Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Fix and optimize StateMonitor template #201

Closed
denisalevi opened this issue Apr 14, 2021 · 1 comment
Closed

Fix and optimize StateMonitor template #201

denisalevi opened this issue Apr 14, 2021 · 1 comment
Labels
bug optimisation

Comments

@denisalevi
Copy link
Member

Statemonitor only works for <1024 recorded neurons and the implementation seems rather inefficient.

This is a follow-up issue from #50. Here are my relevant comments from that issue:

The StateMonitor is also not working working for many neurons. Currently the StateMonitor is called with 1 block and as many threads as there are Neurons (or Synapses etc) to record. So if there are more Neurons then the maximum number of threads, the kernel fails to launch.

Originally posted by @denisalevi in #50 (comment)

And the global memory writes are not coalesced. Currently we have a 2D data structure of dimensions indices x record_times (vector of vectors) for each variable monitor. And we fill that in the kernel like this

monitor[tid][current_iteration] = ...

For coalesced writes we could just "transpose" the monitor data structure so we can use

monitor[current_iteration][tid] = ...

We might have to resort the monitor in the end though since it might then not fit with the format that Brian expects to read back.

Originally posted by @denisalevi in #50 (comment)
denisalevi added a commit that referenced this issue Apr 14, 2021
@denisalevi
Add test that fails for #201
1c743d1
This was referenced Apr 14, 2021
Merged
Closed
denisalevi added a commit that referenced this issue Apr 14, 2021
@denisalevi
Add test that fails for #201
f62749c
denisalevi added a commit that referenced this issue Apr 14, 2021
@denisalevi
Add test that fails for #201
15f97ef
@denisalevi denisalevi mentioned this issue Apr 14, 2021
Closed
denisalevi added a commit that referenced this issue Aug 4, 2021
@denisalevi
Add test that fails for #201
f7d3b42
denisalevi added a commit that referenced this issue Aug 6, 2021
@denisalevi
Add test that fails for #201
056cba7
denisalevi added a commit that referenced this issue Aug 6, 2021
@denisalevi
Add test with SpikeMonitor of >1024 variables
b8fcb0f 
Fails until #201 is fixed
@denisalevi denisalevi mentioned this issue Mar 3, 2022
Closed
denisalevi added a commit that referenced this issue Mar 31, 2022
@denisalevi
Refactor and fix statemonitor template
fd3173c 
Fixes bug in #201, without optimizing it yet.
denisalevi added a commit that referenced this issue Mar 31, 2022
@denisalevi
Modify state monitor test for >1024 threads
0eda13a 
Now it actually tests something instead just failing to notify about
issue #201 (which is solved now)
denisalevi added a commit that referenced this issue Mar 31, 2022
@denisalevi
Refactor and fix statemonitor template
a774976 
Fixes bug in #201, without optimizing it yet.
denisalevi added a commit that referenced this issue Mar 31, 2022
@denisalevi
Modify state monitor test for >1024 threads
2d19fbc 
Now it actually tests something instead just failing to notify about
issue #201 (which is solved now)
This was referenced Mar 31, 2022
Open
Merged
@denisalevi
Copy link
Member Author

Closed with PR #279. Further optimization options are collected in #278.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bug optimisation
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@denisalevi