×
- uci-har-resnet on UCI-HAR0.360×
MobileNet2 runs
RunsGenerated 0.625× · real models 1.295×
MachineTime
- prod-amd64Seed 5486203560.896×Done1 attempt0.622× of the baseline kernels' time
- prod-arm64Seed 5486203560.902×Done1 attempt0.631× of the baseline kernels' time
By model and inputGuard at 1.100×
MobileNet
- Batch 1, normal0.626×
- Batch 1, sparse0.634×
- Batch 1, wide0.618×
Real models
- compact-mobilenet on digits0.920×
- compact-mobilenet on fashion1.752×
- compact-mobilenet on mnist3.083×
- uci-har-mobilenet on UCI-HAR0.566×
Normalized MLP2 runs
RunsGenerated 54.439× · real models 31.032×
MachineTime
- prod-amd64Seed 54862035641.161×Done1 attempt0.572× of the baseline kernels' time
- prod-arm64Seed 54862035641.042×Done1 attempt0.581× of the baseline kernels' time
By model and inputGuard at 1.100×
Normalized MLP
- Batch 1, normal61.225×
- Batch 1, sparse57.942×
- Batch 1, wide57.569×
Real models
- compact-normalized_mlp on digits93.666×
- compact-normalized_mlp on fashion21.156×
- compact-normalized_mlp on mnist23.431×
- uci-har-normalized_mlp on UCI-HAR19.972×
Mixer2 runs
RunsGenerated 162.815× · real models 67.868×
MachineTime
- prod-amd64Seed 548620356106.557×Done1 attempt0.785× of the baseline kernels' time
- prod-arm64Seed 548620356103.700×Done1 attempt0.816× of the baseline kernels' time
By model and inputGuard at 1.100×
Mixer
- Batch 1, normal158.734×
- Batch 1, sparse154.991×
- Batch 1, wide191.028×
Real models
- appliance-energy-mixer on Appliance Energy Prediction208.654×
- compact-mixer on digits64.618×
- compact-mixer on fashion32.595×
- compact-mixer on mnist25.346×
- uci-har-mixer on UCI-HAR129.264×
Recurrent2 runs
RunsGenerated 0.995× · real models 1.001×
MachineTime
- prod-amd64Seed 5486203560.999×Done1 attempt1.003× of the baseline kernels' time
- prod-arm64Seed 5486203560.997×Done1 attempt0.997× of the baseline kernels' time
By model and inputGuard at 1.100×
Recurrent
- Batch 1, normal0.991×
- Batch 1, sparse0.999×
- Batch 1, wide0.995×
Real models
- appliance-gru on appliance-energy1.005×
- mitbih-gru on mitbih1.008×
- mnist-rows-lstm on mnist-rows0.988×
- speech-commands-gru on speech-commands1.011×
- speech-commands-lstm-gru on speech-commands0.998×
- spoken-arabic-digit-lstm on spoken-arabic-digit0.988×
- tiny-shakespeare-lstm on tiny-shakespeare0.991×
- uci-har-gru on uci-har1.036×
- uci-har-lstm on uci-har0.985×