cacosh/csin failures on s390x since CORE-MATH integration

cacosh/csin failures on s390x since CORE-MATH integration

[Mark Wielaard]

Hi,

Since the CORE-MATH integration the s390x buildbot shows the following
4 failures:

FAIL: math/test-float-cacosh
FAIL: math/test-float-csin
FAIL: math/test-float32-cacosh
FAIL: math/test-float32-csin

https://builder.sourceware.org/buildbot/#/builders/glibc-fedora-s390x

Bunsen logs:
https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float-cacosh.out
https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float-csin.out
https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float32-cacosh.out
https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float32-csin.out

Cheers,

Mark

Re: cacosh/csin failures on s390x since CORE-MATH integration

[Adhemerval Zanella Netto]

On 07/11/24 10:06, Mark Wielaard wrote:
> Hi,
> 
> Since the CORE-MATH integration the s390x buildbot shows the following
> 4 failures:
> 
> FAIL: math/test-float-cacosh
> FAIL: math/test-float-csin
> FAIL: math/test-float32-cacosh
> FAIL: math/test-float32-csin
> 
> https://builder.sourceware.org/buildbot/#/builders/glibc-fedora-s390x
> 
> Bunsen logs:
> https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float-cacosh.out
> https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float-csin.out
> https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float32-cacosh.out
> https://builder.sourceware.org/testrun/8e8049073b623360f34a0f72e7bd0beb4ede3891?filename=math%2Ftest-float32-csin.out
> 
> Cheers,
> 
> Mark

It is expected [1], I had to adjust it on x86, arm, and powerpc as well.
The __expm1 is exact for the input range, but due how csin/cacoshf
handles the expm1 result there is addition cancellation errors.

In this case someone will need to run 'make update-ulps' on a real
hardware and update the libm-test-ulps. 

[1] https://sourceware.org/pipermail/libc-alpha/2024-October/161131.html

Re: cacosh/csin failures on s390x since CORE-MATH integration

[Mark Wielaard]

Hi Adhemerval,

On Thu, 2024-11-07 at 13:40 +0000, Adhemerval Zanella Netto wrote:
> It is expected [1], I had to adjust it on x86, arm, and powerpc as well.
> The __expm1 is exact for the input range, but due how csin/cacoshf
> handles the expm1 result there is addition cancellation errors.
> 
> In this case someone will need to run 'make update-ulps' on a real
> hardware and update the libm-test-ulps. 

Aha, found https://sourceware.org/glibc/wiki/Regeneration#Math_files
Running make regen-ulps on an s390x machine produced:

Automatic regeneration of ULPs complete.
Difference between the current baseline and the new baseline is:
--- ../sysdeps/s390/fpu/libm-test-ulps	2024-11-07 16:53:35.036318696 +0000
+++ /home/mjw/glibc-obj/math/NewUlps	2024-11-07 17:09:45.376836323 +0000
@@ -228,7 +228,7 @@
 
 Function: Real part of "cacosh_upward":
 double: 4
-float: 3
+float: 4
 ldouble: 6
 
 Function: Imaginary part of "cacosh_upward":
@@ -720,6 +720,7 @@
 ldouble: 1
 
 Function: Imaginary part of "csin":
+float: 1
 ldouble: 1
 
 Function: Real part of "csin_downward":
Copy /home/mjw/glibc-obj/math/NewUlps to ../sysdeps/s390/fpu/libm-test-ulps (relative to source).

Will sent a patch after another make check run to verify the results.

Thanks,

Mark

> [1] https://sourceware.org/pipermail/libc-alpha/2024-October/161131.html