Skip to content

[NVIDIA] Move cuda graphs instantiation to model compilation#807

Open
ytorzuk-altran wants to merge 3 commits intoopenvinotoolkit:masterfrom
ytorzuk-altran:feature/instantiate-cuda-graphs-during-compilation
Open

[NVIDIA] Move cuda graphs instantiation to model compilation#807
ytorzuk-altran wants to merge 3 commits intoopenvinotoolkit:masterfrom
ytorzuk-altran:feature/instantiate-cuda-graphs-during-compilation

Conversation

@ytorzuk-altran
Copy link
Contributor

Details:

  • Cuda graphs instantiation was moved to model compilation
    Ticket:
  • CVS-119834

@ytorzuk-altran ytorzuk-altran requested a review from a team as a code owner December 27, 2023 17:04
@github-actions github-actions bot added the category: NVIDIA plugin OpenVINO NVIDIA plugin label Dec 27, 2023
@p-durandin
Copy link
Contributor

build_jenkins

@p-durandin
Copy link
Contributor

buid_jenkins

@p-durandin
Copy link
Contributor

build_jenkins

if (!memory_proxy.Get().cudaGraphContext().is_initialized())
memory_proxies.push_back(std::move(memory_proxy));
}
auto cuda_thread_pool = std::dynamic_pointer_cast<CudaThreadPool>(cuda_stream_executor_);
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please add null check.

topology_runner.Run(inferRequestContext, memory);
threadContext.stream().synchronize();
l.lock();
active_threads_num--;
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Isn't it better to use atomic for the counter?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

category: NVIDIA plugin OpenVINO NVIDIA plugin

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants

Comments