On 2015-04-20 08:53, Matthias Klose wrote:
On 04/20/2015 05:27 AM, Yang Chen wrote:
Hi Matthias,
Thanks for reporting the issue.
I couldn't reproduce the GCC failure on my machine running Ubuntu 14.04
(detailed below). Looks like the crash was specific to
powerpc64le-linux-gnu,
which I don't have access to. Do you happen to know which C-Reduce
transformation caused the code expansion, or do you have any saved
log/intermediate files that were produced during the reduction? Thanks.
yes, only seen on this target. I didn't try yet to reproduce this with
a cross
compiler. I had the log of the creduce run attached to the email.
What else
would you need to reproduce this?
Matthias
My apologies - I overlooked the attached log.
The increase was caused by simple-inliner, which doubled the size of the
test after it finished.
Can you attach the file named as something like testcase.best in the dir
where you performed the reduction, which might be stress_cpu.best in
your case? Thanks.
- Yang