Difference between revisions of "GPUs comparison"
From KlayGE
Gongminmin (Talk | contribs) |
Gongminmin (Talk | contribs) |
||
Line 1: | Line 1: | ||
{| class="wikitable sortable" | {| class="wikitable sortable" | ||
|- | |- | ||
− | ! | + | ! !! !! !! !! !! !! ! colspan="2" | Clock rate (MHz) !! ! colspan="4" | Memory !! ! colspan="2" | GFLOPS !! !! ! colspan="2" | GFLOPS/W !! |
|- | |- | ||
− | + | ! Vendor !! Type !! Model !! Fab (nm) !! Bus interface !! Core config !! Core !! Memory !! Size !! Bandwidth (GB/s) !! Bus (bit) !! Type !! Float !! Double !! TDP (watts) !! Float !! Double !! API | |
|- | |- | ||
− | | NVIDIA || Desktop || GeForce GTX | + | | NVIDIA || Desktop || GeForce GTX 680 || 28 || PCIe 3.0 x16 || 1536:128:32 || 1006-1110 || 6008 || 2G || 192 || 256 || GDDR5 || 3090 || 0 || 195 || 15.8 || 0 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || Desktop || GeForce GTX | + | | NVIDIA || Desktop || GeForce GTX 780 || 28 || PCIe 3.0 x16 || 2304:192:48 || 863-1002 || 6008 || 3G || 288 || 384 || GDDR5 || 3977 || 165.7 || 250 || 15.9 || 0.7 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || Desktop || GeForce GTX | + | | NVIDIA || Desktop || GeForce GTX Titan || 28 || PCIe 3.0 x16 || 2688:224:48 || 836-993 || 6008 || 6G || 288 || 384 || GDDR5 || 4500 || 1300-1500 || 250 || 18.0 || 6.0 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || | + | | NVIDIA || Desktop || GeForce GTX 780 Ti || 28 || PCIe 3.0 x16 || 2880:240:48 || 876-928 || 7000 || 3G || 336 || 384 || GDDR5 || 5046 || 210 || 250 || 20.2 || 0.8 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || | + | | NVIDIA || Professional || Quadro K6000 || 28 || PCIe 3.0 x16 || 2880:240:48 || 901.5 || 6008 || 12G || 288 || 384 || GDDR5 || 5196 || 1732 || 225 || 23.1 || 7.7 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || | + | | NVIDIA || Computing || Tesla K20X || 28 || PCIe 3.0 x16 || 2688 FP32, 896 FP64 || 732 || 5200 || 6G || 250 || 384 || GDDR5 || 3935 || 1312 || 235 || 16.7 || 5.6 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | NVIDIA || Mobile || Tegra | + | | NVIDIA || Mobile || Tegra 4 || 28 || Integrated || 72 || 672 || 1866 || N/A || 29.8 || 128 || LPDDR3 || 96.8 || 0 || 4 (SoC) || 24.2 || 0 || OpenGL ES 2.0 |
|- | |- | ||
− | | | + | | NVIDIA || Mobile || Tegra 5 || 28 || Integrated || 192 || 900 || ??? || N/A || ??? || ??? || LPDDR3 || 345.6 || 0 || ??? || ??? || 0 || D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
|- | |- | ||
− | | AMD || Desktop || Fusion APU | + | | AMD || Desktop || Fusion APU 8670D || 32 || Integrated || 384:24:8 || 844-950 || 1066 || N/A || 29.9 || 128 || DDR3 || 648.2 || 0 || 100 (SoC) || 6.5 || 0 || D3D 11.0, OpenGL 4.3, OpenCL 1.2 |
|- | |- | ||
− | | AMD || Desktop || | + | | AMD || Desktop || Fusion APU R7 2??D || 28 || Integrated || 512:32:16 || 720 || ??? || N/A || ??? || 128 || DDR3 || 856 || ??? || 95 (SoC) || ??? || ??? || D3D 11.2, OpenGL 4.3, OpenCL 1.2, Mantle |
|- | |- | ||
− | | AMD || | + | | AMD || Desktop || Radeon HD 8970 || 28 || PCIe 3.0 x16 || 2048:128:32 || 1000-1050 || 6000 || 3G || 288 || 384 || GDDR5 || 4300 || 1075 || 250 || 17.2 || 4.3 || D3D 11.1, OpenGL 4.3, OpenCL 1.2 |
|- | |- | ||
− | | AMD || | + | | AMD || Professional || Radeon W9000 || 28 || PCIe 3.0 x16 || 2048:128:32 || 975 || 5500 || 6G || 264 || 384 || GDDR5 || 3993.6 || 998.4 || 274 || 12.4 || 3.6 || D3D 11.1, OpenGL 4.2, OpenCL 1.2 |
|- | |- | ||
− | | | + | | AMD || Desktop || Radeon R9 290X || 28 || PCIe 3.0 x16 || 2816:176:64 || 800-1000 || 5000 || 4G || 320 || 512 || GDDR5 || 5632 || 704 || 290 || 19.4 || 2.4 || D3D 11.2, OpenGL 4.3, OpenCL 1.2, Mantle |
|- | |- | ||
− | | Intel || | + | | Intel || Desktop || Iris Pro Graphics 5200 || 22 || Integrated || 160:8:4 || 400-1300 || 1600 || N/A || 25.6 || 128 || DDR3 || 832 || 0 || 28 (SoC) || 29.7 || 0 || D3D 11.1, OpenGL 4.0, OpenCL 1.2 |
|- | |- | ||
− | | Intel || Computing || Xeon Phi | + | | Intel || Computing || Xeon Phi 3100P || 22 || PCIe 2.0 x16 || 228 x86 || 1100 || 3750 || 6G || 240 || 512 || GDDR5 || 2000 || 1000 || 300 || 6.7 || 3.3 || OpenMP, OpenCL, MKL |
|- | |- | ||
− | | Intel || Computing || Xeon Phi | + | | Intel || Computing || Xeon Phi 5110P || 22 || PCIe 2.0 x16 || 240 x86 || 1053 || 5000 || 8G || 320 || 512 || GDDR5 || 2022 || 1011 || 225 || 9.0 || 4.5 || OpenMP, OpenCL, MKL |
|- | |- | ||
− | | | + | | Intel || Computing || Xeon Phi 7120P || 22 || PCIe 2.0 x16 || 244 x86 || 1238-1333 || 5500 || 16G || 352 || 512 || GDDR5 || 2416 || 1208 || 300 || 8.1 || 4.0 || OpenMP, OpenCL, MKL |
|- | |- | ||
− | | Qualcomm || Mobile || Adreno | + | | Qualcomm || Mobile || Adreno 330 || 28 (Snapdragon 800) || Integrated || 128 || 450 || 800 || N/A || 12.8 || 128 || LPDDR3 || 129.6 || 0 || 3 (SoC) || 43.2 || 0 || D3D 9.3, OpenGL ES 3.0, OpenCL 1.2 |
|- | |- | ||
− | | | + | | Qualcomm || Mobile || Adreno 420 || 28 (Snapdragon 805) || Integrated || ??? || 500 || 800 || N/A || 25.6 || 256 || LPDDR3 || ??? || ??? || ??? || ??? || ??? || D3D 11.1, OpenGL ES 3.0, OpenCL 1.2 |
|- | |- | ||
− | | Imagination || Mobile || | + | | Imagination || Mobile || PowerVR SGX554 MP4 || 32 (A6X) || Integrated || 128 || 300 || 533 || N/A || 17 || 256 || LPDDR3 || 76.8 || 0 || 4 (SoC) || 19.2 || 0 || D3D 9.3, OpenGL 2.1, OpenGL ES 2.0, OpenCL 1.1 |
|- | |- | ||
− | | | + | | Imagination || Mobile || G6430 || 28 (A7) || Integrated || 256 || 450 || ??? || N/A || ??? || ??? || LPDDR3 || 115.2 || 0 || ??? || ??? || 0 || D3D 10.0, OpenGL 3.2, OpenGL ES 3.0 |
|- | |- | ||
− | | ARM || | + | | ARM || Moblie || Mali-T604 MP4 || 32 (Exynos 5 Dual) || Integrated || 64 || 533 || 800 || N/A || 12.8 || 128 || LPDDR3 || 68 || 0 || 4 (SoC) || 17.0 || 0 || D3D 9.1, OpenGL ES 3.0, OpenCL 1.1 |
|- | |- | ||
− | | Vivante || Mobile || GC4000 || 40 (K3V2) || Integrated | + | | ARM || Mobile || Mali-T760 MP16 || ??? || Integrated || ??? || 600 || ??? || N/A || ??? || ??? || LPDDR3 || 326.4 || 0 || ??? || ??? || 0 || D3D 11.1, OpenGL ES 3.0, OpenCL 1.1 |
+ | |- | ||
+ | | Vivante || Mobile || GC4000 || 40 (K3V2) || Integrated || 32 || 680 || ??? || N/A || ??? || ??? || LPDDR3 || 43.5 || 0 || 4 (SoC) || 10.9 || 0 || D3D 9.3, OpenGL ES 3.0, OpenGL 3.0, OpenCL 1.2 | ||
|} | |} |
Revision as of 03:25, 4 December 2013
Clock rate (MHz) | Memory | GFLOPS | GFLOPS/W | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Vendor | Type | Model | Fab (nm) | Bus interface | Core config | Core | Memory | Size | Bandwidth (GB/s) | Bus (bit) | Type | Float | Double | TDP (watts) | Float | Double | API |
NVIDIA | Desktop | GeForce GTX 680 | 28 | PCIe 3.0 x16 | 1536:128:32 | 1006-1110 | 6008 | 2G | 192 | 256 | GDDR5 | 3090 | 0 | 195 | 15.8 | 0 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Desktop | GeForce GTX 780 | 28 | PCIe 3.0 x16 | 2304:192:48 | 863-1002 | 6008 | 3G | 288 | 384 | GDDR5 | 3977 | 165.7 | 250 | 15.9 | 0.7 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Desktop | GeForce GTX Titan | 28 | PCIe 3.0 x16 | 2688:224:48 | 836-993 | 6008 | 6G | 288 | 384 | GDDR5 | 4500 | 1300-1500 | 250 | 18.0 | 6.0 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Desktop | GeForce GTX 780 Ti | 28 | PCIe 3.0 x16 | 2880:240:48 | 876-928 | 7000 | 3G | 336 | 384 | GDDR5 | 5046 | 210 | 250 | 20.2 | 0.8 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Professional | Quadro K6000 | 28 | PCIe 3.0 x16 | 2880:240:48 | 901.5 | 6008 | 12G | 288 | 384 | GDDR5 | 5196 | 1732 | 225 | 23.1 | 7.7 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Computing | Tesla K20X | 28 | PCIe 3.0 x16 | 2688 FP32, 896 FP64 | 732 | 5200 | 6G | 250 | 384 | GDDR5 | 3935 | 1312 | 235 | 16.7 | 5.6 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
NVIDIA | Mobile | Tegra 4 | 28 | Integrated | 72 | 672 | 1866 | N/A | 29.8 | 128 | LPDDR3 | 96.8 | 0 | 4 (SoC) | 24.2 | 0 | OpenGL ES 2.0 |
NVIDIA | Mobile | Tegra 5 | 28 | Integrated | 192 | 900 | ??? | N/A | ??? | ??? | LPDDR3 | 345.6 | 0 | ??? | ??? | 0 | D3D 11.0, OpenGL 4.4, CUDA 5.5, OpenCL 1.2 |
AMD | Desktop | Fusion APU 8670D | 32 | Integrated | 384:24:8 | 844-950 | 1066 | N/A | 29.9 | 128 | DDR3 | 648.2 | 0 | 100 (SoC) | 6.5 | 0 | D3D 11.0, OpenGL 4.3, OpenCL 1.2 |
AMD | Desktop | Fusion APU R7 2??D | 28 | Integrated | 512:32:16 | 720 | ??? | N/A | ??? | 128 | DDR3 | 856 | ??? | 95 (SoC) | ??? | ??? | D3D 11.2, OpenGL 4.3, OpenCL 1.2, Mantle |
AMD | Desktop | Radeon HD 8970 | 28 | PCIe 3.0 x16 | 2048:128:32 | 1000-1050 | 6000 | 3G | 288 | 384 | GDDR5 | 4300 | 1075 | 250 | 17.2 | 4.3 | D3D 11.1, OpenGL 4.3, OpenCL 1.2 |
AMD | Professional | Radeon W9000 | 28 | PCIe 3.0 x16 | 2048:128:32 | 975 | 5500 | 6G | 264 | 384 | GDDR5 | 3993.6 | 998.4 | 274 | 12.4 | 3.6 | D3D 11.1, OpenGL 4.2, OpenCL 1.2 |
AMD | Desktop | Radeon R9 290X | 28 | PCIe 3.0 x16 | 2816:176:64 | 800-1000 | 5000 | 4G | 320 | 512 | GDDR5 | 5632 | 704 | 290 | 19.4 | 2.4 | D3D 11.2, OpenGL 4.3, OpenCL 1.2, Mantle |
Intel | Desktop | Iris Pro Graphics 5200 | 22 | Integrated | 160:8:4 | 400-1300 | 1600 | N/A | 25.6 | 128 | DDR3 | 832 | 0 | 28 (SoC) | 29.7 | 0 | D3D 11.1, OpenGL 4.0, OpenCL 1.2 |
Intel | Computing | Xeon Phi 3100P | 22 | PCIe 2.0 x16 | 228 x86 | 1100 | 3750 | 6G | 240 | 512 | GDDR5 | 2000 | 1000 | 300 | 6.7 | 3.3 | OpenMP, OpenCL, MKL |
Intel | Computing | Xeon Phi 5110P | 22 | PCIe 2.0 x16 | 240 x86 | 1053 | 5000 | 8G | 320 | 512 | GDDR5 | 2022 | 1011 | 225 | 9.0 | 4.5 | OpenMP, OpenCL, MKL |
Intel | Computing | Xeon Phi 7120P | 22 | PCIe 2.0 x16 | 244 x86 | 1238-1333 | 5500 | 16G | 352 | 512 | GDDR5 | 2416 | 1208 | 300 | 8.1 | 4.0 | OpenMP, OpenCL, MKL |
Qualcomm | Mobile | Adreno 330 | 28 (Snapdragon 800) | Integrated | 128 | 450 | 800 | N/A | 12.8 | 128 | LPDDR3 | 129.6 | 0 | 3 (SoC) | 43.2 | 0 | D3D 9.3, OpenGL ES 3.0, OpenCL 1.2 |
Qualcomm | Mobile | Adreno 420 | 28 (Snapdragon 805) | Integrated | ??? | 500 | 800 | N/A | 25.6 | 256 | LPDDR3 | ??? | ??? | ??? | ??? | ??? | D3D 11.1, OpenGL ES 3.0, OpenCL 1.2 |
Imagination | Mobile | PowerVR SGX554 MP4 | 32 (A6X) | Integrated | 128 | 300 | 533 | N/A | 17 | 256 | LPDDR3 | 76.8 | 0 | 4 (SoC) | 19.2 | 0 | D3D 9.3, OpenGL 2.1, OpenGL ES 2.0, OpenCL 1.1 |
Imagination | Mobile | G6430 | 28 (A7) | Integrated | 256 | 450 | ??? | N/A | ??? | ??? | LPDDR3 | 115.2 | 0 | ??? | ??? | 0 | D3D 10.0, OpenGL 3.2, OpenGL ES 3.0 |
ARM | Moblie | Mali-T604 MP4 | 32 (Exynos 5 Dual) | Integrated | 64 | 533 | 800 | N/A | 12.8 | 128 | LPDDR3 | 68 | 0 | 4 (SoC) | 17.0 | 0 | D3D 9.1, OpenGL ES 3.0, OpenCL 1.1 |
ARM | Mobile | Mali-T760 MP16 | ??? | Integrated | ??? | 600 | ??? | N/A | ??? | ??? | LPDDR3 | 326.4 | 0 | ??? | ??? | 0 | D3D 11.1, OpenGL ES 3.0, OpenCL 1.1 |
Vivante | Mobile | GC4000 | 40 (K3V2) | Integrated | 32 | 680 | ??? | N/A | ??? | ??? | LPDDR3 | 43.5 | 0 | 4 (SoC) | 10.9 | 0 | D3D 9.3, OpenGL ES 3.0, OpenGL 3.0, OpenCL 1.2 |