CARLsim4 segfaults on Ubuntu 18.04

21 views
Skip to first unread message

Vladislav Sekulic

unread,
Oct 16, 2018, 12:44:54 AM10/16/18
to CARLsim: A GPU-accelerated Spiking Neural Network (SNN) Simulator
Hello! I'm trying to use CARLsim4 on a modern Ubuntu 18.04 system with an NVIDIA GTX 1080. I have cloned the git repo and installed the nvidia 410 drivers as well as CUDA Toolkit 9.2. I initially tried using CUDA Toolkit 10 but there were too many problems building CARLsim4, so I reverted to 9.2.

Finally, I got CARLsim4 to build, but the hello_world example segfaults, as follows:


$ make
/usr/local/cuda-9.2/bin/nvcc -ccbin g++ -I/opt/CARL/include -L/opt/CARL/lib -Wno-deprecated-gpu-targets   src/main_hello_world.cpp -o hello_world -lcarlsim -lcurand
$ ./hello_world 
*********************************************************************************
********************      Welcome to CARLsim 4.0      ***************************
*********************************************************************************

***************************** Configuring Network ********************************
Starting CARLsim simulation "hello world" in USER mode
Random number seed: 42
Running COBA mode:
  - AMPA decay time            =     5 ms
  - NMDA rise time (disabled)  =     0 ms
  - GABAa decay time           =     6 ms
  - GABAb rise time (disabled) =     0 ms
  - GABAb decay time           =   150 ms


************************** Global Network Configuration *******************************
The number of neurons in the network (numN) = 126
The number of regular neurons in the network (numNReg:numNExcReg:numNInhReg) = 9:9:0
The number of poisson neurons in the network (numNPois:numNExcPois:numInhPois) = 117:117:0
The maximum axonal delay in the network (maxDelay) = 1
CUDA devices Configuration:
  - Number of CUDA devices          =         1
  - CUDA device ID with max GFLOPs  =         0
  + Use CUDA device[0]              = GeForce GTX 1080
  + CUDA Compute Capability (CC)    =       6.1

+ Local Network (8)
|-+ Group List:
Segmentation fault (core dumped)


C++ is not my forte but if necessary I may try to attach a debugger and see if I can poke around a bit. 

Thanks in advance for any input!
Reply all
Reply to author
Forward
0 new messages