Updated on June 9, 2011: GTX 500 and CUDA 4.0 compute capabilities features.
The Compute Capability describes the features supported by a CUDA hardware. First CUDA capable hardware like the GeForce 8800 GTX have a compute capability (CC) of 1.0 and recent GeForce like the GTX 480 have a CC of 2.0. Knowing the CC can be useful for understanting why a CUDA based demo can’t start on your system.
Here is a comparative table that will help you to know the CC of your NVIDIA graphics card:
Multiprocessors | CUDA cores | Compute Capability | |
GeForce GTX 680 | 8 | 1536 | 3.0 |
GeForce GTX 590 | 2 x 16 | 2 x 512 | 2.0 (R280.28) |
GeForce GTX 580 | 16 | 512 | 2.0 |
GeForce GTX 570 | 15 | 480 | 2.0 |
GeForce GTX 560 Ti | 8 | 384 | 2.1 |
GeForce GTX 560 | 7 | 336 | 2.1 |
GeForce GT 440 | 2 | 96 | 2.1 |
GeForce GTX 480 | 15 | 480 | 2.0 |
GeForce GTX 470 | 14 | 448 | 2.0 |
GeForce GTX 465 | 11 | 352 | 2.0 |
GeForce GTX 295 | 2×30 | 2×240 | 1.3 |
GeForce GTX 280/GTX 285 | 30 | 240 | 1.3 |
GeForce GTX 260 | 24 | 192 | 1.3 |
GeForce 210 | 2 | 16 | 1.2 |
GeForce GT 240 | 12 | 96 | 1.2 |
GeForce GT 220 | 6 | 48 | 1.2 |
GeForce GT 130 | 12 | 96 | 1.1 |
GeForce GT 120 | 4 | 32 | 1.1 |
GeForce GTS 250 | 16 | 128 | 1.1 |
GeForce 9800 GX2 | 2×16 | 2×128 | 1.1 |
GeForce 9800 GTX | 16 | 128 | 1.1 |
GeForce 9800 GT | 14 | 112 | 1.1 |
GeForce 9600 GSO | 12 | 96 | 1.1 |
GeForce 9600 GT | 8 | 64 | 1.1 |
GeForce 8800 GTX/Ultra | 16 | 128 | 1.0 |
GeForce 8800 GT | 14 | 112 | 1.1 |
GeForce 8800 GTS | 12 | 96 | 1.0 |
GeForce 8600 GT/GTS | 4 | 32 | 1.1 |
GeForce 8400 GS/GT | 2 | 16 | 1.1 |
You can use GPU Caps Viewer to know the CC of your card:
Compute capability displayed by GPU Caps Viewer
This table just shows the stupidity of nVidia’s naming and rebranding strategy. Confusion all over the place. I specially like that 8800GTX/Ultra is 1.0 but 8600GT is 1.1 compatible! Does it make any sense? Only in THE WAY IT IS MEANT TO BE FAKED world of nVidia.
Instead of 2.1 as mentioned, GTX 580 is with a compute capability of 2.0
What is meant by “maximum no. of resident warps per multiprocessor”, which is 48 in GF100. A block can have upto 32 warps and there can be 8 blocks/multiprocessor,so it should be 32*8, so what is meant by 48 warps/multiprocessor???
plz help
This table is full with errors. The GTX 570 has the compute capability of 2.0
CC of GTX 580 and 570 fixed.
JeGX, can you please update this for Kepler GPUs? They are Compute Capability 3.0 according to Nvidia.
http://forums.nvidia.com/index.php?showtopic=225312&view=findpost&p=1386379
@Kepler: updated!