| { |
| "op": "com.microsoft.BiasGelu", |
| "tunableSpace": { "WORKGROUP_SIZE": [64, 128, 256] }, |
| "cases": [ |
| { |
| "name": "biasgelu-f32-4096x3072", |
| "preset": "smoke", |
| "vars": { "dtype": "float32" }, |
| "inputs": { |
| "A": { "shape": [4096, 3072], "dtype": "float32", "dist": "normal", "seed": 420, "scale": 2 }, |
| "B": { "shape": [3072], "dtype": "float32", "dist": "normal", "seed": 421, "scale": 1 } |
| }, |
| "outputs": { "C": { "shape": [4096, 3072], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "4096 * 3072 * 2 * dtypeBytes(args.dtype)" }] } |
| }, |
| { |
| "name": "biasgelu-f32-scalar-large-odd-hidden", |
| "preset": "stress", |
| "vars": { "dtype": "float32" }, |
| "inputs": { |
| "A": { "shape": [2097152, 3], "dtype": "float32", "dist": "normal", "seed": 430, "scale": 2 }, |
| "B": { "shape": [3], "dtype": "float32", "dist": "normal", "seed": 431, "scale": 1 } |
| }, |
| "outputs": { "C": { "shape": [2097152, 3], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "2097152 * 3 * 2 * dtypeBytes(args.dtype)" }] } |
| }, |
| { |
| "name": "biasgelu-bert-base-b8-s512-i3072", |
| "preset": "model", |
| "provenance": { |
| "notes": "BERT class defaults (hidden_size 768, intermediate_size 3072): the FFN bias+GELU for a batch of 8 full-length sequences." |
| }, |
| "vars": { "dtype": "float32", "rows": 4096, "inter": 3072 }, |
| "inputs": { |
| "A": { "shape": [4096, 3072], "dtype": "float32", "dist": "normal", "seed": 5500, "scale": 0.5 }, |
| "B": { "shape": [3072], "dtype": "float32", "dist": "normal", "seed": 5501, "scale": 0.1 } |
| }, |
| "outputs": { "C": { "shape": [4096, 3072], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "args.rows * args.inter * 2 * dtypeBytes(args.dtype)" }] } |
| }, |
| { |
| "name": "biasgelu-vit-base-b64-p197-i3072", |
| "preset": "model", |
| "provenance": { |
| "notes": "ViT class defaults (hidden_size 768, intermediate_size 3072, image_size 224, patch_size 16 -> 196 patches + CLS) at a 64-image batch." |
| }, |
| "vars": { "dtype": "float32", "rows": 12608, "inter": 3072 }, |
| "inputs": { |
| "A": { "shape": [12608, 3072], "dtype": "float32", "dist": "normal", "seed": 5600, "scale": 0.5 }, |
| "B": { "shape": [3072], "dtype": "float32", "dist": "normal", "seed": 5601, "scale": 0.1 } |
| }, |
| "outputs": { "C": { "shape": [12608, 3072], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "args.rows * args.inter * 2 * dtypeBytes(args.dtype)" }] } |
| }, |
| { |
| "name": "biasgelu-t5-base-b8-s512-i2048", |
| "preset": "model", |
| "provenance": { "notes": "T5 class defaults (d_model 512, d_ff 2048) at a batch of 8 full-length sequences." }, |
| "vars": { "dtype": "float32", "rows": 4096, "inter": 2048 }, |
| "inputs": { |
| "A": { "shape": [4096, 2048], "dtype": "float32", "dist": "normal", "seed": 5700, "scale": 0.5 }, |
| "B": { "shape": [2048], "dtype": "float32", "dist": "normal", "seed": 5701, "scale": 0.1 } |
| }, |
| "outputs": { "C": { "shape": [4096, 2048], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "args.rows * args.inter * 2 * dtypeBytes(args.dtype)" }] } |
| }, |
| { |
| "name": "biasgelu-deberta-v2-b8-s512-i6144", |
| "preset": "model", |
| "provenance": { "notes": "DeBERTa-v2 class defaults (hidden_size 1536, intermediate_size 6144)." }, |
| "vars": { "dtype": "float32", "rows": 4096, "inter": 6144 }, |
| "inputs": { |
| "A": { "shape": [4096, 6144], "dtype": "float32", "dist": "normal", "seed": 5800, "scale": 0.5 }, |
| "B": { "shape": [6144], "dtype": "float32", "dist": "normal", "seed": 5801, "scale": 0.1 } |
| }, |
| "outputs": { "C": { "shape": [4096, 6144], "dtype": "float32" } }, |
| "bench": { "metrics": [{ "type": "bandwidth", "value": "args.rows * args.inter * 2 * dtypeBytes(args.dtype)" }] } |
| } |
| ] |
| } |
|
|
