open system call in strace

Accelerated Computing CUDA CUDA Programming and Performance
1

hi all,

i am using open system call to open the file in c as well in cuda version.

while running the program in c i am getting 0.8sec open system call time using strace

whereas same thing in cuda version , i am getting 4.1 sec time

and it is showing double the no of calls issued in cuda version whereas i am not changing file open part in both the codes.

please help

2

More specifically in cuda version

open(“/dev/nvidia0”, O_RDWR) and open(“/dev/nvidia1”, O_RDWR) is taking both getting 4 secs …

how can avoid this.