About GPU

Autonomous Vehicles DRIVE PX 2 General
1

Question1: What is the difference between DRIVE PX 2 AutoChauffeur and NVIDIA Tegra X2?

Question2: Which GPU works when I running predict models?

Question3: Why the totalMemory of NVIDIA Tegra X2 is 6.25GiB but the freeMemory is only 2.63GiB.

Here is the program running tips:

name: DRIVE PX 2 AutoChauffeur major: 6 minor: 1 memoryClockRate(GHz): 1.29
pciBusID: 0000:04:00.0
totalMemory: 3.75GiB freeMemory: 3.68GiB

name: NVIDIA Tegra X2 major: 6 minor: 2 memoryClockRate(GHz): 1.275
pciBusID: 0000:00:00.0
totalMemory: 6.25GiB freeMemory: 2.63GiB

2

Dear 763659972,

  1. Both are based on parker SoC. DRIVE PX 2 has an extra PASCAL based dGPU attached to parker SoC. Please run CUDA device query sample to the specs.
  2. CUDA devicequery lists the available GPUs in ascending order of compute power. By default the computation performed on GPU deviceID 0. For ex, on PX 2 It runs on dGPU. You can choose specific GPU by setting CUDA_VISIBLE_DEVICES flag or via cudaSetDevice().
3

Hi 763659972,

Could you file your topic of TX2 to TX2 forum board? It sounds an issue that worth investigation.