CUDA for GeForce GTX 1050 Ti

Hello,

I have a laptop with the GeForce GTX 1050 Ti graphics card. Can I ever expect this card to become supported by the CUDA toolkit or am I out of luck?

Thanks.

What makes you think that the GTX 1050 Ti is not supported by CUDA?

It’s not on the list…

Which list?

https://developer.nvidia.com/cuda-gpus

That would appear to be an inadvertent omission in “the list” then. Such lists are maintained by humans and don’t come with built-in automated tests that check for accuracy.

The currently shipping version of CUDA, CUDA 9.1, supports all GPUs with compute capability 3.0 or higher. The GTX 1050 Ti (which has been on the market for more than a year), has compute capability 6.1, as can be seen in this alternative list, for example:

https://en.wikipedia.org/wiki/CUDA#GPUs_supported

My laptop has a 1050 Ti and this is the deviceQuery output:
Detected 1 CUDA Capable device(s)

Device 0: “GeForce GTX 1050 Ti”
CUDA Driver Version / Runtime Version 9.1 / 8.0
CUDA Capability Major/Minor version number: 6.1
Total amount of global memory: 4096 MBytes (4294967296 bytes)
( 6) Multiprocessors, (128) CUDA Cores/MP: 768 CUDA Cores
GPU Max Clock rate: 1620 MHz (1.62 GHz)
Memory Clock rate: 3504 Mhz
Memory Bus Width: 128-bit
L2 Cache Size: 1048576 bytes
Maximum Texture Dimension Size (x,y,z) 1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
Maximum Layered 1D Texture Size, (num) layers 1D=(32768), 2048 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(32768, 32768), 2048 layers
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 49152 bytes
Total number of registers available per block: 65536
Warp size: 32
Maximum number of threads per multiprocessor: 2048
Maximum number of threads per block: 1024
Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)
Maximum memory pitch: 2147483647 bytes
Texture alignment: 512 bytes
Concurrent copy and kernel execution: Yes with 2 copy engine(s)
Run time limit on kernels: No
Integrated GPU sharing Host Memory: No
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
CUDA Device Driver Mode (TCC or WDDM): WDDM (Windows Display Driver Model)
Device supports Unified Addressing (UVA): Yes
Device PCI Domain ID / Bus ID / location ID: 0 / 1 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.1, CUDA Runtime Version = 8.0, NumDevs = 1, Device0 = GeForce GTX 1050 Ti
Result = PASS

When I go through the installation it gives an error screen after performing the system compatibility check.

“NVIDIA Installer failed” then a big list of items that weren’t installed such as:

  • CUBLAS Development
  • CURAND Runtime
  • Visual Profiler
  • ...etc.

You may want to try https://devtalk.nvidia.com/default/board/58/cuda-setup-and-installation