Err-6:CL_OUT_OF_HOST_MEMORY

[senliu]

Installation was smooth.

When running the example, got the following error:
Setting up the scoring function ... done.
Using heuristic search_depth
Analyzing the binding site ... done.
GPU Platform: NVIDIA CUDA
GPU Device: NVIDIA TITAN V
GPU Device: NVIDIA TITAN V
GPU Device: NVIDIA TITAN V
Using random seed: 1954754933

Build kernel 1 from source
OpenCL version: 3.0

Build kernel 2 from source
OpenCL version: 3.0

Err-6:CL_OUT_OF_HOST_MEMORY

My host has over 200GB+ memory available.

[asarigun]

Hi,

I didn’t run into any issues when allocating resources via Slurm. For example, this works fine:

srun --gpus=1 --ntasks=1 --cpus-per-task=1 --mem=16G --pty bash

The GPU VRAM is around 15 GB in that setup, and I also tested the same code on Google Colab without encountering any errors.

Since your system clearly has plenty of host memory, this might be an issue with the OpenCL installation or configuration. It could help to check the OpenCL runtime or driver versions, and possibly try running with fewer GPUs to isolate the cause.

[senliu]

Just use Slurm to run the test, and the error was the same. My GPU is Titan V, CUDA 12.4.131, OpenCL 3.0 (installed with PocketVina), and the information from clinfo seems normal. The memory of Titan V is 12GB. Should that be the problem? And any modification of the code to make it work?

…

On Tue, Jul 1, 2025 at 8:17 PM Ahmet Sarıgün ***@***.***> wrote: *asarigun* left a comment (BIMSBbioinfo/PocketVina#4) <#4 (comment)> Hi, I didn’t run into any issues when allocating resources via Slurm. For example, this works fine: srun --gpus=1 --ntasks=1 --cpus-per-task=1 --mem=16G --pty bash The GPU VRAM is around 15 GB in that setup, and I also tested the same code on Google Colab <https://colab.research.google.com/drive/10I2Ydu9-5P4uOTRSDNOhP7gA2qUK8-mX?usp=sharing> without encountering any errors. Since your system clearly has plenty of host memory, this might be an issue with the OpenCL installation or configuration. It could help to check the OpenCL runtime or driver versions, and possibly try running with fewer GPUs to isolate the cause. — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA73VNKLLMO7L6PORI5Y5F33GJ35PAVCNFSM6AAAAACAPW5J5SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTAMRTG43DKMRXGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>