msun: Fix incorrect results in hypotl near underflow
#455
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
In the edited code, `t1` is intended to scale subnormal numbers, but
`t1=1` is needed instead of `t1=0` when preparing the scaling.
With this test program, compiled with `-fno-builtin`:
```c
int main() {
long double x = 0x3.6526795f4176ep-16384l;
long double y = 0x3.ffffffffffffep-16352l;
long double z = hypotl(x, y);
printf("z=%La\n", z);
return 0;
}
```
msun generates `z=0x0p+0`, whereas GNU libc generates
`z=0xf.ffffffffffff8p-16354`.
The result from msun is incorrect, as `hypot(x, y) >= |y|`, and thus
cannot be zero for nonzero `y`.
This issue was reported by zimmermann6 on GitHub to openlibm, which
repackages parts of msun. The test case is their work.
|
d3338f3 by dim@ fixed this shortly after the pull request was opened and merged to openlibm... |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This patch is the work of @kargl, fixing a bug reported by @zimmermann6 in JuliaMath/openlibm#224. @kargl's authorship has been preserved in the commit.
In the edited code,
t1is intended to scale subnormal numbers, butt1=1is needed instead oft1=0when preparing the scaling.With this test program, compiled with
-fno-builtin:msun generates
z=0x0p+0, whereas GNU libc generatesz=0xf.ffffffffffff8p-16354.The result from msun is incorrect, as
hypot(x, y) >= |y|, and thus cannot be zero for nonzeroy.This issue was reported by zimmermann6 on GitHub to openlibm, which repackages parts of msun. The test case is their work.