Skip to content

Add safeguard to CUDA runtime bindings for device initialization #100

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

Open
asindarov wants to merge 1 commit into
base: master
Choose a base branch
from
+9 −4
Open

Add safeguard to CUDA runtime bindings for device initialization #100

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 6 additions & 2 deletions source/dcompute/driver/cuda/runtime.d
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,15 +38,19 @@ private static bool _threadReady = false; // safety-net flag
// in module-dependency order. No locking needed here.
shared static this()
{
_initPlatform();
if (__dcompute_reflect(ReflectTarget.CUDA,0)) {

@thewilsonator thewilsonator Jul 2, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

hmm, does this work?
Pretty sure this will always be false for host code.

@thewilsonator thewilsonator Jul 2, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

cc @badnikhil

@asindarov asindarov Jul 2, 2026 •
edited
Loading

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

yes it worked on my mac, I don't have computer with nvidia chip, therefore cannot check the true case

_initPlatform();
}
}

// Per-thread init: thread-local static constructor
// D runtime calls this automatically for every thread (including main) when
// that thread begins. For the main thread it runs after shared static this().
static this()
{
_initThread();
if (__dcompute_reflect(ReflectTarget.CUDA,0)) {
_initThread();
}
}

// Public API
Expand Down
5 changes: 3 additions & 2 deletions source/dcompute/tests/main.d
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,13 +18,14 @@ import std.typecons;
import std.conv : to;
import std.math.traits : isNaN;

import dcompute.driver.cuda.unified_buffer;
import dcompute.tests.dummykernels : saxpy;

version(DComputeTestOpenCL)
import dcompute.driver.ocl;
else version(DComputeTestCUDA)
else version(DComputeTestCUDA) {
import dcompute.driver.cuda.unified_buffer;
import dcompute.driver.cuda;
}
else
static assert(false, "Need to test something!");

Expand Down