Bug 289656 - vex amd64->IR: unhandled instruction bytes: 0xC5 0xF0 0x57 0xC9 0x66 0xF 0x2F 0xC8
Summary: vex amd64->IR: unhandled instruction bytes: 0xC5 0xF0 0x57 0xC9 0x66 0xF 0x2F...
Status: RESOLVED DUPLICATE of bug 273475
Alias: None
Product: valgrind
Classification: Developer tools
Component: vex (show other bugs)
Version: 3.7.0
Platform: Compiled Sources Linux
: NOR normal
Target Milestone: ---
Assignee: Julian Seward
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-12-23 11:12 UTC by Mario Valle
Modified: 2011-12-23 11:24 UTC (History)
1 user (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Mario Valle 2011-12-23 11:12:14 UTC
Version:           3.7.0 (using Devel) 
OS:                Linux

Using Memcheck (valgrind-3.7.0) as below on a Cray XK6 where the computing nodes are AMD Opteron(TM) Processor 6272:

export OMP_NUM_THREADS=16

aprun valgrind --suppressions=./src/fast.supp --gen-suppressions=all --leak-check=full --show-reachable=yes ./my_exec  {its arguments}

Dies with the following error:
vex amd64->IR: unhandled instruction bytes: 0xC5 0xF0 0x57 0xC9 0x66 0xF 0x2F 0xC8

This happens inside a dscal_ BLAS routine as provided by Intel MKL (sequential).


Reproducible: Always

Steps to Reproduce:
Happens in a scientific application compiled with icc 12.0.3 and using for BLAS the Intel MKL (version coming with the compiler) library. MKL used is the sequential version.

Actual Results:  
Memcheck crashes at the first call to dscal_ BLAS routine.

Expected Results:  
The complete check of the application.

On this machine I have no choice than to run it under the Cray aprun dispatcher.
Anyway, if could be helpful to look at the executable (statically linked) take the file valgrind.tgz from ftp.cscs.ch/out/mvalle
Please let me know when you downloaded it so I can delete it.
Comment 1 Tom Hughes 2011-12-23 11:24:17 UTC
That's an AVX instruction.

*** This bug has been marked as a duplicate of bug 273475 ***