{ "benchmark_config": { "iterations": 1000, "cold_iterations": 1000, "benchmarks": [ { "name": "gemm", "function": "gemm", "log_file": "rocblas-gemm_bench.log", "dimensions": "sizes", "extra_args": { "h": { "alpha": "1", "beta": "0", "transposeA": "N" } } }, { "name": "gemv", "function": "gemv", "log_file": "rocblas-gemv_bench.log", "dimensions": "separate" }, { "name": "ger", "function": "ger", "log_file": "rocblas-ger_bench.log", "dimensions": "separate" }, { "name": "dot", "function": "dot", "log_file": "rocblas-dot_bench.log", "dimensions": "simple" }, { "name": "gemm_hpa_hgemm", "function": "gemm_ex", "log_file": "rocblas-gemm_hpa_hgemm_bench.log", "dimensions": "sizes", "has_compute_type": true } ] }, "gemm": { "sizes": [ 2047, 2048, 4095, 4096, 5503, 5504, 5760, 6143, 6144, 7744, 8191, 8192 ], "transpose": [ "T", "N" ], "precision": [ "s", "d", "h" ], "alpha": 1, "beta": 0 }, "gemv": { "m": [ 1000, 10001 ], "n": [ 1000, 10001 ], "lda": [ 1000, 10002 ], "transpose": [ "N", "T" ], "precision": [ "s", "d" ], "gpu_overrides": { "gfx110X": { "m": [ 30001, 20001 ], "n": [ 30001, 20001 ], "lda": [ 30002, 20001 ] }, "gfx94X-dcgpu": { "m": [ 30001, 20001 ], "n": [ 30001, 20001 ], "lda": [ 30002, 20001 ] } } }, "ger": { "m": [ 1000, 10001 ], "n": [ 1000, 10001 ], "lda": [ 1000, 10002 ], "precision": [ "s", "d" ], "gpu_overrides": { "gfx110X": { "m": [ 30001, 20001 ], "n": [ 30001, 20001 ], "lda": [ 30002, 20001 ] }, "gfx94X-dcgpu": { "m": [ 30001, 20001 ], "n": [ 30001, 20001 ], "lda": [ 30002, 20001 ] } } }, "dot": { "n": [ 1000, 100001, 10000002 ], "precision": [ "s", "d" ], "gpu_overrides": { "gfx110X": { "n": [ 1000, 100000002, 50000002 ] }, "gfx94X-dcgpu": { "n": [ 1000, 100000002, 50000002 ] } } }, "gemm_hpa_hgemm": { "sizes": [ 2047, 2048, 4095, 4096, 5503, 5504, 5760, 6143, 6144, 7744, 8191, 8192 ], "transpose": [ "T", "N" ], "precision": "h", "compute_type": "s" } }