-
Notifications
You must be signed in to change notification settings - Fork 2.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Search performance gap between cpu and gpu(nq<gpu_search_threshold) #2437
Comments
compare between CPU and GPU with nq=100, k=100
GPU slower than CPU consists of 3 parts:
==================================================================
|
Need to do:
|
I had compared the disassemble code between CPU and GPU version, they are almost the same. |
cpu version map uids costs time statistics:
cpu version idmap costs time statistics:
gpu version map uids time statistics:
gpu version idmap costs time statistics:
As result shown above, each test's performance are not stable because of the system jitter. |
I use google-perftool to do the profiling on the IDMAP process, the result shows that over 90% time costs on function omp_get_num_procs, it's very strange. |
Search time ranged is expected, put this issue hold |
@cydrain @del-zhenwu |
close |
Describe the bug
performance gap between cpu/gpu(nq<gpu_search_threshold):
cpu:
===
gpu:
Steps/Code to reproduce behavior
dataset:
sift-1b
Expected behavior
A clear and concise description of what you expected to happen.
Environment details
branch: 0.9.1
index:
config:
cpu config:
===
gpu config:
Screenshots
If applicable, add screenshots to help explain your problem.
Additional context
Add any other context about the problem here.
The text was updated successfully, but these errors were encountered: