<div dir="ltr"><br><br><div class="gmail_quote">On Wed, Sep 24, 2008 at 9:22 PM, Anthony Chan <span dir="ltr"><<a href="mailto:chan@mcs.anl.gov">chan@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I tried 10.1 "icc -fast" on a simple program on a AMD box.<br>
<br>
> icc -fast ~/ts_C/ts_main_rc.c ; echo $?<br>
<div class="Ih2E3d">ipo: remark #11001: performing single-file optimizations<br>
</div>ipo: remark #11005: generating object file /tmp/ipo_iccSrFoSL.o<br>
0<br>
<br>
> a.out<br>
<br>
Fatal Error: This program was not built to run on the processor in your system.<br>
The allowed processors are: Intel(R) Core(TM) Duo processors and compatible Intel processors with supplemental Streaming SIMD Extensions 3 (SSSE3) instruction support.<br>
<br>
I tried -xT and -xP as well, both give the same error message. It appear<br>
-fast can't be used on at least AMD box.<br>
<div><div></div><div class="Wj3C7c"><br>
</div></div></blockquote><div><br>Hello, <br><br>There was a discussion about Intel's compiler creating slower code + locking down the application to specific CPU's(Intel's) on slashdot, if any of you has time, I recommend you read it.<br>
<br>Someone also pointed out a way to make icc compiled code to run on AMD with target CPU option turned on here :<br><br><a href="http://yro.slashdot.org/comments.pl?sid=155593&cid=13044057">http://yro.slashdot.org/comments.pl?sid=155593&cid=13044057</a><br>
<br><br>And this link points to the original thread :<br><br><a href="http://yro.slashdot.org/article.pl?sid=05/07/12/1320202&tid=142">http://yro.slashdot.org/article.pl?sid=05/07/12/1320202&tid=142</a><br><br> </div>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div><div class="Wj3C7c"><br>
<br>
----- "Rajeev Thakur" <<a href="mailto:thakur@mcs.anl.gov">thakur@mcs.anl.gov</a>> wrote:<br>
<br>
> I am able to reproduce the problem if I use the -fast option to icc.<br>
> But<br>
> without -fast it works. Wonder if it is a problem with icc. I used<br>
> version<br>
> 10.1.<br>
><br>
> Rajeev<br>
><br>
> > -----Original Message-----<br>
> > From: <a href="mailto:owner-mpich-discuss@mcs.anl.gov">owner-mpich-discuss@mcs.anl.gov</a><br>
> > [mailto:<a href="mailto:owner-mpich-discuss@mcs.anl.gov">owner-mpich-discuss@mcs.anl.gov</a>] On Behalf Of Robert<br>
> Kubrick<br>
> > Sent: Wednesday, September 24, 2008 11:54 AM<br>
> > To: <a href="mailto:mpich-discuss@mcs.anl.gov">mpich-discuss@mcs.anl.gov</a><br>
> > Subject: Re: [mpich-discuss] Compiling mpich2 with icc<br>
> ><br>
> > Compiling with ipcp returns this error:<br>
> ><br>
> > make[4]: Entering directory `/home/pts/archives/mpich2-1.1.0a1/src/<br>
> > mpid/common/locks'<br>
> > CC mpidu_process_locks.c<br>
> > ../../../include/pmi.h(740): error: parameter type involves<br>
> > pointer to array of unknown bound<br>
> > int PMI_Args_to_keyval(int *argcp, char *((*argvp)[]),<br>
> > PMI_keyval_t **keyvalp, int *size);<br>
> > ^<br>
> ><br>
> > compilation aborted for mpidu_process_locks.c (code 2)<br>
> > make[4]: *** [mpidu_process_locks.o] Error 2<br>
> > make[4]: Leaving directory `/home/pts/archives/mpich2-1.1.0a1/src/<br>
> > mpid/common/locks'<br>
> > make[3]: *** [all-redirect] Error 2<br>
> > make[3]: Leaving directory `/home/pts/archives/mpich2-1.1.0a1/src/<br>
> > mpid/common'<br>
> > make[2]: *** [all-redirect] Error 2<br>
> > make[2]: Leaving directory<br>
> > `/home/pts/archives/mpich2-1.1.0a1/src/mpid'<br>
> > make[1]: *** [all-redirect] Error 2<br>
> > make[1]: Leaving directory `/home/pts/archives/mpich2-1.1.0a1/src'<br>
> > make: *** [all-redirect] Error 2<br>
> ><br>
> ><br>
> > I tried nm -a libmpich.a and it didn't return the required<br>
> > symbols. I am not too familiar with nm, but I am not sure<br>
> > this symbol table is<br>
> > right:<br>
> ><br>
> > me@harper1:~/archives/mpich2-1.1.0a1> nm -a lib/libmpich.a<br>
> > |grep MPIR_Version_date me@harper1:~/archives/mpich2-1.1.0a1><br>
> > nm -a lib/libmpich.a<br>
> ><br>
> > mpidu_process_locks.o:<br>
> > 00000000 a .strtab<br>
> > 00000000 a .symtab<br>
> > 00000000 n __ildata<br>
> > w __ildata_included<br>
> > 00000000 a mpidu_process_locks.c<br>
> ><br>
> > mpidu_atomic_primitives.o:<br>
> > 00000000 a .strtab<br>
> > 00000000 a .symtab<br>
> > 00000000 n __ildata<br>
> > w __ildata_included<br>
> > 00000000 a mpidu_atomic_primitives.c<br>
> ><br>
> > mpidu_queue.o:<br>
> > 00000000 a .strtab<br>
> > 00000000 a .symtab<br>
> > 00000000 n __ildata<br>
> > w __ildata_included<br>
> > 00000000 a mpidu_queue.c<br>
> ><br>
> > dataloop.o:<br>
> > 00000000 a .strtab<br>
> > 00000000 a .symtab<br>
> > 00000000 n __ildata<br>
> > w __ildata_included<br>
> > 00000000 a dataloop.c<br>
> > ...<br>
> > ...<br>
> > ...<br>
> ><br>
> > On Sep 24, 2008, at 12:00 PM, Rajeev Thakur wrote:<br>
> ><br>
> > > Can you do an "nm" on libmpich.a and see if those symbols are in<br>
> > > there.<br>
> > > Also, use icpc as the C++ compiler.<br>
> > ><br>
> > > Rajeev<br>
> > ><br>
> > >> -----Original Message-----<br>
> > >> From: <a href="mailto:owner-mpich-discuss@mcs.anl.gov">owner-mpich-discuss@mcs.anl.gov</a><br>
> > >> [mailto:<a href="mailto:owner-mpich-discuss@mcs.anl.gov">owner-mpich-discuss@mcs.anl.gov</a>] On Behalf Of<br>
> > Robert Kubrick<br>
> > >> Sent: Wednesday, September 24, 2008 10:55 AM<br>
> > >> To: <a href="mailto:mpich-discuss@mcs.anl.gov">mpich-discuss@mcs.anl.gov</a><br>
> > >> Subject: Re: [mpich-discuss] Compiling mpich2 with icc<br>
> > >><br>
> > >> initinfo.c compiles fine, but there are several warnings during<br>
> > >> compilation. I am attaching here the complete make log.<br>
> > >><br>
> > >><br>
> > >><br>
> > ><br>
> ><br>
> ><br>
<br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Mert TÜRKAL<br>
</div>