-
Notifications
You must be signed in to change notification settings - Fork 21.4k
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.
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
Fix structured kernel codegen #49244
Conversation
see https://fb.quip.com/ceEdANd5iVsO RegisterMkldnnCPU kernels incorrectly used makeUnboxedOnly() calls to register add_.Tensor kernels. This is because the codegen incorrectly thought they're not c10-full. This PR fixes that. Differential Revision: [D25500246](https://our.internmc.facebook.com/intern/diff/D25500246/) [ghstack-poisoned]
💊 CI failures summary and remediationsAs of commit 7790021 (more details on the Dr. CI page):
🕵️ 2 new failures recognized by patternsThe following CI failures do not appear to be due to upstream breakages: pytorch_linux_bionic_py3_6_clang9_test (1/2)Step: "Run tests" (full log | diagnosis details | 🔁 rerun)
|
see https://fb.quip.com/ceEdANd5iVsO RegisterMkldnnCPU kernels incorrectly used makeUnboxedOnly() calls to register add_.Tensor kernels. This is because the codegen incorrectly thought they're not c10-full. This PR fixes that. Differential Revision: [D25500246](https://our.internmc.facebook.com/intern/diff/D25500246/) ghstack-source-id: 118411117 Pull Request resolved: #49244
@ezyang calling back to thread about explicit plumbing vs. dynamic scoping like I imagine you figured I would |
Stack from ghstack:
see https://fb.quip.com/ceEdANd5iVsO
RegisterMkldnnCPU kernels incorrectly used makeUnboxedOnly() calls to register add_.Tensor kernels. This is because the codegen incorrectly thought they're not c10-full.
This PR fixes that.
Differential Revision: D25500246