opengl:gpuflops
文書の過去の版を表示しています。
GPU FLOPS
GPU FLOPS
| GPU | Clock | highp FLOPS | mediump FLOPS | core | fop/clock | op |
|---|---|---|---|---|---|---|
| RADEON R9 Fury X (GCN1.3) | 1050MHz | 8601.6 GFLOPS | 4096sp | 8192 | 2(mad) x 4096(core) x 1.05(clock) = 8601.6 GFLOPS | |
| RADEON R9 290X (GCN1.1) | 1000MHz | 5632.0 GFLOPS | 2816sp | 5632 | 2(mad) x 2816(core) x 1.0(clock) = 5632 GFLOPS | |
| GeFOrce GTX 980 Ti (Maxwell) | 1000MHz | 5632.0 GFLOPS | 2816sp | 5632 | 2(mad) x 2816(core) x 1.0(clock) = 5632 GFLOPS | |
| GeForce GTX 780 Ti (Kepler) | 875MHz | 5040.0 GFLOPS | 2880sp | 5760 | 2(mad) x 2880(core) x 0.875(clock) = 5040 GFLOPS | |
| GeForce GTX 780 (Kepler) | 863MHz | 3976.7 GFLOPS | 2304sp | 4608 | 2(mad) x 2304(core) x 0.863(clock) = 3976.7 GFLOPS | |
| RADEON R9 280X (GCN1.0) | 850MHz | 3481.6 GFLOPS | 2048sp | 4096 | 2(mad) x 2048(core) x 0.85(clock) = 3481.6 GFLOPS | |
| RADEON R9 285 (GCN1.2) | 918MHz | 3290.0 GFLOPS | 1792sp | 3584 | 2(mad) x 1792(core) x 0.918(clock) = 3290 GFLOPS | |
| PS4 (GCN) | 800MHz | 1843.2 GFLOPS | 1152sp | 2304 | 2(mad) x 1152(core) x 0.8(clock) = 1843.2 GFLOPS | |
| Xbox One (GCN) | 853MHz | 1310.2 GFLOPS | 768sp | 1536 | 2(mad) x 768(core) x 0.853(clock) = 1310.2 GFLOPS | |
| Intel Iris Pro Graphics 6200 (8th) | 1150MHz | 883.2 GFLOPS | 48eu | 768 | 2(mad) x 4(vec4) x 2(op) x 48(eu) x (clock) | |
| Intel Iris Pro Graphics 5200 (7.5) | 200-1300 | 182-832 GFLOPS | 40eu | 640 | 2(mad) x 4(vec4) x 2(op) x 40(eu) x (clock) | |
| Radeon R7 (Kaveri A10-7850K) | 720MHz | 737.3 GFLOPS | 512sp | 1024 | 2(mad) x 512(core) x 0.72(clock) = 737.28 GFLOPS | |
| Tegra X1 (Maxwell) | 1000MHz | 512.0 GFLOPS | 1024 GFLOPS | 256sp | 512 | 2(mad) x 256(core) x 1.0(clock) = 512 GFLOPS |
| Intel HD Graphics 4600 (7.5th) | 350-1200 | 112-384 GFLOPS | 20eu | 320 | 2(mad) x 4(vec4) x 2(op) x 20(eu) x (clock) | |
| Tegra K1 (Kepler) | 950MHz | 364.8 GFLOPS | 192sp | 384 | 2(mad) x 192(core) x 0.95(clock) = 384.8 GFLOPS | |
| Intel HD Graphics 4000 (7th) | 650-1150 | 166.4-294.4 | 16eu | 256 | 2(mad) x 4(vec4) x 2(op) x 16(eu) x (clock) | |
| Xbox 360 Xenos (RADEON) | 500MHz | 240.0 GFLOPS | 240sp | 480 | 2(mad) x 5(vec5) x 48(unit) x 0.5(clock) = 240 GFLOPS | |
| PowerVR GX6850? A8X (RogueXT) | 450MHz | 230.4 GFLOPS | 512 | 2(mad) x 32(simd) x 8(core) x 0.45(clock) = 230.4 GFLOPS | ||
| PS3 RSX (G70) | 550MHz | 192.0 GFLOPS | 350 | ( 2(mad) x 5(vec5) x 8(unit) + 2(mad) x 5(vec5) x 24(unit) ) x 0.55(clock) = ? | ||
| RADEON R3 GCN1.1 (Kabini Athlon 5350) | 600MHz | 153.6 GFLOPS | 128 | 256 | 2(mad) x 128(core) x 0.6(clock) = 153.6 GFLOPS | |
| APQ8084 Adreno 420 | 500MHz | 144.0 GFLOPS | 128 | 288 | ( 2(mad) x 4(simd) x 32(unit) + 32(scalar) ) x 0.5(clock) = 144 GFLOPS | |
| MSN8974 Adreno 330 (128) | 450MHz | 129.6 GFLOPS | 128 | 288 | ( 2(mad) x 4(simd) x 32(unit) + 32(scalar) ) x 0.45(clock) = 129.6 GFLOPS | |
| Intel HD Graphics 3000 (6th) | 850-1350 | 81.6-129.6 | 12eu | 96 | 2(mad) x 4(simd) x 12(eu) x (clock) | |
| PowerVR G6430 (Rogue) | 115.2 GFLOPS | 256 | 2(mad) x 32(simd) x 4(core) x (clock) | |||
| APQ8064T Adreno 320 (96) | 400MHz | 86.4 GFLOPS | 96 | 216 | ( 2(mad) x 4(simd) x 24(unit) + 24(scalar) ) x 0.4(clock) = 86.4 GFLOPS | |
| Exynos 5 Dual Mali-T604 | 533MHz | 72.5 GFLOPS | 4core | 136 | ( 2(mad) x 5(simd) + 7(dp) ) x 2(alu) x 4(core) x 0.533(clock) = 72.5 GFLOPS | |
| PowerVR SGX554MP4 A6X | 266MHz | 69.0 GFLOPS | 137.9 GFLOPS? | 4core | 256 | 2(mad) x 4(simd) x 8(unit) x 4(core) x 0.266(clock) = 68.96 GFLOPS |
| APQ8064 Adreno 320 (64) | 400MHz | 57.6 GFLOPS | 64 | 144 | ( 2(mad) x 4(simd) x 16(unit) + 16(scalar) ) x 0.4(clock) = 57.6 GFLOPS | |
| Celeron J1900 Intel HG Graphics (7th) | 688-854 | 44.0-54.7 | 4eu | 64 | 2(mad) x 4(vec4) x 2(op) x 4(eu) x (clock) | |
| Atom Z3740 Intel HG Graphics (7th) | 311-667 | 19.9-42.7 | 4eu | 64 | 2(mad) x 4(vec4) x 2(op) x 4(eu) x (clock) | |
| PowerVR SGX543MP4 A5X | 250MHz | 32.0 GFLOPS | 64.0 GFLOPS? | 4core | 128 | 2(mad) x 4(simd) x 4(unit) x 4(core) x 0.250(clock) = 32 GFLOPS |
| Tegra 4 ULP GeForce(72) | 672MHz | 32.3 GFLOPS | 96.8 GFLOPS | 72sp | 144 (48) | 2(mad) x 72(core) x 0.672(clock) = 96.77 GFLOPS |
| K3V2 Vivante GC4000 | 480MHz | 30.7 GFLOPS | 32 | 64 | 2(mad) x 32(core) x 0.48(clock) = 30.72 GFLOPS | |
| Atom Z2760 PowerVR SGX545 | 533MHz | 8.5 GFLOPS | 17.0 GFLOPS? | 1core | 16 | |
| VideoCore IV | 250MHz | 8.0 GFLOPS | 32 | 2(mad) x 4(simd) x 4(qpu) x 0.25(clock) = 8.0 GFLOPS | ||
| Tegra 3 ULP GeForce(12) | 416MHz | 3.3 GFLOPS | 10.0 GFLOPS | 12sp | 24 (8) | 2(mad) x 12(core) x 0.416(clock) = 9.98 GFLOPS |
| Tegra 2 ULP GeForce(8) | 333MHz | 2.7 GFLOPS | 5.3 GFLOPS | 8sp | 16 (8) | 2(mad) x 8(core) x 0.333(clock) = 5.33 GFLOPS |
| RK3188 Mali-400MP4 | 533MHz | 2.1 GFLOPS | 19.2 GFLOPS | 4+1 | 36 (4) | ( 2(mad) x 4(simd) x 4(core) + 2(mad) x 2(simd) x 1(core) ) x 0.533(clock) = 19.2 GFLOPS |
| RK3066 Mali-400MP4 | 250MHz | 1.0 GFLOPS | 9.0 GFLOPS | 4+1 | 36 (4) | ( 2(mad) x 4(simd) x 4(core) + 2(mad) x 2(simd) x 1(core) ) x 0.25(clock) = 9 GFLOPS |
- 原則として Shader Unit のみ
参照page
GPU spec
| GPU | ARCH | SP | Clock | GFLOPS | TMU | ROP | Mem | Fill | Tex | B/W | API Windows | OSX | Linux | SM | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RADEON HD 5850 | 40nm | Cypress VLIW5 | 1440 | 725MHz | 2088.0G | 72 | 32 | GDDR5 | 256bit | 4000MHz | 1GB | 23.2Gpix/s | 52.2GT/s | 128.0GB/s | D3D11.0 / GL4.4 | – | GL4.4 | 5.0 |
| RADEON HD 6750M | 40nm | Whistler VLIW5 | 480 | 600MHz | 576.0G | 24 | 8 | GDDR5 | 128bit | 3176MHz | 1GB | 4.8Gpix/s | 14.4GT/s | 50.8GB/s | D3D11.0 / GL4.4 | GL4.1 | GL4.3 | 5.0 |
| RADEON HD 7750 | 28nm | GCN 1.0 | 512 | 800MHz | 819.2G | 40 | 16 | GDDR5 | 128bit | 4500MHz | 1GB | 72.0GB/s | D3D11.1 / GL4.5 | – | GL4.4 | 5.0 | ||
| RADEON R3 (HD8400) (Athlon5350) | 28nm | Kabini GCN 1.1 + HSA | 128 | 600MHz | 153.6G | 8 | 4 | DDR3 | 128bit | 1600MHz | – | 2.0Gpix/s | 5.0GT/s | 12.8GB/s | D3D12.0 / GL4.5 | – | GL4.4 | 5.0 |
| RADEON R9 285 | 28nm | Tonga PRO GCN 1.2 | 1792 | 918MHz | 3290.1G | 112 | 32 | GDDR5 | 256bit | 5500MHz | 29.4Gpix/s | 102.8GT/s | 176.0GB/s | D3D12.0 / GL4.5 | – | GL4.4 | 5.0 | |
| RADEON R9 290X | 28nm | Hawaii XT GCN 1.1 | 2816 | 1000MHz | 5632.0G | 176 | 64 | GDDR5 | 512bit | 5000MHz | 64.0Gpix/s | 176.0GT/s | 320.0GB/s | D3D12.0 / GL4.5 | – | GL4.4 | 5.0 | |
| RADEON Fury X | 28nm | Fiji GCN 1.3 | 4096 | 1050MHz | 8601.6G | 256 | 64 | HBM | 4096bit | 1000MHz | 4GB | 67.2Gpix/s | 268.8GT/s | 512.0GB/s | D3D12.0 / GL4.5 | – | 5.0 | |
| GeForce GT 240 | 40nm | GT215 | (96sp) 144 | 1340MHz | 385.9G | 32 | 8 | DDR3 | 128bit | 1580MHz | 1GB | 4.4Gpix/s | 17.6GT/s | 25.3GB/s | D3D10.1 / GL3.3 | – | GL3.3 | 4.1 |
| GeForce GTX 650 | 28nm | GK107 Kepler | 384 | 1059MHz | 813.3G | 32 | 16 | GDDR5 | 128bit | 5000MHz | 1GB | 16.9Gpix/s | 33.9GT/s | 80.0GB/s | D3D11.0 / GL4.5 | – | GL4.5 | 5.0 |
| GeForce GTX 750 Ti | 28nm | GM107 Maxwell | 640 | 1033MHz | 1322.2G | 40 | 32 | GDDR5 | 128bit | 5400MHz | 2GB | 33.1Gpix/s | 41.3GT/s | 86.4GB/s | D3D11.0 / GL4.5 | – | GL4.5 | 5.0 |
| Tegra K1 | 28nm | Kepler | 192 | 950MHz | 364.8G | 8 | 4 | LPDDR3 | 64bit | 1860MHz | – | 14.9GB/s | D3D11.0 / GL4.5 | – | GL4.5 | 5.0 | ||
| Tegra X1 | 20nm | Maxwell | 256 | 1000MHz | 512.0G | 16 | 16 | LPDDR4 | 64bit | 3200MHz | – | 25.6GB/s | D3D11.0 / GL4.5 | – | GL4.5 | 5.0 | ||
| Intel HD Graphcis 4600 | 22nm | GT2 | (20EU) 160 | 350-1250MHz | 400.0G | 8 | 4 | DDR3 | 128bit | 1600MHz | – | 25.6GB/s | D3D11.0 / GL4.3 | GL4.1 | GL3.3 | 5.0 | ||
| Intel HD Graphcis 4000 | 22nm | GT2 | (16EU) 128 | 650MHz | 166.4G | 8 | 4 | DDR3 | 128bit | 1600MHz | – | 25.6GB/s | D3D11.0 / GL4.0 | GL4.1 | GL3.3 | 5.0 | ||
| Intel HD Graphics (BayTrail Z3740) | 22nm | (4EU) 32 | 311-667MHz | 42.7G | 4 | 2 | LPDDR3 | 128bit | 1066MHz | – | 17.1GB/s | D3D11.0 / GL4.0 | – | GL3.3 | 5.0 | |||
| Intel HD Graphics (Celeron J1900) | 22nm | (4EU) 32 | 688-854MHz | 54.6G | 4 | 2 | DDR3L | 128bit | 1333MHz | – | 21.3GB/s | D3D11.0 / GL4.0 | – | GL3.3 | 5.0 | |||
- GPU-Z 使用, Intel HD Graphics は公式の spec 表より
opengl/gpuflops.1434888968.txt.gz · 最終更新: 2015/06/21 21:16 by oga
