GTX 560 Ti number of cores 256 vs 384?

[AndyCUDA]

Hi there.

Just testing a new machine with a GTX 560 Ti, which has 384 cores, right?

Device query (Cuda 4.0) is reporting 256 (see below)… Is this a device query issue or do I have to do something to enable all 384?

Thanks for any help!

Andrew

CUDA Device Query (Runtime API) version (CUDART static linking)

There is 1 device supporting CUDA

Device 0: “GeForce GTX 560 Ti”

CUDA Driver Version: 4.0

CUDA Runtime Version: 3.10

CUDA Capability Major revision number: 2

CUDA Capability Minor revision number: 1

Total amount of global memory: 1008271360 bytes

Number of multiprocessors: 8

Number of cores: 256

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 49152 bytes

Total number of registers available per block: 32768

Warp size: 32

Maximum number of threads per block: 1024

Maximum sizes of each dimension of a block: 1024 x 1024 x 64

Maximum sizes of each dimension of a grid: 65535 x 65535 x 65535

Maximum memory pitch: 2147483647 bytes

Texture alignment: 512 bytes

Clock rate: 1.66 GHz

Concurrent copy and execution: Yes

Run time limit on kernels: Yes

Integrated: No

Support host page-locked memory mapping: Yes

Compute mode: Default (multiple host threads

can use this device simultaneously)

Concurrent kernel execution: Yes

Device has ECC support enabled: No

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 4.0, CUDA Runtime Versi

on = 3.10, NumDevs = 1, Device = GeForce GTX 560 Ti

PASSED

Press to Quit…

[seibert]

You are using an old copy of deviceQuery. The CUDA API actually does not have any function to return the number of CUDA cores. Instead, it returns the number of streaming multiprocessors and the compute capability. The deviceQuery program then multiplies the number of SMs by some factor depending on the compute capability of the card. Your version of deviceQuery predates the existence of capability 2.1, which has 48 CUDA cores per SM. Instead the deviceQuery program is assuming that all 2.x devices have 32 CUDA cores per SM.

So you don’t need to worry, although you might want to update your CUDA toolkit and SDK. :)