k-Wave

1. Administratorte
   

   Member
   Joined: Aug '19
   Posts: 2

   Good Day.

   I'm trying to help a prof run the kspaceFirstOrder3D-CUDA v1.2
   program from matlab and after a bunch of library trickery
   to get it to work on CentOS 7.6, it runs but then:

   !!! K-Wave experienced a fatal error !!! │
   ├───────────────────────────────────────────────────────────────┤
   │ Error: Insufficient CUDA driver version. Install the latest │
   │ drivers. │
   ├───────────────────────────────────────────────────────────────┤
   │ Execution terminated

   The machine is running the very latest NVIDIA GPU driver
   (430.40) and the CUDA that is installed is 10.1.

   Is this too new? Do we need to drop back on both the driver
   to the latest 390.x and perhaps an earlier version of CUDA?

   Any help greatly appreciated.

   Cheers.

   Posted 5 years ago #

2. Jiri Jaros
   

   Developer
   Joined: Feb '12
   Posts: 115

   Hi Administratorte,
   Could you run nvidia-smi and and post the result?
   and also kspace... --version?

   Posted 5 years ago #

3. Administratorte
   

   Member
   Joined: Aug '19
   Posts: 2

   Hi. Thanks very much for getting back to me. The two outputs
   are below. Thanks again for any help you can provide.

   Cheers.

   **

   Thu Aug 22 06:41:32 2019
   +-----------------------------------------------------------------------------+
   | NVIDIA-SMI 430.40 Driver Version: 430.40 CUDA Version: 10.1 |
   |-------------------------------+----------------------+----------------------+
   | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
   | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
   |===============================+======================+======================|
   | 0 Quadro P400 Off | 00000000:01:00.0 On | N/A |
   | 34% 34C P8 N/A / N/A | 334MiB / 1999MiB | 0% Default |
   +-------------------------------+----------------------+----------------------+

   +-----------------------------------------------------------------------------+
   | Processes: GPU Memory |
   | GPU PID Type Process name Usage |
   |=============================================================================|
   | 0 2129 G /usr/bin/X 143MiB |
   | 0 23212 G /usr/bin/gnome-shell 188MiB |
   +-----------------------------------------------------------------------------+

   ┌───────────────────────────────────────────────────────────────┐
   │ kspaceFirstOrder3D-CUDA v1.2 │
   ├───────────────────────────────────────────────────────────────┤
   │ Selected GPU device id: Failed │
   ├───────────────────────────────────────────────────────────────┤
   │ Build information │
   ├───────────────────────────────────────────────────────────────┤
   │ Build number: kspaceFirstOrder3D v3.5 │
   │ Build date: Sep 4 2017 │
   │ Build time: 23:31:29 │
   │ Git hash: cc750a58cc6b3d43f96f2c3bba959fbb9e69556f │
   ├───────────────────────────────────────────────────────────────┤
   │ Operating system: Linux x64 │
   │ Compiler name: GNU C++ 5.4.0 │
   │ Instruction set: Intel SSE 2 │
   ├───────────────────────────────────────────────────────────────┤
   │ CUDA runtime: N/A │
   │ CUDA driver: 0.0 │
   │ CUDA code arch: N/A │
   ├───────────────────────────────────────────────────────────────┤
   │ CUDA device id: N/A │
   │ CUDA device name: N/A │
   │ CUDA capability: N/A │
   ├───────────────────────────────────────────────────────────────┤
   │ Contact email: jarosjir@fit.vutbr.cz │
   │ Contact web: http://www.k-wave.org │
   ├───────────────────────────────────────────────────────────────┤
   │ Copyright (C) 2017 Jiri Jaros and Bradley Treeby │
   └───────────────────────────────────────────────────────────────┘
   ┌───────────────────────────────────────────────────────────────┐
   │ !!! K-Wave experienced a fatal error !!! │
   ├───────────────────────────────────────────────────────────────┤
   │ Error: Insufficient CUDA driver version. Install the latest │
   │ drivers. │
   ├───────────────────────────────────────────────────────────────┤
   │ Execution terminated │
   └───────────────────────────────────────────────────────────────┘

   Posted 5 years ago #

4. Jiri Jaros
   

   Developer
   Joined: Feb '12
   Posts: 115

   Hi Administratorte,

   Since you have a Pascal GPU, you will have to download source codes and recompile the code because by the time the code was released this GPU did not exist.

   Please download latest CUDA, open the makefile and change CUDA_ARCH variable

   CUDA_ARCH = --generate-code arch=compute_30,code=sm_30 \
               --generate-code arch=compute_32,code=sm_32 \
               --generate-code arch=compute_35,code=sm_35 \
               --generate-code arch=compute_37,code=sm_37 \
               --generate-code arch=compute_50,code=sm_50 \
               --generate-code arch=compute_52,code=sm_52 \
               --generate-code arch=compute_53,code=sm_53 \
               --generate-code arch=compute_60,code=sm_60 \
               --generate-code arch=compute_61,code=sm_61 \
               --generate-code arch=compute_62,code=sm_62 \
               --generate-code arch=compute_70,code=sm_70 \
               --generate-code arch=compute_72,code=sm_72 \
               --generate-code arch=compute_75,code=sm_75 \

   Before you do so, try to run the CPU code first. The P400 GPU you have is very bad at CUDA computation. It is a very basic low-end card. The CPU will probably be much faster. Moreover, the tiny amount fo GPU memory will limit the size of the simulation to something like 200x200x200 gridpoints.

   Best
   Jiri

   Posted 5 years ago #

RSS feed for this topic

Reply

You must log in to post.