| {"ts": "2025-12-19T18:57:39Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B1_S512_E2", "batch": 1, "seq_len": 512, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 2, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 155.7981640000321, "p50": 157.7297640000097, "p90": 159.48504900001126, "mean": 158.39911260001145, "iqr": 2.223896000032255, "raw_times": [161.72143300002517, 157.261152999979, 155.7981640000321, 159.48504900001126, 157.7297640000097], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 159.10347999999885, "peak_bytes": 416866816, "ok": true, "absmax": 2.765655517578125e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 2.765655517578125e-05, "mae": 2.0696452338597737e-06, "mse": 7.332408985538663e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T18:58:03Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B1_S512_E4", "batch": 1, "seq_len": 512, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 4, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 199.79041199997027, "p50": 204.82147100000248, "p90": 205.0451750000093, "mean": 203.32668460000605, "iqr": 3.4747309999829668, "raw_times": [205.40592100002186, 199.79041199997027, 201.57044400002633, 205.0451750000093, 204.82147100000248], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 200.72428899999295, "peak_bytes": 632035840, "ok": true, "absmax": 1.621246337890625e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 1.621246337890625e-05, "mae": 9.61917862696282e-07, "mse": 1.59423277530657e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T18:58:47Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B1_S1024_E2", "batch": 1, "seq_len": 1024, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 2, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 372.8170420000083, "p50": 383.31174900002907, "p90": 392.9121939999618, "mean": 385.07766660000016, "iqr": 10.251173999961338, "raw_times": [393.68632800000114, 392.9121939999618, 382.66102000000046, 383.31174900002907, 372.8170420000083], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 393.062126000018, "peak_bytes": 643844608, "ok": true, "absmax": 2.6226043701171875e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 2.6226043701171875e-05, "mae": 2.0501920516835526e-06, "mse": 7.1848811622476916e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T18:59:36Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B1_S1024_E4", "batch": 1, "seq_len": 1024, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 4, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 418.8624209999716, "p50": 421.41534400002456, "p90": 422.4395519999007, "mean": 421.30189059998884, "iqr": 1.8283119999296105, "raw_times": [423.18089600007625, 421.41534400002456, 418.8624209999716, 420.6112399999711, 422.4395519999007], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 421.8970150000132, "peak_bytes": 823386112, "ok": true, "absmax": 1.3589859008789062e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 1.3589859008789062e-05, "mae": 9.400179123986163e-07, "mse": 1.5130355735665235e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T19:01:05Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B4_S512_E2", "batch": 4, "seq_len": 512, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 2, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 766.098573000022, "p50": 773.6994500000947, "p90": 774.865274999911, "mean": 772.9573942000115, "iqr": 8.746404999897095, "raw_times": [766.1188700000139, 773.6994500000947, 766.098573000022, 774.865274999911, 784.0048030000162], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 782.9079639999463, "peak_bytes": 1036112384, "ok": true, "absmax": 3.2901763916015625e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 3.2901763916015625e-05, "mae": 2.0572656467265915e-06, "mse": 7.247809123700488e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T19:02:49Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B4_S512_E4", "batch": 4, "seq_len": 512, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 4, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 828.9166780000414, "p50": 840.0145479999992, "p90": 848.174653000001, "mean": 841.7884347999916, "iqr": 11.353517000088686, "raw_times": [855.0151590000041, 828.9166780000414, 848.174653000001, 836.8211359999123, 840.0145479999992], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 857.4785790000305, "peak_bytes": 1235263488, "ok": true, "absmax": 1.430511474609375e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 1.430511474609375e-05, "mae": 9.400343401466671e-07, "mse": 1.5107844445957919e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T19:05:50Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B4_S1024_E2", "batch": 4, "seq_len": 1024, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 2, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 1492.7651169999763, "p50": 1513.7102520000099, "p90": 1522.1755649999977, "mean": 1513.4781133999923, "iqr": 10.99431700004061, "raw_times": [1492.7651169999763, 1511.1812479999571, 1522.1755649999977, 1527.5583850000203, 1513.7102520000099], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 1532.0516410000664, "peak_bytes": 1861947904, "ok": true, "absmax": 2.6226043701171875e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 2.6226043701171875e-05, "mae": 2.060702854578267e-06, "mse": 7.262949790198814e-12, "ref": "naive_moe"}, "err": null} | |
| {"ts": "2025-12-19T19:09:07Z", "run": "aec86efe9ed1483a979a8427ac940bd6", "impl": "binned_torch", "tags": {"family": "pytorch", "backend": "eager"}, "wl": {"name": "cuda_B4_S1024_E4", "batch": 4, "seq_len": 1024, "hidden_dim": 2880, "expert_dim": 5760, "num_experts": 4, "top_k": 2, "dtype": "float32", "device": "cuda"}, "env": {"torch": "2.8.0+cu128", "cuda": "12.8", "gpu": "NVIDIA L40S", "sm": "8.9", "py": "3.11.14", "plat": "Linux-6.12.58-82.121.amzn2023.x86_64-x86_64-with-glibc2.35"}, "lat_ms": {"p10": 1654.5569229999728, "p50": 1658.7427389999903, "p90": 1665.0588319999997, "mean": 1660.4780848000016, "iqr": 7.11779099992782, "raw_times": [1658.7427389999903, 1665.0588319999997, 1666.0908889999746, 1657.941041000072, 1654.5569229999728], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 1670.381679000002, "peak_bytes": 2062163968, "ok": true, "absmax": 1.5974044799804688e-05, "corr": {"ok": true, "rtol": 0.01, "atol": 0.01, "absmax": 1.5974044799804688e-05, "mae": 9.529014732834185e-07, "mse": 1.5621694476192216e-12, "ref": "naive_moe"}, "err": null} | |