Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix alignment check for non 16-byte alignments #175

Merged
merged 1 commit into from
Nov 15, 2023

Conversation

thomasfaingnaert
Copy link
Member

No description provided.

@maleadt
Copy link
Member

maleadt commented Nov 15, 2023

Benchmark results for commit 8caa172 (comparing to 912fa05):

test master PR Δmin
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.16 μs ± 1.73% (8.82 … 9.54 μs) 8.11 μs ± 2.02% (7.87 … 8.34 μs) -10.8% ✅
FPU GEMM Float32*Float32=Float64 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 180.0 μs ± 0.704% (175.0 … 181.0 μs) 156.0 μs ± 0.68% (155.0 … 159.0 μs) -11.1% ✅
WMMA GEMM Float16*Float16+Float16=Float16 (2048×2048) · (2048×2048) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 221.0 μs ± 4.43% (207.0 … 239.0 μs) 231.0 μs ± 1.07% (224.0 … 238.0 μs) +8.2% ❌
FPU GEMM Float64*Float64=Float64 (128×128) · (128×128) (NN) OP (8, 16, 2), base shape (4, 8, 1) 144.0 μs ± 4.94% (129.0 … 149.0 μs) 148.0 μs ± 0.156% (147.0 … 149.0 μs) +13.6% ❌
FPU GEMM Float32*Float32=Float64 (128×128) · (128×128) (NT) OP (8, 16, 2), base shape (4, 8, 1) 178.0 μs ± 0.556% (175.0 … 180.0 μs) 156.0 μs ± 0.462% (154.0 … 158.0 μs) -11.9% ✅
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.11 μs ± 1.82% (8.82 … 9.3 μs) 8.05 μs ± 2.06% (7.87 … 8.34 μs) -10.8% ✅
WMMA GEMM+bias Float16*Float16+Float32=Float32 (128×128) · (128×128) (TT) OP (16, 16, 16) 10.3 μs ± 1.76% (9.78 … 11.0 μs) 9.04 μs ± 1.9% (8.82 … 9.54 μs) -9.8% ✅
WMMA diagonal GEMM diag(Float16)*Float16=Float32 (128×128) · (128×128) (NN) OP (16, 16, 16) 11.2 μs ± 1.14% (11.0 … 11.7 μs) 9.94 μs ± 1.24% (9.54 … 10.0 μs) -13.0% ✅
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.42 μs ± 1.68% (8.11 … 8.82 μs) 7.76 μs ± 6.46% (7.15 … 9.06 μs) -11.8% ✅
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.99 μs ± 1.61% (8.58 … 9.3 μs) 7.95 μs ± 1.61% (7.63 … 8.34 μs) -11.1% ✅
Tropical GEMM Float32*Float32=Float32 (128×256) · (256×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 122.0 μs ± 5.07% (109.0 … 126.0 μs) 125.0 μs ± 0.129% (125.0 … 126.0 μs) +13.9% ❌
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.11 μs ± 1.83% (8.82 … 9.3 μs) 8.0 μs ± 2.02% (7.87 … 8.34 μs) -10.8% ✅
Tropical GEMM Float32*Float32=Float32 (256×256) · (256×256) (NT) OP (8, 16, 2), base shape (4, 8, 1) 127.0 μs ± 0.137% (127.0 … 128.0 μs) 112.0 μs ± 0.147% (112.0 … 113.0 μs) -11.8% ✅
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 7.8 μs ± 2.14% (7.63 … 8.11 μs) 6.88 μs ± 1.3% (6.44 … 7.15 μs) -15.6% ✅
Tropical GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 56.0 μs ± 0.252% (55.8 … 56.5 μs) 63.8 μs ± 0.266% (63.4 … 64.1 μs) +13.7% ❌
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 8, 1), base shape (4, 8, 1) 15.3 μs ± 1.1% (15.0 … 15.7 μs) 13.5 μs ± 1.07% (13.1 … 13.8 μs) -12.7% ✅
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.47 μs ± 1.78% (8.11 … 8.82 μs) 7.47 μs ± 1.84% (7.15 … 7.87 μs) -11.8% ✅
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 9.27 μs ± 1.6% (9.06 … 9.54 μs) 8.17 μs ± 1.63% (7.87 … 8.34 μs) -13.2% ✅
Tropical GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 55.4 μs ± 0.245% (55.1 … 55.8 μs) 63.1 μs ± 0.26% (62.7 … 63.4 μs) +13.9% ❌
FPU GEMM Float32*Float32=Float64 (128×128) · (128×128) (NN) OP (8, 16, 2), base shape (4, 8, 1) 180.0 μs ± 0.201% (178.0 … 182.0 μs) 174.0 μs ± 5.57% (157.0 … 181.0 μs) -11.8% ✅

Comparison with baseline

test GemmKernels Baseline %
FPU GEMM Float32*Float32=Float32 (256×256) · (256×256) (TT) OP (8, 16, 2), base shape (4, 8, 1) 15.0 μs ± 1.13% (14.8 … 15.3 μs) 7.21 μs ± 1.67% (6.91 … 7.63 μs) 46.8
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (4, 32, 1), base shape (1, 32, 1) 16.1 μs ± 1.03% (15.7 … 16.2 μs) 4.18 μs ± 3.7% (4.05 … 4.53 μs) 25.8
WMMA GEMM Float16*Float16+Float16=Float16 (2048×2048) · (2048×2048) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 230.0 μs ± 0.675% (225.0 … 234.0 μs) 189.0 μs ± 2.47% (180.0 … 195.0 μs) 80.1
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 7.95 μs ± 1.61% (7.63 … 8.34 μs) 2.34 μs ± 6.13% (2.15 … 2.62 μs) 28.1
WMMA GEMM Float16*Float16+Float16=Float16 (2048×2048) · (2048×2048) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 223.0 μs ± 1.68% (205.0 … 229.0 μs) 186.0 μs ± 2.48% (171.0 … 192.0 μs) 83.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (16, 16, 1), base shape (8, 4, 1) 14.5 μs ± 1.07% (14.3 … 14.8 μs) 4.02 μs ± 3.4% (3.81 … 4.29 μs) 26.7
FPU GEMM Float32*Float32=Float32 (2048×2048) · (2048×2048) (NN) OP (8, 16, 2), base shape (4, 8, 1) 1.84 ms ± 2.57% (1.82 … 2.02 ms) 996.0 μs ± 1.3% (936.0 … 1020.0 μs) 51.5
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (32, 4, 1), base shape (32, 1, 1) 13.1 μs ± 0.947% (12.6 … 13.4 μs) 4.04 μs ± 3.56% (3.81 … 4.29 μs) 30.2
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (32, 4, 1), base shape (16, 2, 1) 19.1 μs ± 0.917% (18.8 … 19.3 μs) 4.08 μs ± 2.74% (3.81 … 4.29 μs) 20.3
FPU GEMM Float16*Float16=Float32 (2048×2048) · (2048×2048) (TT) OP (8, 16, 2), base shape (4, 8, 1) 1.83 ms ± 2.67% (1.8 … 2.02 ms) 291.0 μs ± 0.206% (289.0 … 293.0 μs) 16.1
FPU GEMM Float64*Float64=Float64 (2048×2048) · (2048×2048) (TN) OP (8, 16, 2), base shape (4, 8, 1) 46.2 ms ± 0.728% (46.0 … 52.4 ms) 41.0 ms ± 0.00317% (41.0 … 41.0 ms) 89.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (32, 4, 1), base shape (16, 2, 1) 19.1 μs ± 0.887% (18.8 … 19.3 μs) 4.28 μs ± 3.46% (4.05 … 4.53 μs) 21.5
FPU GEMM Float64*Float64=Float64 (256×256) · (256×256) (TN) OP (8, 16, 2), base shape (4, 8, 1) 285.0 μs ± 3.8% (255.0 … 293.0 μs) 109.0 μs ± 0.161% (109.0 … 109.0 μs) 42.7
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 7.42 μs ± 1.88% (7.15 … 7.87 μs) 3.81 μs ± 3.08% (3.58 … 4.29 μs) 50.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (8, 16, 2), base shape (4, 8, 1) 15.3 μs ± 1.16% (15.0 … 15.7 μs) 4.07 μs ± 2.89% (3.81 … 4.29 μs) 25.4
WMMA GEMM Float16*Float16+Float16=Float16 (512×256) · (256×512) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.74 μs ± 1.89% (8.34 … 9.06 μs) 5.18 μs ± 3.06% (4.77 … 5.48 μs) 57.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (4, 8, 1), base shape (4, 8, 1) 15.2 μs ± 1.09% (15.0 … 15.5 μs) 4.3 μs ± 3.46% (4.05 … 4.53 μs) 27.0
FPU GEMM Float32*Float32=Float32 (2048×2048) · (2048×2048) (TT) OP (8, 16, 2), base shape (4, 8, 1) 1.87 ms ± 2.72% (1.84 … 2.05 ms) 1.1 ms ± 1.71% (1.03 … 1.14 ms) 56.2
FPU GEMM Float64*Float64=Float64 (256×256) · (256×256) (NN) OP (8, 16, 2), base shape (4, 8, 1) 281.0 μs ± 5.58% (256.0 … 292.0 μs) 109.0 μs ± 0.156% (109.0 … 109.0 μs) 42.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (4, 8, 1), base shape (4, 8, 1) 15.2 μs ± 1.05% (15.0 … 15.5 μs) 4.03 μs ± 3.53% (3.81 … 4.53 μs) 25.4
WMMA GEMM Float16*Float16+Float32=Float32 (2048×2048) · (2048×2048) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 362.0 μs ± 4.97% (343.0 … 390.0 μs) 292.0 μs ± 0.323% (290.0 … 294.0 μs) 84.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (16, 16, 1), base shape (4, 8, 1) 15.3 μs ± 0.904% (15.0 … 15.7 μs) 4.08 μs ± 2.61% (3.81 … 4.29 μs) 25.4
FPU GEMM Float32*Float32=Float32 (2048×2048) · (2048×2048) (TN) OP (8, 16, 2), base shape (4, 8, 1) 1.45 ms ± 2.21% (1.43 … 1.53 ms) 1.11 ms ± 1.79% (1.04 … 1.15 ms) 72.6
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.11 μs ± 2.02% (7.87 … 8.34 μs) 2.23 μs ± 7.3% (1.91 … 2.38 μs) 24.2
FPU GEMM Float16*Float16=Float32 (2048×2048) · (2048×2048) (NN) OP (8, 16, 2), base shape (4, 8, 1) 1.86 ms ± 3.45% (1.81 … 2.03 ms) 292.0 μs ± 0.314% (290.0 … 294.0 μs) 16.1
FPU GEMM Float64*Float64=Float64 (128×128) · (128×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 131.0 μs ± 0.39% (129.0 … 132.0 μs) 28.8 μs ± 0.558% (28.6 … 29.1 μs) 22.2
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.0 μs ± 2.02% (7.87 … 8.34 μs) 2.46 μs ± 6.56% (2.15 … 2.62 μs) 27.3
FPU GEMM Float64*Float64=Float64 (2048×2048) · (2048×2048) (TT) OP (8, 16, 2), base shape (4, 8, 1) 46.3 ms ± 0.945% (46.0 … 52.4 ms) 41.0 ms ± 0.00328% (41.0 … 41.0 ms) 89.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (32, 4, 1), base shape (32, 1, 1) 12.6 μs ± 1.32% (12.4 … 12.9 μs) 4.19 μs ± 3.71% (4.05 … 4.53 μs) 32.7
WMMA GEMM Float16*Float16+Float32=Float32 (2048×2048) · (2048×2048) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 381.0 μs ± 4.58% (350.0 … 397.0 μs) 291.0 μs ± 0.335% (289.0 … 293.0 μs) 82.6
FPU GEMM Float16*Float16=Float32 (256×256) · (256×256) (TN) OP (8, 16, 2), base shape (4, 8, 1) 13.0 μs ± 1.07% (12.6 … 13.4 μs) 3.8 μs ± 3.19% (3.58 … 4.29 μs) 28.3
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.43 μs ± 1.87% (8.11 … 8.82 μs) 3.7 μs ± 3.87% (3.58 … 4.05 μs) 44.1
WMMA GEMM Float16*Float16+Float32=Float32 (2048×2048) · (2048×2048) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 365.0 μs ± 5.11% (341.0 … 389.0 μs) 292.0 μs ± 0.303% (290.0 … 293.0 μs) 84.8
FPU GEMM Float16*Float16=Float32 (2048×2048) · (2048×2048) (NT) OP (8, 16, 2), base shape (4, 8, 1) 2.3 ms ± 3.32% (2.26 … 2.56 ms) 291.0 μs ± 0.313% (289.0 … 293.0 μs) 12.8
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 6.88 μs ± 1.3% (6.44 … 7.15 μs) 2.34 μs ± 6.77% (1.91 … 2.62 μs) 29.6
WMMA GEMM Float16*Float16+Float32=Float32 (256×128) · (128×256) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.05 μs ± 1.88% (8.82 … 9.3 μs) 2.88 μs ± 4.9% (2.62 … 3.1 μs) 29.7
WMMA GEMM Float16*Float16+Float32=Float32 (128×256) · (256×128) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.3 μs ± 1.17% (12.2 … 12.6 μs) 2.92 μs ± 5.27% (2.62 … 3.1 μs) 21.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (4, 8, 2), base shape (4, 8, 1) 16.0 μs ± 1.03% (15.7 … 16.5 μs) 4.17 μs ± 3.75% (4.05 … 4.53 μs) 25.8
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (8, 16, 2), base shape (4, 8, 1) 16.4 μs ± 1.01% (16.0 … 16.7 μs) 4.18 μs ± 3.69% (4.05 … 4.53 μs) 25.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (32, 4, 1), base shape (32, 1, 1) 31.6 μs ± 0.465% (31.5 … 31.9 μs) 4.07 μs ± 2.83% (3.81 … 4.29 μs) 12.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 13.5 μs ± 0.95% (13.1 … 13.8 μs) 4.03 μs ± 3.53% (3.81 … 4.29 μs) 29.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 32, 1), base shape (1, 32, 1) 15.7 μs ± 1.1% (15.5 … 16.0 μs) 4.08 μs ± 2.52% (3.81 … 4.29 μs) 24.6
FPU GEMM Float16*Float16=Float32 (256×256) · (256×256) (NT) OP (8, 16, 2), base shape (4, 8, 1) 18.1 μs ± 0.867% (17.9 … 18.4 μs) 3.9 μs ± 3.6% (3.58 … 4.29 μs) 20.0
WMMA GEMM Float16*Float16+Float16=Float16 (2048×2048) · (2048×2048) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 221.0 μs ± 3.1% (206.0 … 229.0 μs) 189.0 μs ± 1.94% (178.0 … 195.0 μs) 86.4
WMMA GEMM Float16*Float16+Float16=Float16 (2048×2048) · (2048×2048) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 231.0 μs ± 1.07% (224.0 … 238.0 μs) 185.0 μs ± 2.2% (175.0 … 190.0 μs) 78.3
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (4, 32, 1), base shape (2, 16, 1) 21.9 μs ± 0.761% (21.7 … 22.2 μs) 4.29 μs ± 3.37% (4.05 … 4.53 μs) 18.7
WMMA GEMM Float16*Float16+Float32=Float32 (128×256) · (256×128) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.1 μs ± 1.43% (11.9 … 12.4 μs) 3.1 μs ± 3.64% (2.86 … 3.34 μs) 24.0
FPU GEMM Float64*Float64=Float64 (128×128) · (128×128) (NT) OP (8, 16, 2), base shape (4, 8, 1) 132.0 μs ± 1.16% (129.0 … 134.0 μs) 30.1 μs ± 0.529% (29.8 … 30.3 μs) 23.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (16, 16, 1), base shape (8, 4, 1) 14.5 μs ± 1.06% (14.3 … 14.8 μs) 4.08 μs ± 2.8% (3.81 … 4.29 μs) 26.7
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (4, 8, 2), base shape (4, 8, 1) 12.4 μs ± 1.4% (12.2 … 12.6 μs) 4.03 μs ± 3.56% (3.81 … 4.53 μs) 31.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (16, 16, 1), base shape (4, 8, 1) 15.4 μs ± 1.08% (15.3 … 15.7 μs) 4.03 μs ± 3.45% (3.81 … 4.29 μs) 25.0
WMMA GEMM Float16*Float16+Float16=Float16 (256×512) · (512×256) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.0 μs ± 1.37% (11.7 … 12.4 μs) 5.21 μs ± 3.25% (5.01 … 5.72 μs) 42.9
WMMA GEMM Float16*Float16+Float16=Float16 (512×512) · (512×512) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.3 μs ± 1.47% (11.9 … 12.6 μs) 7.23 μs ± 1.81% (6.91 … 7.63 μs) 58.0
WMMA GEMM Float16*Float16+Float16=Float16 (512×256) · (256×512) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.68 μs ± 2.17% (8.34 … 9.06 μs) 4.96 μs ± 3.41% (4.77 … 5.25 μs) 57.1
FPU GEMM Float64*Float64=Float64 (128×128) · (128×128) (NN) OP (8, 16, 2), base shape (4, 8, 1) 148.0 μs ± 0.156% (147.0 … 149.0 μs) 28.7 μs ± 0.481% (28.6 … 29.1 μs) 19.4
WMMA GEMM Float16*Float16+Float32=Float32 (256×256) · (256×256) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.4 μs ± 1.36% (12.2 … 12.6 μs) 3.81 μs ± 3.13% (3.58 … 4.29 μs) 29.4
WMMA GEMM Float16*Float16+Float32=Float32 (256×256) · (256×256) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.2 μs ± 1.31% (11.9 … 12.4 μs) 3.94 μs ± 3.94% (3.58 … 4.53 μs) 30.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (4, 32, 1), base shape (2, 16, 1) 21.9 μs ± 0.75% (21.7 … 22.2 μs) 4.03 μs ± 3.56% (3.81 … 4.29 μs) 17.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (32, 4, 1), base shape (16, 2, 1) 19.1 μs ± 0.879% (18.8 … 19.3 μs) 4.02 μs ± 3.53% (3.81 … 4.29 μs) 20.3
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (4, 8, 2), base shape (4, 8, 1) 11.9 μs ± 1.37% (11.7 … 12.2 μs) 4.29 μs ± 3.42% (4.05 … 4.53 μs) 34.7
WMMA GEMM Float16*Float16+Float16=Float16 (256×512) · (512×256) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.2 μs ± 1.43% (11.9 … 12.6 μs) 5.18 μs ± 3.06% (5.01 … 5.48 μs) 42.0
WMMA GEMM Float16*Float16+Float16=Float16 (512×256) · (256×512) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.69 μs ± 2.08% (8.34 … 9.06 μs) 5.1 μs ± 3.06% (4.77 … 5.48 μs) 57.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (4, 16, 1), base shape (4, 8, 1) 15.2 μs ± 1.09% (15.0 … 15.5 μs) 4.29 μs ± 3.35% (4.05 … 4.53 μs) 27.0
WMMA GEMM Float16*Float16+Float32=Float32 (256×256) · (256×256) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.2 μs ± 1.28% (11.9 … 12.4 μs) 3.96 μs ± 3.89% (3.81 … 4.29 μs) 32.0
FPU GEMM Float64*Float64=Float64 (2048×2048) · (2048×2048) (NT) OP (8, 16, 2), base shape (4, 8, 1) 46.6 ms ± 0.869% (46.4 … 52.8 ms) 41.0 ms ± 0.00258% (41.0 … 41.0 ms) 88.5
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (8, 8, 1), base shape (4, 8, 1) 15.4 μs ± 0.836% (15.0 … 15.7 μs) 4.09 μs ± 2.68% (3.81 … 4.29 μs) 25.4
WMMA GEMM Float16*Float16+Float16=Float16 (512×256) · (256×512) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.77 μs ± 2.05% (8.34 … 9.06 μs) 4.93 μs ± 3.22% (4.77 … 5.25 μs) 57.1
WMMA GEMM Float16*Float16+Float16=Float16 (512×512) · (512×512) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.4 μs ± 1.41% (12.2 … 12.6 μs) 7.38 μs ± 1.74% (7.15 … 7.87 μs) 58.8
WMMA GEMM Float16*Float16+Float32=Float32 (2048×2048) · (2048×2048) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 367.0 μs ± 4.8% (349.0 … 395.0 μs) 293.0 μs ± 0.295% (291.0 … 295.0 μs) 83.2
WMMA GEMM Float16*Float16+Float16=Float16 (512×512) · (512×512) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.3 μs ± 1.43% (11.9 … 12.6 μs) 7.29 μs ± 2.01% (7.15 … 7.63 μs) 60.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 8, 1), base shape (4, 8, 1) 13.5 μs ± 1.07% (13.1 … 13.8 μs) 4.08 μs ± 2.66% (3.81 … 4.29 μs) 29.1
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (NT) OP (8, 16, 2), base shape (4, 8, 1) 9.2 μs ± 1.51% (8.82 … 9.54 μs) 2.04 μs ± 5.96% (1.67 … 2.38 μs) 18.9
WMMA GEMM Float16*Float16+Float32=Float32 (256×128) · (128×256) (NN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.29 μs ± 1.81% (9.06 … 9.54 μs) 2.94 μs ± 5.31% (2.62 … 3.1 μs) 28.9
WMMA GEMM Float16*Float16+Float32=Float32 (128×128) · (128×128) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 8.05 μs ± 2.06% (7.87 … 8.34 μs) 2.04 μs ± 6.49% (1.67 … 2.38 μs) 21.2
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (16, 16, 1), base shape (8, 4, 1) 14.6 μs ± 0.905% (14.3 … 14.8 μs) 4.19 μs ± 3.76% (4.05 … 4.53 μs) 28.3
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (4, 32, 1), base shape (1, 32, 1) 33.0 μs ± 0.414% (32.7 … 33.4 μs) 4.03 μs ± 3.31% (3.81 … 4.29 μs) 11.7
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (4, 16, 1), base shape (4, 8, 1) 15.2 μs ± 1.13% (15.0 … 15.5 μs) 4.03 μs ± 3.55% (3.81 … 4.29 μs) 25.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (32, 4, 1), base shape (16, 2, 1) 19.2 μs ± 0.801% (18.8 … 19.3 μs) 4.18 μs ± 3.66% (4.05 … 4.53 μs) 21.5
FPU GEMM Float32*Float32=Float32 (2048×2048) · (2048×2048) (NT) OP (8, 16, 2), base shape (4, 8, 1) 2.36 ms ± 3.46% (2.32 … 2.63 ms) 865.0 μs ± 1.13% (816.0 … 876.0 μs) 35.1
FPU GEMM Float64*Float64=Float64 (256×256) · (256×256) (TT) OP (8, 16, 2), base shape (4, 8, 1) 282.0 μs ± 5.73% (257.0 … 295.0 μs) 109.0 μs ± 0.161% (109.0 … 110.0 μs) 42.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TT) OP (8, 8, 1), base shape (4, 8, 1) 15.3 μs ± 1.01% (15.0 … 15.5 μs) 4.03 μs ± 3.5% (3.81 … 4.53 μs) 25.4
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 7.47 μs ± 1.84% (7.15 … 7.87 μs) 3.63 μs ± 2.96% (3.58 … 4.05 μs) 50.0
WMMA GEMM Float16*Float16+Float32=Float32 (128×256) · (256×128) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.2 μs ± 1.46% (11.9 … 12.4 μs) 2.98 μs ± 4.97% (2.62 … 3.34 μs) 22.0
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (TT) OP (8, 16, 2), base shape (4, 8, 1) 8.17 μs ± 1.63% (7.87 … 8.34 μs) 2.17 μs ± 7.94% (1.91 … 2.38 μs) 24.2
WMMA GEMM Float16*Float16+Float16=Float16 (512×512) · (512×512) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.4 μs ± 1.44% (11.9 … 12.9 μs) 7.28 μs ± 2.02% (7.15 … 7.63 μs) 60.0
WMMA GEMM Float16*Float16+Float32=Float32 (256×256) · (256×256) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.4 μs ± 1.39% (12.2 … 12.6 μs) 3.91 μs ± 3.77% (3.58 … 4.29 μs) 29.4
FPU GEMM Float32*Float32=Float32 (256×256) · (256×256) (NN) OP (8, 16, 2), base shape (4, 8, 1) 15.3 μs ± 1.12% (15.0 … 15.5 μs) 6.77 μs ± 2.01% (6.44 … 6.91 μs) 42.9
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (8, 8, 1), base shape (4, 8, 1) 15.4 μs ± 1.07% (15.3 … 15.7 μs) 4.18 μs ± 3.71% (4.05 … 4.53 μs) 26.6
FPU GEMM Float64*Float64=Float64 (256×256) · (256×256) (NT) OP (8, 16, 2), base shape (4, 8, 1) 269.0 μs ± 6.22% (256.0 … 293.0 μs) 109.0 μs ± 0.158% (109.0 … 109.0 μs) 42.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (4, 16, 1), base shape (4, 8, 1) 15.3 μs ± 0.964% (15.0 … 15.7 μs) 4.19 μs ± 3.76% (4.05 … 4.53 μs) 27.0
WMMA GEMM Float16*Float16+Float32=Float32 (256×128) · (128×256) (TT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.2 μs ± 1.6% (8.82 … 9.54 μs) 2.91 μs ± 5.13% (2.62 … 3.1 μs) 29.7
FPU GEMM Float16*Float16=Float32 (128×128) · (128×128) (NN) OP (8, 16, 2), base shape (4, 8, 1) 9.22 μs ± 1.31% (8.82 … 9.54 μs) 2.22 μs ± 7.27% (1.91 … 2.38 μs) 21.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (8, 8, 1), base shape (4, 8, 1) 15.2 μs ± 1.14% (15.0 … 15.5 μs) 4.28 μs ± 3.59% (4.05 … 4.77 μs) 27.0
FPU GEMM Float16*Float16=Float32 (256×256) · (256×256) (TT) OP (8, 16, 2), base shape (4, 8, 1) 15.8 μs ± 1.08% (15.5 … 16.0 μs) 3.97 μs ± 4.07% (3.58 … 4.29 μs) 23.1
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (16, 16, 1), base shape (4, 8, 1) 15.3 μs ± 0.926% (15.0 … 15.7 μs) 4.17 μs ± 3.62% (4.05 … 4.53 μs) 27.0
WMMA GEMM Float16*Float16+Float32=Float32 (128×256) · (256×128) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.4 μs ± 1.34% (12.2 … 12.6 μs) 2.74 μs ± 5.58% (2.38 … 2.86 μs) 19.6
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 8, 2), base shape (4, 8, 1) 14.9 μs ± 1.11% (14.5 … 15.3 μs) 4.09 μs ± 2.79% (3.81 … 4.29 μs) 26.2
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (32, 4, 1), base shape (32, 1, 1) 31.6 μs ± 0.455% (31.5 … 31.9 μs) 4.29 μs ± 3.36% (4.05 … 4.53 μs) 12.9
WMMA GEMM Float16*Float16+Float32=Float32 (256×128) · (128×256) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 9.17 μs ± 1.66% (8.82 … 9.3 μs) 2.91 μs ± 5.19% (2.62 … 3.1 μs) 29.7
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (4, 32, 1), base shape (2, 16, 1) 21.9 μs ± 0.76% (21.7 … 22.2 μs) 4.19 μs ± 3.66% (4.05 … 4.53 μs) 18.7
FPU GEMM Float16*Float16=Float32 (256×256) · (256×256) (NN) OP (8, 16, 2), base shape (4, 8, 1) 15.7 μs ± 1.1% (15.5 … 16.0 μs) 3.95 μs ± 3.89% (3.81 … 4.29 μs) 24.6
FPU GEMM Float16*Float16=Float32 (2048×2048) · (2048×2048) (TN) OP (8, 16, 2), base shape (4, 8, 1) 1.54 ms ± 2.56% (1.52 … 1.67 ms) 292.0 μs ± 0.172% (290.0 … 295.0 μs) 19.1
FPU GEMM Float64*Float64=Float64 (2048×2048) · (2048×2048) (NN) OP (8, 16, 2), base shape (4, 8, 1) 46.5 ms ± 0.73% (46.3 … 52.7 ms) 41.0 ms ± 0.00326% (41.0 … 41.0 ms) 88.6
WMMA GEMM Float16*Float16+Float16=Float16 (256×512) · (512×256) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.1 μs ± 1.36% (11.9 … 12.4 μs) 5.09 μs ± 3.48% (4.77 … 5.72 μs) 40.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 32, 1), base shape (2, 16, 1) 22.0 μs ± 0.639% (21.7 … 22.4 μs) 4.08 μs ± 2.92% (3.81 … 4.29 μs) 17.6
WMMA GEMM Float16*Float16+Float16=Float16 (256×512) · (512×256) (NT) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 12.0 μs ± 1.38% (11.7 … 12.4 μs) 4.48 μs ± 3.63% (4.29 … 4.77 μs) 36.7
FPU GEMM Float32*Float32=Float32 (256×256) · (256×256) (NT) OP (8, 16, 2), base shape (4, 8, 1) 18.2 μs ± 0.77% (17.9 … 18.6 μs) 6.76 μs ± 2.08% (6.44 … 6.91 μs) 36.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NN) OP (4, 16, 1), base shape (4, 8, 1) 15.2 μs ± 1.13% (15.0 … 15.5 μs) 4.08 μs ± 2.76% (3.81 … 4.29 μs) 25.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (NT) OP (4, 8, 1), base shape (4, 8, 1) 15.3 μs ± 1.09% (15.0 … 15.5 μs) 4.18 μs ± 3.66% (4.05 … 4.53 μs) 27.0
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (4, 32, 1), base shape (1, 32, 1) 33.0 μs ± 0.474% (32.7 … 33.4 μs) 4.29 μs ± 3.53% (4.05 … 4.53 μs) 12.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 12.1 μs ± 1.2% (11.7 … 12.4 μs) 4.3 μs ± 3.47% (4.05 … 4.53 μs) 34.7
FPU GEMM Float64*Float64=Float64 (128×128) · (128×128) (TN) OP (8, 16, 2), base shape (4, 8, 1) 140.0 μs ± 6.34% (129.0 … 152.0 μs) 30.7 μs ± 0.381% (30.3 … 31.0 μs) 23.4
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (16, 16, 1), base shape (4, 8, 1) 15.2 μs ± 1.13% (15.0 … 15.5 μs) 4.29 μs ± 3.5% (4.05 … 4.53 μs) 27.0
WMMA GEMM Float16*Float16+Float16=Float16 (256×256) · (256×256) (TN) Block (128, 128, 64) Warps (4, 2) OP (16, 16, 16) 7.76 μs ± 6.46% (7.15 … 9.06 μs) 3.59 μs ± 2.9% (3.1 … 4.05 μs) 43.3
FPU GEMM Float32*Float32=Float32 (128×128) · (128×128) (TN) OP (16, 16, 1), base shape (8, 4, 1) 14.5 μs ± 1.06% (14.3 … 14.8 μs) 4.3 μs ± 3.45% (4.05 … 4.53 μs) 28.3
FPU GEMM Float32*Float32=Float32 (256×256) · (256×256) (TN) OP (8, 16, 2), base shape (4, 8, 1) 12.6 μs ± 1.31% (12.4 … 12.9 μs) 7.16 μs ± 1.19% (6.68 … 7.63 μs) 53.8

Copy link

codecov bot commented Nov 15, 2023

Codecov Report

Attention: 4 lines in your changes are missing coverage. Please review.

Comparison is base (912fa05) 34.16% compared to head (8caa172) 34.16%.

Files Patch % Lines
src/layout.jl 0.00% 4 Missing ⚠️
Additional details and impacted files
@@           Coverage Diff           @@
##           master     #175   +/-   ##
=======================================
  Coverage   34.16%   34.16%           
=======================================
  Files          11       11           
  Lines         922      922           
=======================================
  Hits          315      315           
  Misses        607      607           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@thomasfaingnaert thomasfaingnaert merged commit 8b52f30 into master Nov 15, 2023
1 check passed
@thomasfaingnaert thomasfaingnaert deleted the tf/fix-alignment-check branch November 15, 2023 13:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants