faster cluster, now does not requires to know number of modules #68
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
VinInn
changed the title
|
Thanks! |
fwyzard
reviewed
Jun 4, 2018
| auto niter = (msize-first)/blockDim.x; | ||
| assert(niter<10); | ||
| for (int i=0; i<niter+1; ++i) jmax[i]=msize; | ||
| // for (int k=0; k<niter+1; ++k) jmin[k]=first+k*blockDim.x+1; | ||
| for (int k=0; k<niter+1; ++k) jmax[k]=msize; | ||
|
|
||
| while (go) { |
There was a problem hiding this comment.
I know it was like this before, but I'm wondering - is it actually safe to use __syncthreads() inside a while loop?
Isn't it a problem if some threads run a different number of loops, or take different branches (i.e. continue) ?
There was a problem hiding this comment.
cuda-memcheck did not detect any race
There was a problem hiding this comment.
All threads goes through the while loop together.
the continue is in the for and there is no __syncthreads() in the for-loop.
So no divergencies besides the for loop
felicepantaleo
requested changes
Jun 4, 2018
| // if (h_clus[i]==i) seeds.push_back(i); // only if no renumbering | ||
| } | ||
|
|
||
| std::cout << "found " << std::accumulate(nclus,nclus+MaxNumModules,0) << ' ' << clids.size() << " clusters" << std::endl; |
| d_clus.get(), | ||
| d_debug.get(), | ||
| n | ||
| ); | ||
|
|
||
| cuda::memory::copy(&nModules,d_moduleStart.get(),sizeof(uint32_t)); | ||
| std::cout << "found " << nModules << " Modules active" << std::endl; |
There was a problem hiding this comment.
I don't mind keeping the cout since this is a just a test file, not one used in production.
| d_clus.get(), | ||
| d_debug.get(), | ||
| n | ||
| ); | ||
|
|
||
| cuda::memory::copy(&nModules,d_moduleStart.get(),sizeof(uint32_t)); |
| threadsPerBlock = 256; | ||
| blocksPerGrid = MaxNumModules; //nModules; | ||
|
|
||
|
|
||
|
|
||
| std::cout |
| // Launch CUDA Kernels | ||
|
|
||
|
|
||
| int threadsPerBlock = 256; | ||
| int blocksPerGrid = (numElements + threadsPerBlock - 1) / threadsPerBlock; | ||
| std::cout |
There was a problem hiding this comment.
this is a test; it is normal to print something!
| uint32_t nModules=0; | ||
| cuda::memory::copy(d_moduleStart.get(),&nModules,sizeof(uint32_t)); | ||
|
|
||
|
|
||
| cuda::memory::copy(d_id.get(), h_id.get(), size16); | ||
| cuda::memory::copy(d_x.get(), h_x.get(), size16); | ||
| cuda::memory::copy(d_y.get(), h_y.get(), size16); | ||
| cuda::memory::copy(d_adc.get(), h_adc.get(), size8); |
| cuda::memory::copy(d_id.get(), h_id.get(), size16); | ||
| cuda::memory::copy(d_x.get(), h_x.get(), size16); | ||
| cuda::memory::copy(d_y.get(), h_y.get(), size16); | ||
| cuda::memory::copy(d_adc.get(), h_adc.get(), size8); | ||
| cuda::memory::device::zero(d_debug.get(),size32); | ||
| cuda::memory::device::zero(d_debug.get(),size32); |
There was a problem hiding this comment.
this is a cuda memset to zero? If yes, is there a async version with stream?
fwyzard
approved these changes
Jun 4, 2018
fwyzard
pushed a commit
that referenced
this pull request
Jun 4, 2018
fwyzard
pushed a commit
that referenced
this pull request
Jun 4, 2018
fwyzard
pushed a commit
that referenced
this pull request
Oct 8, 2020
fwyzard
pushed a commit
that referenced
this pull request
Oct 19, 2020
fwyzard
pushed a commit
that referenced
this pull request
Oct 20, 2020
fwyzard
pushed a commit
that referenced
this pull request
Oct 23, 2020
fwyzard
pushed a commit
that referenced
this pull request
Nov 6, 2020
fwyzard
pushed a commit
that referenced
this pull request
Nov 16, 2020
fwyzard
pushed a commit
that referenced
this pull request
Dec 25, 2020
fwyzard
pushed a commit
that referenced
this pull request
Dec 29, 2020
fwyzard
pushed a commit
that referenced
this pull request
Dec 29, 2020
fwyzard
pushed a commit
that referenced
this pull request
Feb 12, 2021
Co-authored-by: Anders <aryd@cern.ch>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
addresses #66 as well