Dear Flexus users,

Attached to this message is an updated patch for compiling and running  
Flexus on 64-bit Linux.  This patch should be applied using:

patch -p0 < flexus_64b.diff

while you are in your $FLEXUS_ROOT directory.  This patch supersedes  
earlier patches sent to the list.  You don't need to do apply any of  
those (and shouldn't).   Please check that you have a 64-bit version  
of simics by running the command:

find $PATH_TO_SIMICS -name simics-common

If you get a file under amd64-linux, then you have 64-bit Simics;  
otherwise, x86-linux indicates you have a 32-bit version.

You will need to edit the environment section at the top of  
makefile.defs.  The relevant line is SIMICS_ARCH.  Use x86-linux for  
32-bit systems, and amd64-linux for 64-bit.

Let us know if you have problems.  We were able to verify with Chris  
Bienia that this patch solved his problems.  We plan to release a new  
Flexus version that will incorporate these and other fixes this Summer.

Thanks,
Brian

-------------- next part --------------
A non-text attachment was scrubbed...
Name: flexus_64b.diff
Type: application/octet-stream
Size: 28424 bytes
Desc: not available
URL: 
<http://sos.ece.cmu.edu/pipermail/simflex/attachments/20090420/87b85575/attachment-0001.obj>
-------------- next part --------------




Begin forwarded message:
>
> On Apr 15, 2009, at 7:47 PM, Christian Bienia wrote:
>
>> Brian,
>>
>> I applied the patch, but I get compilation errors. Here's what I
>> experienced:
>>
>> * When I apply the diff the patch tool complains that a few of the
>> referenced files don't exist. I tried to find them but they either  
>> got
>> deleted or renamed in a non-obvious way. I skipped the files.
>>
>> * When I tried to build the patched source it breaks right at the  
>> beginning
>> because it generates a warning and "-Werror" was added to the  
>> compiler flags
>> by the patch. I removed "-Werror" and started to build everything  
>> again.
>>
>> * The build process then stops at a more serious error involving  
>> missing
>> members in a structure. Here's the output:
>>
>> ==========
>> Checking simics v9_iface_gcc core
>> Building Boost.DateTime
>> Building Boost.Regex
>> Building Boost.Serialization
>> Building Boost.Iostreams
>> /usr/local/gcc-4.2.1/bin/g++ -isystem
>> /n/fs/parsec/bin/flexus-3.0.0/boost_1_33_1 -isystem
>> /n/fs/parsec/bin/virtutech/simics-3.0.31/src/include
>> -I/u/cbienia/parsec/bin/flexus-3.0.0 -c -O3 -funroll-loops
>> -fno-strict-aliasing -Wall -fmessage-length=160 -x c++ -fPIC -DX86_64
>> -DTARGET_PLATFORM=v9 -DSELECTED_DEBUG=iface wiring.cpp -o
>> wiring.v9_iface_gcc_o
>> wiring.cpp: In function ?bool  
>> Flexus::Wiring::initializeParameters()?:
>> wiring.cpp:142: error: ?struct uArchConfiguration? has no member  
>> named
>> ?InstructionsPerCheckpoint?
>> wiring.cpp:143: error: ?struct uArchConfiguration? has no member  
>> named
>> ?StoreAddressesPerCheckpoint?
>> In file included from wiring.cpp:377:
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp: At global scope:
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:57: warning:  
>> ?Flexus::SharedTypes::<unnamed>::MemoryMessageTag?
>> defined
>>  but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:65: warning:  
>> ?Flexus::SharedTypes::<unnamed>::ExecuteStateTag?
>> defined
>>  but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:73: warning:  
>> ?Flexus::SharedTypes::<unnamed>::uArchStateTag? defined
>>
>>  but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:81: warning: ?Flexus::SharedTypes::<unnamed>::MuxTag?  
>> defined but
>> not
>>  used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:89: warning: ?Flexus::SharedTypes::<unnamed>::BusTag?  
>> defined but
>> not
>>  used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:98: warning:  
>> ?Flexus::SharedTypes::<unnamed>::DirectoryEntryTag?
>>  defined but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> MemoryTransp
>> ort.hpp:107: warning:
>> ?Flexus::SharedTypes::<unnamed>::TransactionTrackerTag?
>>  defined but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> DirectoryTra
>> nsport.hpp:51: warning:
>> ?Flexus::SharedTypes::<unnamed>::DirectoryMessageTag?
>>  defined but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> DirectoryTra
>> nsport.hpp:69: warning:  
>> ?Flexus::SharedTypes::<unnamed>::DirMux2ArbTag?
>> defined
>>  but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> NetworkTrans
>> port.hpp:55: warning:  
>> ?Flexus::SharedTypes::<unnamed>::NetworkMessageTag?
>>  defined but not used
>> /u/cbienia/parsec/bin/flexus-3.0.0/components/Common/Transports/ 
>> NetworkTrans
>> port.hpp:64: warning:  
>> ?Flexus::SharedTypes::<unnamed>::ProtocolMessageTag?
>>  defined but not used
>> make[9]: *** [wiring.v9_iface_gcc_o] Error 1
>> make[8]: *** [simics-v9] Error 2
>> make[7]: *** [DSMFlex.OoO] Error 2
>> make[6]: *** [DSMFlex.OoO] Error 2
>> make[5]: *** [DSMFlex.OoO] Error 2
>> make[4]: *** [all] Error 2
>> make[3]: *** [all] Error 2
>> make[2]: *** [all] Error 2
>> make[1]: *** [all] Error 2
>> make: *** [all] Error 2
>> spin.CS.Princeton.EDU%
>> ==========
>>
>> Do you see a way to fix this? I looked at the code but it's not  
>> clear to me
>> how it's supposed to be.
>>
>> - Chris
>>
>>
>> -----Original Message-----
>> From: simflex-bounces at ece.cmu.edu [mailto:simflex- 
>> bounces at ece.cmu.edu] On
>> Behalf Of Brian Gold
>> Sent: Wednesday, April 15, 2009 2:36 PM
>> To: SimFlex software support
>> Subject: Re: [Simflex] flexus on 64-bit
>>
>> Chris, a quick follow-up: see the attached diff for a more up-to- 
>> date patch
>> to run on 64-bit platforms.  If you hit any errors, send the output  
>> to me.
>>
>> Thanks,
>> Brian
>>
>>
>

Reply via email to