brian-team / brian2cuda

A brian2 extension to simulate spiking neural networks on GPUs

https://brian2cuda.readthedocs.io/

brian-team/brian2cuda Issues

Why the volatile GPU utilization is so low? I run the same code respectively on the CPU(cpp_standalone) and GPU(brian2cuda), and i expect that it can run quickly by using brian2ccuda. However, it runs much slower... Can someone explain this?
Closed 7 months ago1
Bypass nvidia-smi when using a Jetson platform
Closed 8 months ago6
Support Windows
Updated a year ago1
Optimise brian2CUDA
Updated 2 years ago2
Brian2Hears and Brian2CUDA
Closed 2 years ago3
Brian2Cuda Uninstalls Brian2 2.5.1 and Installs 2.4.2 Which Won't Import
Closed 2 years ago3
Recent Brian2 update PR broke benchmark scripts
Updated 2 years ago1
Circular eventspaces fail when source group runs on a different clock than synapses
Updated 2 years ago
Spikes are lost when changing delays between `run` calls
Updated 2 years ago3
Fix memory leak when having multiple `run` calls
Updated 2 years ago
Fix `ReferencError` in spatialneuron tests
Closed 2 years ago3
Refactor test suite scripts
Updated 2 years ago
Optimize our `SpikeMonitor` for `Subgroups`
Updated 2 years ago
Fix `SpikeMonitor` for `Subgroup`s
Closed 2 years ago1
Make sure default compile flags are used correctly and are well documented
Updated 2 years ago
Investigate and document performance effects when working with `Subgroup`s
Updated 3 years ago1
Consider partitioning eventspaces when using `Subgroup`s
Updated 3 years ago4
Copy all eventspace counters to host efficiently at each time step
Updated 3 years ago
Optimize `PopulationRateMonitor`
Updated 3 years ago
Impelement brian2cuda preference file support
Updated 3 years ago
Investigate if reducing registers by modifying code generation is possible
Updated 3 years ago1
Optimize `StateMonitor`
Updated 3 years ago
Implement special case connectivity matrices with pre/post IDs for models that effect only pre/post variables
Updated 3 years ago5
Needs patch to run with Brian 2.4.2
Closed 3 years ago2
Check if storing the size of synapse groups is necessary?
Updated 3 years ago1
Refactor benchmarking scripts and update generated plots
Updated 3 years ago
Compilation is failing with CUDA 11.4
Closed 3 years ago6
Call reset kernel only with as many threads as there are spiking neurons (not as there are neurons in total)
Updated 3 years ago
Parallelize effect application in heterogeneous delay mode over additional CUDA blocks
Updated 3 years ago1
Reduce number of threads in push kernel for heterogeneous delays in bundle mode
Updated 3 years ago
Report when models are using too many registers for full occupancy
Updated 3 years ago
Block dimension calculation can lead to non-optimal occupancy
Updated 3 years ago1
Single precision mode is broken
Closed 3 years ago3
Fix remaining issues in single precision mode
Updated 3 years ago
Check if single precision math functions are used correctly
Updated 3 years ago
Single precision f-literal replacement fails for variables with double digit + dot in name
Updated 3 years ago3
Make inverted assertion tests single precision compatible
Updated 3 years ago
[how to] changing Synapses Weights after run with cpp/cuda
Closed 3 years ago3
Failing getting CUDA code
Closed 3 years ago2
Implement better metric for automatically choosing the best GPU
Updated 3 years ago
Support GSL integration
Updated 3 years ago1
Profiling failed on SCIoI GPU node
Updated 3 years ago1
Implement testst that make sure the compiler, CUDA and preferences are compatible
Updated 3 years ago
Clean up and simplify thresholder template
Updated 3 years ago
Decide on debug mode for CUDA Standalone
Closed 3 years ago1
Make single-precision version of `exprel` and `clip` typesafe
Updated 3 years ago1
Use `size_t` datatype instead of `int` for indices and sizes
Updated 3 years ago
Verify that our checks for poisson function calls with scalar lambda are correct
Updated 3 years ago
SpikeMonitor fails for multiple runs
Updated 3 years ago1
Call kernels with maximum the number of units instead of always multiples of 1024
Updated 3 years ago