Someone had that problem on the mailing list before, and it looks like
they solved it by using a newer version of the cross compiler,
specifically 4.2.1.
Since this is at least the second time someone has had this problem,
it would be nice if we updated the version on the wiki (NUDGE, NUDGE,
POKE, POKE). It looks like Ali put the link there on February 29th,
2008, and since I can't find any evidence to the contrary and since
Ali has historically been involved with SPARC, I assume he also
prepared the cross compiler.
Gabe
Quoting IC <[email protected]>:
Thank Ali and Malek.
I try SPARC architecture now.
I downloaded sparc64 cross compiler from M5 wiki and used it to compile the
test program of the m5threads
But the error message happened
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_stackgrow.o
test_stackgrow.cpp
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o
test_pthreadbasic.o test_pthreadbasic.cpp
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_pthread.o
test_pthread.cpp
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_atomic.o
test_atomic.cpp
test_atomic.cpp: In function `int main(int, const char**)':
test_atomic.cpp:102: warning: cast to pointer from integer of different size
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_barrier.o
test_barrier.cpp
test_barrier.cpp: In function `int main(int, const char**)':
test_barrier.cpp:53: warning: cast to pointer from integer of different size
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_lock.o
test_lock.cpp
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_malloc.o
test_malloc.cpp
test_malloc.cpp: In function `int main(int, const char**)':
test_malloc.cpp:89: warning: cast to pointer from integer of different size
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test_sieve.o
test_sieve.cpp
sparc64-unknown-linux-gnu-g++ -g -O3 -DM5_PROFILING -c -o test___thread.o
test___thread.cpp
test___thread.cpp: In function `void* run(void*)':
test___thread.cpp:54: warning: cast to pointer from integer of different
size
test___thread.cpp: In function `int main(int, char**)':
test___thread.cpp:77: warning: cast to pointer from integer of different
size
sparc64-unknown-linux-gnu-gcc -g -O3 -DM5_PROFILING -c ../pthread.c -o
../pthread.o
sparc64-unknown-linux-gnu-g++ -static -o test_stackgrow test_stackgrow.o
../pthread.o
/home/ic/x-tools/SPARC64/gcc-3.4.5-glibc-2.3.5/sparc64-unknown-linux-gnu/bin/../lib/gcc/sparc64-unknown-linux-gnu/3.4.5/../../../../sparc64-unknown-linux-gnu/bin/ld:
. ./pthread.o: check_relocs:
unhandled reloc type 72
../pthread.o: could not read symbols: File format not recognized
collect2: ld returned 1 exit status
make: *** [test_stackgrow] Error 1
Would anyone know the reason of the error?
thanks.
2010/10/31 Malek Musleh <[email protected]>
The error seems somewhat generic in the sense that it could be the
result of 1 of several things. I believe I have encoutered such an
error message previously. In my case, I think the issue was due to
when I had unaligned data variables in my benchmark. Alpha requires
aligned accesses.
I resolved the issue when I manually inserted the __attribute__
(alligned(X))) feature to shared structures/variables, where X is
number of bytes.
Also, I believe that m5 threads was more developed/tested for the
SPARC ISA, and still needed some work for ALPHA (reading the m5threads
README). If that's still the case, your error could be because of
that.
Malek
On Sat, Oct 30, 2010 at 5:07 PM, Ali Saidi <[email protected]> wrote:
> You're benchmark likely has a bug in it.
>
> Ali
>
> On Oct 29, 2010, at 1:20 AM, IC wrote:
>
>> Hello:
>>
>> I download m5threads code and using alpha crosscompiler (from m5 wiki
gcc-4.3.2, glibc-2.6.1 (NPTL,x86/32)) to make.
>>
>> All test program are built successfully.
>>
>> But when I run these test programs on ALPHA_SE, it occurs error.
>>
>> (modified splash2/run.py for these test program)
>>
---------------------------------------------------------------------------------
>> command line: ./build/ALPHA_SE/m5.opt ./configs/splash2/runMyBench.py
--rootdir /home/ic/M5sim/M5_SVN/M5/tags/m5threads -b MyBench -n 2
>> Global frequency set at 1000000000000 ticks per second
>> 0: system.remote_gdb.listener: listening for remote gdb #0 on port 7000
>> 0: system.remote_gdb.listener: listening for remote gdb #1 on port 7001
>> info: Entering event queue @ 0. Starting simulation...
>> info: Increasing stack size by one page.
>> panic: Tried to access unmapped address 0x70.
>> @ cycle 644000
>> [invoke:build/ALPHA_SE/arch/alpha/faults.cc, line 206]
>> Memory Usage: 153884 KBytes
>> For more information see: http://www.m5sim.org/panic/5932f339
>> Program aborted at cycle 644000
>> Aborted
>>
-------------------------------------------------------------------------------------
>>
>> Would somebody know this problem?
>>
>> Please help , Thanks
>>
>> _______________________________________________
>> 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
>
_______________________________________________
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
