CUDA on Ubuntu 12.04 x86_64

Message boards : Number crunching : CUDA on Ubuntu 12.04 x86_64
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1454838 - Posted: 17 Dec 2013, 16:04:44 UTC

Hello,

have a CUDA capable GPU in my notebook.

BOINC-Manger Messages:
Di 17 Dez 2013 13:31:09 CET | | CUDA: NVIDIA GPU 0: GeForce GT 330M (driver version unknown, CUDA version 5.5, compute capability 1.2, 1023MB, 600MB available, 182 GFLOPS peak)
Di 17 Dez 2013 13:31:09 CET | | OpenCL: NVIDIA GPU 0: GeForce GT 330M (driver version 319.37, device version OpenCL 1.0 CUDA, 1023MB, 600MB available, 182 GFLOPS peak)

If I attach to SETI@Home project in Boinc-Manger, does it use the CUDA-Seti@Home automatically or do I have to
compile the CUDA version by myself?

Thanks

ID: 1454838 · Report as offensive
Profile ivan
Volunteer tester
Avatar

Send message
Joined: 5 Mar 01
Posts: 783
Credit: 348,560,338
RAC: 223
United Kingdom
Message 1454871 - Posted: 17 Dec 2013, 20:10:18 UTC - in response to Message 1454838.  

I have a CUDA capable GPU in my notebook.

If I attach to SETI@Home project in Boinc-Manger, does it use the CUDA-Seti@Home automatically or do I have to compile the CUDA version by myself?

It should use the Astropulse application automatically (see this computer for example), but AP workunits tend to be scarce. To run MB you'd need to compile it or acquire a copy from elsewhere, at the moment. One of the big problems is the huge variety of Linux environments, especially versioning of libraries -- I've even just run into problems compiling the latest BOINC because the developers have moved to a more modern version of autoconf than I have on some of my machines.
ID: 1454871 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1454880 - Posted: 17 Dec 2013, 20:30:32 UTC - in response to Message 1454838.  

Hello,

have a CUDA capable GPU in my notebook.

BOINC-Manger Messages:
Di 17 Dez 2013 13:31:09 CET | | CUDA: NVIDIA GPU 0: GeForce GT 330M (driver version unknown, CUDA version 5.5, compute capability 1.2, 1023MB, 600MB available, 182 GFLOPS peak)
Di 17 Dez 2013 13:31:09 CET | | OpenCL: NVIDIA GPU 0: GeForce GT 330M (driver version 319.37, device version OpenCL 1.0 CUDA, 1023MB, 600MB available, 182 GFLOPS peak)

If I attach to SETI@Home project in Boinc-Manger, does it use the CUDA-Seti@Home automatically or do I have to
compile the CUDA version by myself?

Thanks



The easy way: just try. Click, point and shoot.
Die einfachste: probieren.

The hard way: download the source code, compile, ask, recompile, ask, ...
Die meisst einfachst: selbst machen und fragen die Leuten hier danach ein möglicher Error kommst.

I compiled my own, many others have. We are all here. Just try.
Ich habe mein compiliert, wie so ein ganze menge andere mit ihren versionen gemacht haben. Die Hilfe ist hier.

I'd like think You do this for fun. So ...
Ich denke du habst Seti als ein Hobby. Es benötigt nur Zeit ...

Go on and give it a try. I have this quite old Fedora 14 linux and I had to install a zillion of things and compile a new compiler but I got it working.
The source code is fine.

--
It's been some 25+ years since I learned German.
Es ist mehr als 25 Jahren seit dem Ich Deutsch gelernt habe.

I had to try.
Ja, nur drei Jahre es gelest.
--


To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1454880 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1454881 - Posted: 17 Dec 2013, 20:32:12 UTC - in response to Message 1454871.  

Thanks Ivan, I knew You'd come to help. I got all started with your advice.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1454881 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1454886 - Posted: 17 Dec 2013, 20:39:12 UTC

I been using MB v7 on my Ubuntu 12.04 with no problem, Lunatics_x41g_linux64_cuda32.7z can be downloaded here:

http://www.arkayn.us/forum/index.php?action=tpmod;dl=cat5;dlsort=id;desc;p=10
ID: 1454886 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1455835 - Posted: 20 Dec 2013, 20:58:57 UTC

sorry for late reply. last days were a little busy.

tried to compile seti@home by myself. but was not successful :-(

Did the following steps:

- Downloaded the source code with "svn checkout https://setisvn.ssl.berkeley.edu/svn/seti_boinc" (described in http://setiathome.berkeley.edu/sah_porting.php)

- in directory seti_boinc I executed ./_autosetup

- export BOINCDIR=/usr/bin (otherwise ./configure complains)
- ./configure (for beginning without any options)
- make

Make fails almost just after it started:
make
make all-recursive
make[1]: Entering directory '/home/user/temp/setiathome/seti_boinc'
Making all in client
make[2]: Entering directory '/home/user/temp/setiathome/seti_boinc/client'
if g++ --include ../sah_config.h -DHAVE_CONFIG_H -I. -I.. -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I/usr/local/include -I/usr/local/include -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/ -I//api -I//lib -I//sched -I//db -pthread -msse2 -mfpmath=sse -DUSE_SSE -DUSE_SSE2 -I../db -I.. -I/ -I//api -I//lib -I//sched -I//db -pthread -O2 -Wall -MT seti_boinc-main.o -MD -MP -MF ".deps/seti_boinc-main.Tpo" -c -o seti_boinc-main.o main.cpp; then mv -f ".deps/seti_boinc-main.Tpo" ".deps/seti_boinc-main.Po"; else rm -f ".deps/seti_boinc-main.Tpo"; exit 1; fi
In file included from <command-line>:0:0:
./../sah_config.h:665:23: fatal error: std_fixes.h: No such file or directory
#include "std_fixes.h"
^
compilation terminated.
Makefile:1680: recipe for target 'seti_boinc-main.o' failed
make[2]: *** [seti_boinc-main.o] Error 1
make[2]: Leaving directory '/home/user/temp/setiathome/seti_boinc/client'
Makefile:520: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/home/user/temp/setiathome/seti_boinc'
Makefile:443: recipe for target 'all' failed




on http://pastebin.com/Lpykpd1C you can see the steps I did and the results of each step.

Maybe somebody could give me some hints how to successfully compile setiathome.





ID: 1455835 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1455840 - Posted: 20 Dec 2013, 21:25:52 UTC - in response to Message 1455835.  
Last modified: 20 Dec 2013, 21:36:46 UTC

I've got that file in two places.

/home/petri/sah_v7_opt/src/boinc
and
/home/petri/boinc_repo/lib

I think the second one is used. To build an application you'll need BOINC source too. Try to add your BOINC directory to the include path too. You may need to download and compile BOINC first.


I used the configure_cmdline.txt files as a template for my configure command. The was a lot of options in them.

Something like: (for setiathome v7)
./configure BOINCDIR=/home/petri/boinc_repo --enable-sse3 CFLAGS='-O3 -march=core2 -mtune=core2 -msse2avx -mavx -mpreferred-stack-boundary=8 -fexceptions -fno-rounding-math -fno-signaling-nans -fcx-limited-range -fno-math-errno -fno-trapping-math --param inline-unit-growth=3000 -DNDEBUG -DHAVE_STRCASECMP -fpeel-loops -funroll-loops -fgcse-sm -fgcse-las -fweb -I/usr/local/cuda-5.5/include -L/opt/lib-4.12/lib ' LIBS="/opt/lib-4.12/lib/libm.so.6 /opt/lib-4.12/lib/libc.so /opt/lib-4.12/lib/libpthread.so /usr/lib64/libstdc++.so /usr/lib64/lib/libm.so.6"

and

for Astropulse:
./configure --enable-bitness=64 --build=x86_64-pc-linux-gnu --target=x86_64-pc-linux-gnu --with-boinc-platform=x86_64-pc-linux-gnu --enable-static --enable-static-client --enable-avx --disable-shared --disable-graphics --enable-intrinsics CXXFLAGS=" -O3 -march=core2 -mtune=core2 -mfpmath=sse -mavx --param inline-unit-growth=3000 -I/root/NVIDIA_GPU_Computing_SDK/OpenCL/common/inc -I/root/NVIDIA_GPU_Computing_SDK/shared/inc" CPPFLAGS=" -DUSE_FFTW -DUSE_CONVERSION_OPT -DUSE_INCREASED_PRECISION -DSMALL_CHIRP_TABLE -DUSE_OPENCL -DUSE_AVX -DUSE_OPENCL_NV -DOPENCL_WRITE -DCOMBINED_DECHIRP_KERNEL -DOCL_ZERO_COPY -DAP_CLIENT" LIBS=" -L/usr/lib64 -lOpenCL -L/opt/lib-4.12/lib" LDFLAGS=" -static-libgcc -static-libstdc++" BOINCDIR=" /home/petri/boinc_repo" SETI_BOINC_DIR=" ../../AKv8" LIBS="/opt/lib-4.12/lib/libm.so.6 /opt/lib-4.12/lib/libc.so /opt/lib-4.12/lib/libpthread.so /usr/lib64/libstdc++.so /usr/lib64/lib/libm.so.6 "

My source is from here
svn checkout https://setisvn.ssl.berkeley.edu/svn/seti_boinc
a few months ago. Im running linux (fc14) and CUDA.

The config file templates are there too.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1455840 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 1455957 - Posted: 21 Dec 2013, 5:45:32 UTC - in response to Message 1455835.  
Last modified: 21 Dec 2013, 5:45:48 UTC



- Downloaded the source code with "svn checkout https://setisvn.ssl.berkeley.edu/svn/seti_boinc" (described in http://setiathome.berkeley.edu/sah_porting.php)


If you trying to build GPU SETI app of any kind (OpenCL or CUDA) then this info is obsolete. Right branch for GPU builds is sah_v7_opt .
You ckecked out only CPU version of MultiBeam app, even w/o CPU AstroPulse.

And yes, it would be nice if SETI web pages info was updated more frequently...
SETI apps news
We're not gonna fight them. We're gonna transcend them.
ID: 1455957 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1456121 - Posted: 21 Dec 2013, 16:07:56 UTC - in response to Message 1455957.  
Last modified: 21 Dec 2013, 16:11:29 UTC



- Downloaded the source code with "svn checkout https://setisvn.ssl.berkeley.edu/svn/seti_boinc" (described in http://setiathome.berkeley.edu/sah_porting.php)


If you trying to build GPU SETI app of any kind (OpenCL or CUDA) then this info is obsolete. Right branch for GPU builds is sah_v7_opt .
You ckecked out only CPU version of MultiBeam app, even w/o CPU AstroPulse.

And yes, it would be nice if SETI web pages info was updated more frequently...


Thanks for correcting. I got that info from the pages since i did not remember where i got the right source code to my computer. sah_v7_opt for current cuda source.

This is the web page https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1456121 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1456148 - Posted: 21 Dec 2013, 17:28:13 UTC

Thanks for claryifing.
Will try petri33's in the next days and provide feedback.


Raistmer wrote:

MAD__Mane wrote:

- Downloaded the source code with "svn checkout https://setisvn.ssl.berkeley.edu/svn/seti_boinc" (described in http://setiathome.berkeley.edu/sah_porting.php)


If you trying to build GPU SETI app of any kind (OpenCL or CUDA) then this info is obsolete. Right branch for GPU builds is sah_v7_opt .
You ckecked out only CPU version of MultiBeam app, even w/o CPU AstroPulse.


I knew this was CPU only. Because build of GPU version didn't work I tried if it will work with "normal" version to eliminate the possibility that the failure was due to wrong CUDA installation or something like that.

But thanks for the hint to sah_v7_opt branch.


ID: 1456148 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1460087 - Posted: 2 Jan 2014, 20:48:48 UTC

as I started trying to build seti@home by myself I didn't expect that it can get this annoying :-(
in the meantime I was able to build a seti@home executable. I think the build was successful when I can execute the binary with options -version or -help:


$ ./setiathome-7.19.x86_64-pc-linux-gnu -version
SETI@home client.
Version: 7.19

SETI@home is sponsored by individual donors around the world.
If you'd like to contribute to the project,
please visit the SETI@home web site at
http://setiathome.ssl.berkeley.edu.
The project is also sponsored by the Planetary Society,
the University of California, Sun Microsystems, Paramount Pictures,
Fujifilm Computer Products, Informix, Engineering Design Team Inc,
The Santa Cruz Operation (SCO), Intel, Quantum Corporation,
and the SETI Institute.

SETI@home was developed by David Gedye (Founder),
David Anderson (Director), Dan Werthimer (Chief Scientist),
Hiram Clawson, Jeff Cobb, Charlie Fenton,
Eric Heien, Eric Korpela, Matt Lebofsky,
Tetsuji 'Maverick' Rai and Rom Walton


./setiathome-7.19.x86_64-pc-linux-gnu -help
bad arg: -help
options:
 -nographics run without graphics
 -notranspose do not transpose data for pulse and gaussian searches
 -default_functions  use the safe unoptimized default functions
 -standalone (implies -nographics)
 -version  show version info
 -verbose  print running status



But when I try it to use like described in http://boinc.berkeley.edu/wiki/Anonymous_platform then I only see this in the logs:
02-Jan-2014 21:42:06 [SETI@home] Sending scheduler request: To fetch work.
02-Jan-2014 21:42:06 [SETI@home] Requesting new tasks for CPU                                                                    
02-Jan-2014 21:42:08 [SETI@home] Scheduler request completed: got 0 new tasks                                                    
02-Jan-2014 21:42:08 [SETI@home] No tasks sent                                                                                   
02-Jan-2014 21:42:08 [SETI@home] No tasks are available for SETI@home Enhanced                                                   
02-Jan-2014 21:42:08 [SETI@home] No tasks are available for SETI@home v7                                                         
02-Jan-2014 21:42:08 [SETI@home] No tasks are available for AstroPulse v6                                                        
02-Jan-2014 21:42:08 [SETI@home] Your app_info.xml file doesn't have a usable version of SETI@home v7.        



Maybe somebody could give me a hint for my app_info.xml.
(caution: for my first tries I built a CPU version of Seti@Home, not CUDA)

<app_info>
<app>
    <name>setiathome_v7</name>
</app>
<file_info>
    <name>setiathome-7.19.x86_64-pc-linux-gnu</name>
    <executable/>
</file_info>
<file_info>
    <name>seti_graphics</name>
    <executable/>
</file_info>
<app_version>
    <app_name>setiathome_v7</app_name>
    <version_num>719</version_num>
    <api_version>7.2.33</api_version>
    [
      <coproc>
          <type>cpu</type>
          <count>1</count>
      </coproc>
    ]
    [ <plan_class>cpu</plan_class> ]
    [ <flops>10000</flops> ]
    [ <avg_ncpus>1</avg_ncpus> ]
    [ <max_ncpus>1</max_ncpus> ]
    [ <cmdline>args</cmdline> ]
    <file_ref>
        <file_name>setiathome-7.19.x86_64-pc-linux-gnu</file_name>
        <main_program/>
    </file_ref>
    <file_ref>
        <file_name>seti_graphics</file_name>
        <open_name>graphics_app</open_name>
    </file_ref>
</app_version>
</app_info>



Thanks in advance.
ID: 1460087 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1460091 - Posted: 2 Jan 2014, 21:01:58 UTC - in response to Message 1460087.  

Maybe somebody could give me a hint for my app_info.xml.

Square brackets are a programming manual convention to signify an optional block or an optional item.

If you're not using that feature, delete the whole thing (square brackets and contents)
If you want to use the feature, remove the square brackets and tweak the contents as needed.

The documentation is written as an example only.
ID: 1460091 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1460096 - Posted: 2 Jan 2014, 21:06:47 UTC - in response to Message 1460091.  

ok, my app_info.xml now looks this:
<app_info>
<app>
    <name>setiathome_v7</name>
</app>
<file_info>
    <name>setiathome-7.19.x86_64-pc-linux-gnu</name>
    <executable/>
</file_info>
<file_info>
    <name>seti_graphics</name>
    <executable/>
</file_info>
<app_version>
    <app_name>setiathome_v7</app_name>
    <version_num>719</version_num>
    <api_version>7.2.33</api_version>
    <plan_class>cpu</plan_class>
    <flops>10000</flops>
    <avg_ncpus>1</avg_ncpus>
    <max_ncpus>1</max_ncpus>
    <file_ref>
        <file_name>setiathome-7.19.x86_64-pc-linux-gnu</file_name>
        <main_program/>
    </file_ref>
    <file_ref>
        <file_name>seti_graphics</file_name>
        <open_name>graphics_app</open_name>
    </file_ref>
</app_version>
</app_info>


but error messages stays the same :-/
02-Jan-2014 22:04:33 [SETI@home] update requested by user
02-Jan-2014 22:04:38 [SETI@home] Sending scheduler request: Requested by user.
02-Jan-2014 22:04:38 [SETI@home] Requesting new tasks for CPU
02-Jan-2014 22:04:40 [SETI@home] Scheduler request completed: got 0 new tasks
02-Jan-2014 22:04:40 [SETI@home] No tasks sent
02-Jan-2014 22:04:40 [SETI@home] No tasks are available for SETI@home Enhanced
02-Jan-2014 22:04:40 [SETI@home] No tasks are available for SETI@home v7
02-Jan-2014 22:04:40 [SETI@home] No tasks are available for AstroPulse v6
02-Jan-2014 22:04:40 [SETI@home] Your app_info.xml file doesn't have a usable version of SETI@home v7.

ID: 1460096 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1460097 - Posted: 2 Jan 2014, 21:09:00 UTC - in response to Message 1460096.  
Last modified: 2 Jan 2014, 21:13:22 UTC

Start by dropping these lines & see what happens:

<api_version>7.2.33</api_version>
<plan_class>cpu</plan_class>
<flops>10000</flops>
<avg_ncpus>1</avg_ncpus>
<max_ncpus>1</max_ncpus>


[If still nonworking, can look for other issues more easily], and you may then you need a <platfornm></platform> tag as well, filled in with the correct platform.
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1460097 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1460099 - Posted: 2 Jan 2014, 21:13:26 UTC - in response to Message 1460096.  

Did you build and install a screensaver file called 'seti_graphics'?

If not, remove both the <file_info> and <file_ref> blocks for it. I very much doubt you'll ever need it, if you're heading towards CUDA crunching - most GPGPU programmers want to use the power of the GPU for science, not eye-candy.

And read all the event log, especially the BOINC start-up section. BOINC may report app_info errors long before you get to the stage of testing a project update.
ID: 1460099 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1460113 - Posted: 2 Jan 2014, 21:45:36 UTC - in response to Message 1460099.  

guys, thanks for your immediate help.
my app_info.xml now looks this:
<app_info>
<app>
    <name>setiathome_v7</name>
</app>
<file_info>
    <name>setiathome-7.19.x86_64-pc-linux-gnu</name>
    <executable/>
</file_info>
<app_version>
    <app_name>setiathome_v7</app_name>
    <version_num>719</version_num>
    <file_ref>
        <file_name>setiathome-7.19.x86_64-pc-linux-gnu</file_name>
        <main_program/>
    </file_ref>
</app_version>
</app_info>


now I can download work for Seti@Home...
ID: 1460113 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1460117 - Posted: 2 Jan 2014, 21:52:50 UTC - in response to Message 1460113.  
Last modified: 2 Jan 2014, 21:55:38 UTC

now I can download work for Seti@Home...


Turns out for the Cuda build I updated & simplified the build process and promptly forgot as well ;)

When you get to the point of a Cuda build, you'll likely want the Xbranch (under sah_v7_opt IIRC) revision just prior to my current Cuda6 fiddling around. If everything's [ i.e. Cuda toolkit, boinc, cuda sources,] in the right place you'll want to do the regular simple autosetup, configure, make process, using the configure line:
./configure BOINCDIR=/your/repo/path/for/boinc --enable-sse2 --enable-fast-math
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1460117 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1460858 - Posted: 5 Jan 2014, 13:43:48 UTC
Last modified: 5 Jan 2014, 13:44:14 UTC

today I tried to build the CUDA version of S@H.


# checkout sources
svn checkout https://setisvn.ssl.berkeley.edu/svn/branches/sah_v7_opt

cd sah_v7_opt/Xbranch

# configure
./configure BOINCDIR=/home/user/temp/setiathome/boinc_repo --enable-sse2 --enable-fast-math

# make
[.....]
g++ -DHAVE_CONFIG_H -I. -I..  -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math  -fPIC -DPIC -DHAVE_CONFIG_H  -I/usr/local/include -g -O2 -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/home/manfred/temp/setiathome/boinc_repo -I/home/manfred/temp/setiathome/boinc_repo/api -I/home/manfred/temp/setiathome/boinc_repo/lib -I/home/manfred/temp/setiathome/boinc_repo/sched -I/home/manfred/temp/setiathome/boinc_repo/db -pthread  -msse2 -mfpmath=sse -DUSE_SSE -DUSE_SSE2   -MT seti_cuda-x86_float4.o -MD -MP -MF .deps/seti_cuda-x86_float4.Tpo -c -o seti_cuda-x86_float4.o `test -f 'vector/x86_float4.cpp' || echo './'`vector/x86_float4.cpp
mv -f .deps/seti_cuda-x86_float4.Tpo .deps/seti_cuda-x86_float4.Po
g++ -DHAVE_CONFIG_H -I. -I..  -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math  -fPIC -DPIC -DHAVE_CONFIG_H  -I/usr/local/include -g -O2 -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/home/manfred/temp/setiathome/boinc_repo -I/home/manfred/temp/setiathome/boinc_repo/api -I/home/manfred/temp/setiathome/boinc_repo/lib -I/home/manfred/temp/setiathome/boinc_repo/sched -I/home/manfred/temp/setiathome/boinc_repo/db -pthread  -msse2 -mfpmath=sse -DUSE_SSE -DUSE_SSE2   -MT seti_cuda-hires_timer.o -MD -MP -MF .deps/seti_cuda-hires_timer.Tpo -c -o seti_cuda-hires_timer.o `test -f 'vector/hires_timer.cpp' || echo './'`vector/hires_timer.cpp
mv -f .deps/seti_cuda-hires_timer.Tpo .deps/seti_cuda-hires_timer.Po
g++ -DHAVE_CONFIG_H -I. -I..  -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math  -fPIC -DPIC -DHAVE_CONFIG_H  -I/usr/local/include -g -O2 -msse2 -mfpmath=sse -m64 -g -O2 -I/usr/local/cuda/include -ffast-math -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/home/manfred/temp/setiathome/boinc_repo -I/home/manfred/temp/setiathome/boinc_repo/api -I/home/manfred/temp/setiathome/boinc_repo/lib -I/home/manfred/temp/setiathome/boinc_repo/sched -I/home/manfred/temp/setiathome/boinc_repo/db -pthread  -msse2 -mfpmath=sse -DUSE_SSE -DUSE_SSE2   -MT seti_cuda-analyzeFuncs.o -MD -MP -MF .deps/seti_cuda-analyzeFuncs.Tpo -c -o seti_cuda-analyzeFuncs.o `test -f 'analyzeFuncs.cpp' || echo './'`analyzeFuncs.cpp
In file included from analyzeFuncs.cpp:73:0:
cuda/cudaAcc_utilities.h:4:37: warning: 'cudaAcc_SafeCall_No_Sync' initialized and declared 'extern' [enabled by default]
cuda/cudaAcc_utilities.h:4:37: error: 'cudaError_t' was not declared in this scope
cuda/cudaAcc_utilities.h:4:54: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:4:72: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:4:84: error: expected primary-expression before 'int'
cuda/cudaAcc_utilities.h:4:92: error: expression list treated as compound expression in initializer [-fpermissive]
cuda/cudaAcc_utilities.h:5:34: warning: 'cudaAcc_SafeCall_Sync' initialized and declared 'extern' [enabled by default]
cuda/cudaAcc_utilities.h:5:34: error: 'cudaError_t' was not declared in this scope
cuda/cudaAcc_utilities.h:5:51: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:5:69: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:5:81: error: expected primary-expression before 'int'
cuda/cudaAcc_utilities.h:5:89: error: expression list treated as compound expression in initializer [-fpermissive]
cuda/cudaAcc_utilities.h:6:43: warning: 'cudaAcc_SafeCall_No_Sync_wExit' initialized and declared 'extern' [enabled by default]
cuda/cudaAcc_utilities.h:6:43: error: 'cudaError_t' was not declared in this scope
cuda/cudaAcc_utilities.h:6:60: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:6:78: error: expected primary-expression before 'char'
cuda/cudaAcc_utilities.h:6:90: error: expected primary-expression before 'int'
cuda/cudaAcc_utilities.h:6:98: error: expression list treated as compound expression in initializer [-fpermissive]
cuda/cudaAcc_utilities.h:8:75: error: 'int cudaAcc_SafeCall_No_Sync(char*, char*, int)' redeclared as different kind of symbol
cuda/cudaAcc_utilities.h:4:12: error: previous declaration of 'int cudaAcc_SafeCall_No_Sync'
cuda/cudaAcc_utilities.h:9:72: error: 'int cudaAcc_SafeCall_Sync(char*, char*, int)' redeclared as different kind of symbol
cuda/cudaAcc_utilities.h:5:12: error: previous declaration of 'int cudaAcc_SafeCall_Sync'
analyzeFuncs.cpp:101:33: note: #pragma message: ----ooura----
analyzeFuncs.cpp: In function 'int seti_analyze(ANALYSIS_STATE&)':
analyzeFuncs.cpp:978:6: error: 'cudaAcc_SafeCall_No_Sync_wExit' cannot be used as a function
make[2]: *** [seti_cuda-analyzeFuncs.o] Error 1
make[2]: Leaving directory `/home/manfred/temp/setiathome/sah_v7_opt/Xbranch/client'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/manfred/temp/setiathome/sah_v7_opt/Xbranch'
make: *** [all] Error 2




Thanks in advance
ID: 1460858 · Report as offensive
Kamu

Send message
Joined: 19 Jan 02
Posts: 56
Credit: 11,009,499
RAC: 0
Finland
Message 1460863 - Posted: 5 Jan 2014, 14:07:55 UTC - in response to Message 1460858.  

Hi

try following, you only need Xbranch for cuda client and older revision:

cd /home/user/temp/setiathome
svn checkout -r1921 https://setisvn.ssl.berkeley.edu/svn/branches/sah_v7_opt/Xbranch
cd Xbranch
sh ./_autosetup
sh ./configure BOINCDIR=/home/user/temp/setiathome/boinc_repo --enable-sse2 --enable-fast-math
make
ID: 1460863 · Report as offensive
Profile MAD__Mane

Send message
Joined: 6 Jun 99
Posts: 9
Credit: 599,529
RAC: 0
Germany
Message 1460928 - Posted: 5 Jan 2014, 17:43:14 UTC

thanks, was able to build the CUDA version.
the first tasks are running at the moment :-)
ID: 1460928 · Report as offensive
1 · 2 · Next

Message boards : Number crunching : CUDA on Ubuntu 12.04 x86_64


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.