[AMBER] Amber16 on GPUs and speed differences between CUDA_VISIBLE_DEVICES=0, 1 or 0, 2

From: Neale, Christopher Andrew <cneale.lanl.gov>
Date: Mon, 19 Dec 2016 18:31:19 +0000

I learned how to get a bit more information, which perhaps will help:

$ nvidia-smi topo -m
        GPU0 GPU1 GPU2 GPU3 CPU Affinity
GPU0 X PHB SOC SOC 0-5,12-17
GPU1 PHB X SOC SOC 0-5,12-17
GPU2 SOC SOC X PHB 6-11,18-23
GPU3 SOC SOC PHB X 6-11,18-23


  X = Self
  SOC = Path traverses a socket-level link (e.g. QPI)
  PHB = Path traverses a PCIe host bridge
  PXB = Path traverses multiple PCIe internal switches
  PIX = Path traverses a PCIe internal switch

Thank you,

AMBER mailing list
Received on Mon Dec 19 2016 - 11:00:02 PST
Custom Search