-
Notifications
You must be signed in to change notification settings - Fork 2.3k
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
[llvm] Further improve LLVM 10 compatibility by using RTLD_DEEPBIND #1355
[llvm] Further improve LLVM 10 compatibility by using RTLD_DEEPBIND #1355
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for fixing this - I'm not sure if this will break Taichi on other Linux/OS X environment, since how RTLD_DEEPBIND
works is unclear to me.
At least we should only do this for Linux? Here RTLD_DEEPBIND
is used for both Linux and Mac.
Also note that LLVM 10 is statically linked into libtaichi.so
- there's no libLLVM-10.so
.
Well, but the stack trace tells me:
See? The stupid |
Codecov Report
@@ Coverage Diff @@
## master #1355 +/- ##
=======================================
Coverage 85.58% 85.58%
=======================================
Files 19 19
Lines 3371 3371
Branches 624 624
=======================================
Hits 2885 2885
Misses 356 356
Partials 130 130 Continue to review full report at Codecov.
|
Actually this might be the reason of this failure. I found rust-lang/rust#18671 (comment) and https://xamarin.github.io/bugzilla-archives/57/57742/bug.html. Both reported that statically linked LLVM has caused this problem. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good. Let's see how this works. Thank you!
Co-authored-by: Yuanming Hu <[email protected]>
TI_ASSERT(var.snode()->num_active_indices == 0); | ||
TI_ASSERT_INFO( | ||
var.snode()->num_active_indices == 0, | ||
"Please always use 'x[None]' (instead of simply 'x') to access any 0-D tensor." |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
"Please always use 'x[None]' (instead of simply 'x') to access any 0-D tensor." | |
"Please always use 'x[None]' (instead of simply 'x') to access any 0-D tensor."); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry for merging without proofread! Did a OFT fix in #1356.
* Revert "[skip ci] revert ti.chain_compare" This reverts commit d1a356e. * add func eval twice test according to @yuanming-hu * fix #1355 (comment) * [skip ci] enforce code format Co-authored-by: Taichi Gardener <[email protected]>
Related issue = #1326
[Click here for the format server]
Failed to start OpenGL on a very old NVIDIA card which doesn't support compute shader, RTLD_DEEPBIND seems fix that issue.