adam27x / hybrid_bc Goto Github PK

Hi @Adam27X ,

I run your code on a node which equipped with 4 NVIDIA K20c GPUs. The host machine is running RedHat OS version 4.4.5-6. The CUDA version is 9.0 using the “-arch=sm35” compute compatibility flag. But the algorithm just show 1 GPU is running. How can I run the code on multi-GPUs and how can I run the code on multi-node with multi-GPUs?

Hi guys,

i think the link is borked unfortunately :(

Hi,
I am testing your code with small hand-written instances but the executable crash.

Can you provide small instances (max 16 nodes) ?

Best,
Flavio