Group Abstract

Message Boards

WOLFRAM COMMUNITY

362 Views

2 Replies

2 Total Likes

View groups...

Follow this post

Share this post:

GROUPS:

Why is my GPU running so slowly?

Jürgen Kanz

Jürgen Kanz, @ JuergenKanz

Posted 1 month ago

I recently upgraded my GPU to an NVIDIA RTX 3060 to handle more local calculations. Therefore, I conducted a simple test to calculate the *AbsoluteTiming* using *TargetDevice, considering "CPU", "GPU", and "CUDA" as an alternative to "GPU". I've included the code I used: Clear[img] img = RandomImage[] AbsoluteTiming[ NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "CPU"]] (0.0306286,"igneous rock") AbsoluteTiming[ NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "GPU"]] (0.0575578,"igneous rock") AbsoluteTiming[ NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "CUDA"]] (0.05472, "igneous rock"*) Can anyone tell me why my GPU is running so slowly?

I recently upgraded my GPU to an NVIDIA RTX 3060 to handle more local calculations. Therefore, I conducted a simple test to calculate the AbsoluteTiming using TargetDevice, considering "CPU", "GPU", and "CUDA" as an alternative to "GPU". I've included the code I used:

Clear[img]
img = RandomImage[]
AbsoluteTiming[
 NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "CPU"]]
(*0.0306286,"igneous rock"*)
AbsoluteTiming[
 NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "GPU"]]
(*0.0575578,"igneous rock"*)
AbsoluteTiming[
 NetModel["Wolfram ImageIdentify Net V1"][img, TargetDevice -> "CUDA"]]
(*0.05472, "igneous rock"*)

Can anyone tell me why my GPU is running so slowly?

POSTED BY: Jürgen Kanz

2 Replies

Sort By:

Vladimir Ivanov

Vladimir Ivanov, Pulkovo observatory

Posted 12 days ago

Hi! My results for NVIDIA RTX 3060 are: {0.072155, "IgneousRock"} {0.0358761, "IgneousRock"} {0.03677, "IgneousRock"} Are you sure that your GPU is used? Try Needs["CUDALink`"] CUDAInformation[] Here is my output: {1 -> {"Name" -> "NVIDIA GeForce RTX 3060", "Clock Rate" -> 1837000, "Compute Capabilities" -> 8.6, "GPU Overlap" -> 1, "Maximum Block Dimensions" -> {1024, 1024, 64}, "Maximum Grid Dimensions" -> {2147483647, 65535, 65535}, "Maximum Threads Per Block" -> 1024, "Maximum Shared Memory Per Block" -> 49152, "Total Constant Memory" -> 65536, "Warp Size" -> 32, "Maximum Pitch" -> 2147483647, "Maximum Registers Per Block" -> 65536, "Texture Alignment" -> 512, "Multiprocessor Count" -> 28, "Core Count" -> 3584, "Execution Timeout" -> 1, "Integrated" -> False, "Can Map Host Memory" -> True, "Compute Mode" -> "Default", "Texture1D Width" -> 131072, "Texture2D Width" -> 131072, "Texture2D Height" -> 65536, "Texture3D Width" -> 16384, "Texture3D Height" -> 16384, "Texture3D Depth" -> 16384, "Texture2D Array Width" -> 32768, "Texture2D Array Height" -> 32768, "Texture2D Array Slices" -> 2048, "Surface Alignment" -> 512, "Concurrent Kernels" -> True, "ECC Enabled" -> False, "TCC Enabled" -> False, "Total Memory" -> 12884246528}}

Hi!

My results for NVIDIA RTX 3060 are:

{0.072155, "IgneousRock"}
{0.0358761, "IgneousRock"}
{0.03677, "IgneousRock"}

Are you sure that your GPU is used? Try

Needs["CUDALink`"]
CUDAInformation[]

Here is my output:

{1 -> {"Name" -> "NVIDIA GeForce RTX 3060", "Clock Rate" -> 1837000, 
   "Compute Capabilities" -> 8.6, "GPU Overlap" -> 1, 
   "Maximum Block Dimensions" -> {1024, 1024, 64}, 
   "Maximum Grid Dimensions" -> {2147483647, 65535, 65535}, 
   "Maximum Threads Per Block" -> 1024, 
   "Maximum Shared Memory Per Block" -> 49152, 
   "Total Constant Memory" -> 65536, "Warp Size" -> 32, 
   "Maximum Pitch" -> 2147483647, 
   "Maximum Registers Per Block" -> 65536, "Texture Alignment" -> 512,
    "Multiprocessor Count" -> 28, "Core Count" -> 3584, 
   "Execution Timeout" -> 1, "Integrated" -> False, 
   "Can Map Host Memory" -> True, "Compute Mode" -> "Default", 
   "Texture1D Width" -> 131072, "Texture2D Width" -> 131072, 
   "Texture2D Height" -> 65536, "Texture3D Width" -> 16384, 
   "Texture3D Height" -> 16384, "Texture3D Depth" -> 16384, 
   "Texture2D Array Width" -> 32768, 
   "Texture2D Array Height" -> 32768, 
   "Texture2D Array Slices" -> 2048, "Surface Alignment" -> 512, 
   "Concurrent Kernels" -> True, "ECC Enabled" -> False, 
   "TCC Enabled" -> False, "Total Memory" -> 12884246528}}

POSTED BY: Vladimir Ivanov

Jürgen Kanz

Jürgen Kanz, @ JuergenKanz

Posted 4 days ago

Sorry for late response. I didn’t receive a message about your information. Meantime it was possible for me to complain the card and I will go for a new one. Your results are on a speed level I would expect from a GPU support . Thank you.

POSTED BY: Jürgen Kanz

Reply to this discussion

Reply Preview

Attachments

Remove Add a file to this post

Follow this discussion

or Discard

Feedback