This is the mail archive of the
mailing list for the GCC project.
Re: [patch] libstdc++: Limit Riemann zeta testcases on simulator
- From: Jesper Nilsson <jesper dot nilsson at axis dot com>
- To: Ed Smith-Rowland <3dw4rd at verizon dot net>
- Cc: gcc-patches at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org, hans-peter dot nilsson at axis dot com
- Date: Thu, 11 Oct 2007 17:16:03 +0200
- Subject: Re: [patch] libstdc++: Limit Riemann zeta testcases on simulator
- References: <20071011074127.GC23461@axis.com> <470E1D6D.firstname.lastname@example.org>
On Thu, Oct 11, 2007 at 08:56:13AM -0400, Ed Smith-Rowland wrote:
> I might be OK to limit the number of iterations for *one* target. The
> data set that is being looped over is also a numeric accuracy check so
> I'm a little reluctant to truncate the tests for *all* targets.
Yes, I suspected as much. Could there be a smaller set that covers
the same corner cases?
> OTOH, I'm a little shocked at how long it takes to run the computations!
> I don't have access to this architecture. Does this arch have floating
> point? Or is it simulated?
CRIS has no hardware floating point support, and running it in
the simulator makes it even slower.
> Maybe we might want to xfail on targets where you really aren't going to
> be crunching numbers.
Yes, but the real problem is not that the test fails, but that it takes
a very long time to complete it. In fact the test can be made to
pass by increasing the timeout to more than 700 seconds.
However, it might be argued that running a full numeric accuracy test when
cross-compiling is somewhat unnecessary. :-)
> I'll go check on my machines to see if Zeta is one of the more slow
Thank you. I must confess that my primary motivation is making the
testsuite run faster, but I'd like to do this without (unecessarily)
losing test coverage.
/^JN - Jesper Nilsson
Jesper Nilsson -- email@example.com