| {"ts": "2026-01-04T21:51:00Z", "run": "508523ec81a34de19375bc1e9bf856ae", "impl": "torch_layer_norm", "tags": {"family": "torch", "op": "layer_norm"}, "wl": {"name": "LN_B16_S2048_D4096", "batch": 16, "seq_len": 2048, "hidden_dim": 4096, "dtype": "bfloat16", "device": "cuda", "seed": 0}, "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": 0.8105489998797566, "p50": 0.8139180001762725, "p90": 0.8163179998064152, "mean": 0.8145365999553178, "iqr": 0.0031489998946199194, "raw_times": [0.8187290000023495, 0.8105489998797566, 0.8163179998064152, 0.8139180001762725, 0.8131689999117953], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 0.8199989999866375, "peak_bytes": 2415935488, "ok": true, "absmax": 0.03125, "corr": {"ok": true, "rtol": 0.001, "atol": 0.03125, "absmax": 0.03125, "mae": 0.0015411376953125, "mse": 1.1205673217773438e-05, "ref": "layer_norm_ref"}, "err": null} | |
| {"ts": "2026-01-04T21:51:00Z", "run": "508523ec81a34de19375bc1e9bf856ae", "impl": "torch_layer_norm", "tags": {"family": "torch", "op": "layer_norm"}, "wl": {"name": "LN_B16_S2048_D8192", "batch": 16, "seq_len": 2048, "hidden_dim": 8192, "dtype": "bfloat16", "device": "cuda", "seed": 0}, "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": 1.663946999997279, "p50": 1.6735270000935998, "p90": 1.6742169998451573, "mean": 1.6716250000172295, "iqr": 0.0027299997782392893, "raw_times": [1.6742169998451573, 1.663946999997279, 1.6735270000935998, 1.6749470000831934, 1.671487000066918], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 1.675217000183693, "peak_bytes": 4831870976, "ok": true, "absmax": 0.03125, "corr": {"ok": true, "rtol": 0.001, "atol": 0.03125, "absmax": 0.03125, "mae": 0.0015106201171875, "mse": 1.1086463928222656e-05, "ref": "layer_norm_ref"}, "err": null} | |
| {"ts": "2026-01-04T21:51:00Z", "run": "508523ec81a34de19375bc1e9bf856ae", "impl": "torch_layer_norm", "tags": {"family": "torch", "op": "layer_norm"}, "wl": {"name": "LN_B16_S4096_D4096", "batch": 16, "seq_len": 4096, "hidden_dim": 4096, "dtype": "bfloat16", "device": "cuda", "seed": 0}, "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": 1.6077049999694282, "p50": 1.6113560000121652, "p90": 1.6143249999913678, "mean": 1.6125273999932688, "iqr": 0.004708999995273189, "raw_times": [1.6096159999960946, 1.6113560000121652, 1.6143249999913678, 1.6077049999694282, 1.619634999997288], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 1.632315999813727, "peak_bytes": 4831854592, "ok": true, "absmax": 0.03125, "corr": {"ok": true, "rtol": 0.001, "atol": 0.03125, "absmax": 0.03125, "mae": 0.0015411376953125, "mse": 1.1205673217773438e-05, "ref": "layer_norm_ref"}, "err": null} | |
| {"ts": "2026-01-04T21:51:01Z", "run": "508523ec81a34de19375bc1e9bf856ae", "impl": "torch_layer_norm", "tags": {"family": "torch", "op": "layer_norm"}, "wl": {"name": "LN_B16_S4096_D8192", "batch": 16, "seq_len": 4096, "hidden_dim": 8192, "dtype": "bfloat16", "device": "cuda", "seed": 0}, "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": 3.314662000093449, "p50": 3.3266129998992255, "p90": 3.335402999937287, "mean": 3.327090999982829, "iqr": 0.013279999848236912, "raw_times": [3.314662000093449, 3.336653999895134, 3.335402999937287, 3.3266129998992255, 3.32212300008905], "has_warnings": false, "reps": 5, "warmup": 2}, "compile_ms": 3.3250040000893932, "peak_bytes": 9663709184, "ok": true, "absmax": 0.03125, "corr": {"ok": true, "rtol": 0.001, "atol": 0.03125, "absmax": 0.03125, "mae": 0.0015106201171875, "mse": 1.1026859283447266e-05, "ref": "layer_norm_ref"}, "err": null} | |