thanks ali!
i have a follow up question, this time regarding the SPLASH-2 benchmarks.
reading that running SPLASH-2 in FS mode is less problematic than running it in
SE, I tried running it in FS mode. I downloaded the SPLASH-2 benchmarks from
m5sim.org, from a link in the mailing list archives. There are binaries
included, but they're not compiled for alpha-linux, so I have to compile* them
from the source code. running "make" at the command line while at the source
folder yields the following command line output:
-----------------------------------------------------------------------------
cc -c -I../../../../parmacs.upc.3/pthreads code.c
code.C: In function âmainâ:
code.C:116: warning: incompatible implicit declaration of built-in function
âexitâ
code.C: In function âLog_base_2â:
code.C:227: warning: incompatible implicit declaration of built-in function
âexitâ
code.C: In function âstartrunâ:
code.C:342: warning: comparison between pointer and integer
cc -c -I../../../../parmacs.upc.3/pthreads code_io.c
cc -c -I../../../../parmacs.upc.3/pthreads load.c
load.C: In function âprinttreeâ:
load.C:168: warning: incompatible implicit declaration of built-in function
âexitâ
cc -c -I../../../../parmacs.upc.3/pthreads grav.c
cc -c -I../../../../parmacs.upc.3/pthreads getparam.c
getparam.C: In function âgetparamâ:
getparam.C:57: warning: incompatible implicit declaration of built-in function
âstrlenâ
getparam.C: In function âgetiparamâ:
getparam.C:76: warning: comparison between pointer and integer
getparam.C: In function âgetlparamâ:
getparam.C:88: warning: comparison between pointer and integer
getparam.C: In function âgetbparamâ:
getparam.C:98: warning: comparison between pointer and integer
getparam.C:100: warning: incompatible implicit declaration of built-in
function âstrchrâ
getparam.C: In function âgetdparamâ:
getparam.C:115: warning: comparison between pointer and integer
getparam.C: In function âmatchnameâ:
getparam.C:155: warning: comparison between pointer and integer
getparam.C: In function âextrvalueâ:
getparam.C:168: warning: comparison between pointer and integer
cc -c -I../../../../parmacs.upc.3/pthreads util.c
util.C: In function âerrorâ:
util.C:97: warning: incompatible implicit declaration of built-in function
âexitâ
cc code.o code_io.o load.o grav.o getparam.o util.o
-I../../../../parmacs.upc.3/pthreads -o BARNES -non_shared -lm
-L../../../../parmacs.upc.3/pthreads -lparmacs_pthreads -lpthread -lexc
cc: unrecognized option '-non_shared'
../../../../parmacs.upc.3/pthreads/libparmacs_pthreads.a: could not read
symbols: Archive has no index; run ranlib to add one
collect2: ld returned 1 exit status
make: *** [BARNES] Error 1
------------------------------------------------------------------------------
taking clue from the "unrecognized option" line, i then modified the makefile,
removing the "-non_shared" argument. running make then yielded...
------------------------------------------------------------------------------
cc code.o code_io.o load.o grav.o getparam.o util.o
-I../../../../parmacs.upc.3/pthreads -o BARNES -lm
-L../../../../parmacs.upc.3/pthreads -lparmacs_pthreads -lpthread -lexc
../../../../parmacs.upc.3/pthreads/libparmacs_pthreads.a: could not read
symbols: Archive has no index; run ranlib to add one
collect2: ld returned 1 exit status
make: *** [BARNES] Error 1
----------------------------------------------------------------------------
switching to the right folder i then ran
ranlib libparmacs_pthreads.a
returning to the barnes folder and running make again...
------------------------------------------------------------------------------
cc code.o code_io.o load.o grav.o getparam.o util.o
-I../../../../parmacs.upc.3/pthreads -o BARNES -lm
-L../../../../parmacs.upc.3/pthreads -lparmacs_pthreads -lpthread -lexc
../../../../parmacs.upc.3/pthreads/libparmacs_pthreads.a: could not read
symbols: Archive has no index; run ranlib to add one
collect2: ld returned 1 exit status
make: *** [BARNES] Error 1
------------------------------------------------------------------------------
what seems to be the problem? again, thanks for taking time to read my message
and answer my questions.
*-i have compiled the crosscompiler successfully, but after seeing the first
make run fail, i decided to try compiling it for my host machine first
Ali Saidi <[EMAIL PROTECTED]> wrote: The last line is very important, and no
the binary wasn't executed. It's not a question of input files or anything
else, the problem is that binary. I would guess that its a OSF/1 binary and not
a Linux binary or it's not a binary at all. Just take some random file of bits,
make it executable and run it. You'll get the same error.
Ali
On Oct 7, 2007, at 6:10 AM, w. tan wrote:
actually the entire command line output went like
# ./twolf00.peak.ev6 test
modprobe: FATAL: Could not load /lib/modules/2.6.13/modules.dep: No such file
or directory
modprobe: FATAL: Could not load /lib/modules/2.6.13/modules.dep: No such file
or directory
bash: ./twolf00.peak.ev6: cannot execute binary file
(sorry for not including the last line in my last email, my bad)
so i think the benchmark was not executed. and besides, no output file was
produced. the name of the executable changed since i switched to the
precompiled binary i downloaded off the Internet(so the crosscompiler can't be
the culprit).
am i missing something? i mean, what i did basically was place the binary and
the input files in the image(latest-linux.img, the one that came with the m5
distro), boot the simulator, and then run the executable.
is that the correct procedure for running benchmarks in FS mode?
thanks a lot for taking time to answer my questions.
Ali Saidi <[EMAIL PROTECTED]> wrote: That error shouldn't be a problem. The
benchmark is almost certainly still running, just some syscall it did attempted
to load a module, however that module isn't necessary.
Ali
On Oct 5, 2007, at 9:49 PM, w. tan wrote:
i am trying to run cpu2000 benchmarks in full system mode, and already compiled
the benchmarks using the crosscompiler. after copying the binaries to the image
i then tried running the benchmarks within the simulated system. i got the
error
# ../TimberWolfSC stdcell
modprobe: FATAL: Could not load /lib/modules/2.6.13/modules.dep: No such file
or directory
modprobe: FATAL: Could not load /lib/modules/2.6.13/modules.dep: No such file
or directory
after looking around the net for a while i came upon this message wherein steve
advised another user to have the benchmark(it was not exactly the same
benchmark though) statically linked to get around this. i then tried statically
linking twolf by adding "-static" to the makefile's "gcc" argument and then
crosscompiling again. the binary's size did increase, but the results are still
the same, it still looks for modules.dep. any ideas? thanks a lot.
---------------------------------
Take the Internet to Go: Yahoo!Go puts the Internet in your pocket: mail, news,
photos & more. _______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
---------------------------------
Moody friends. Drama queens. Your life? Nope! - their life, your story.
Play Sims Stories at Yahoo! Games.
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
---------------------------------
Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user panel
and lay it on us._______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
