PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

error for a simple OPENACC program
Goto page Previous  1, 2, 3, 4, 5  Next
 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling
View previous topic :: View next topic  
Author Message
mkcolg



Joined: 30 Jun 2004
Posts: 6119
Location: The Portland Group Inc.

PostPosted: Fri May 10, 2013 8:45 am    Post subject: Reply with quote

Quote:
Can i install it in a 32 bits ubuntu version ?
Yes, you can install the compilers on a 32-bit OS. I just don't have a GPU attached to a system with pure 32-bit OS and this may or may not account for the difference in what we're each seeing. Let me talk with my IT folks and see if I can get 32-bit Ubuntu installed here next week.

Quote:
Do you want to i install the 64 bits version of the compiler to test ?
If you can, that would be great.

Just curious why did you choose a 32-bit OS?

- Mat
Back to top
View user's profile
alechand



Joined: 14 May 2012
Posts: 21

PostPosted: Fri May 10, 2013 10:12 am    Post subject: Reply with quote

Hello.

I use 32 bits because i found lots of problems with 64 bits.
You are not reading my messages very well, i was asking if i can
use the 64 bits version of the compiler in the 32 bits kubuntu version, because you asked me to use that -m64 parameter in the compilation, even if i have a 32 bits system.

I am thinking ... is this compiler really easier to use than use directly CUDA ?
I cannot even put this to work ...
I am very disapointed
Whatever, i will try to install all again.

Thanks !
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6119
Location: The Portland Group Inc.

PostPosted: Fri May 10, 2013 11:15 am    Post subject: Reply with quote

Quote:
You are not reading my messages very well, i was asking if i can
use the 64 bits version of the compiler in the 32 bits kubuntu version, because you asked me to use that -m64 parameter in the compilation, even if i have a 32 bits system.
No, it's because I didn't know you were use a 32-bit OS. Granted, I should not have assumed that you're running a 64-bit OS, but given your only the second 32-bit OS user I've encountered in many years, hopefully its a forgiveable assumption.
Quote:

I am thinking ... is this compiler really easier to use than use directly CUDA ?
Yes, OpenACC is easier to program than CUDA. However, massively parallel programming is still difficult, no matter which programming paradigm you use.

Quote:
I cannot even put this to work ...
Again, we haven't established that is a problem with the PGI compilers. It's possible, but since you have a very unusual configuration it's hard to know. I'd still like you to try running the NVIDIA CUDA SDK examples. If you can get those to run, then it's a PGI configuration error. If you can't, then it's an NVIDIA problem or a problem with your system.

- Mat
Back to top
View user's profile
alechand



Joined: 14 May 2012
Posts: 21

PostPosted: Tue May 14, 2013 8:07 pm    Post subject: Reply with quote

Hello Mat.

Just to inform you what is happening now...
I followed your suggestion and installed a 64 bits version of ubuntu 11.10 and all upgrades available.
Then i installed the last nvidia 64 bits driver for my GTX680 graphic card.
I installed the pgi compiler 12.10, also for 64 btis, with all options avaiable ( including the cuda tool kit ). I obtained the same sequence of errors as before (see below)

############################

alechand@pcsantos2:~/test$ more picalc.f90
program picalc
implicit none
integer, parameter :: n=1000000
integer :: i
real(kind=8) :: t, pi
pi = 0.0
!$acc parallel loop
do i=0, n-1
t = (i+0.5_8)/n
pi = pi + 4.0/(1.0 + t*t)
end do
!$acc end parallel loop
print *, 'pi=', pi/n
end program picalc

alechand@pcsantos2:~/test$ pgf90 -ta=nvidia -Minfo=all -o PI picalc.f90
picalc:
7, Accelerator kernel generated
7, CC 1.3 : 23 registers; 32 shared, 36 constant, 0 local memory bytes
CC 2.0 : 23 registers; 0 shared, 60 constant, 0 local memory bytes
8, !$acc loop gang, vector(256) ! blockidx%x threadidx%x
10, Sum reduction generated for pi
7, Generating compute capability 1.3 binary
Generating compute capability 2.0 binary
alechand@pcsantos2:~/test$ export PGI_ACC_DEBUG=1
alechand@pcsantos2:~/test$ ./PI
__pgi_cu_init() found 1 devices
__pgi_cu_init( file=/home/alechand/test/picalc.f90, function=picalc, line=7, startline=1, endline=14 )
__pgi_cu_init() will use device 0 (V3.0)
__pgi_cu_init() compute context created
__pgi_cu_module3( lineno=7 )
__pgi_cu_module3 module loaded at 0x8d8390
__pgi_cu_module_function( name=0x6735f2=picalc_7_gpu, lineno=7, argname=(nil)=, argsize=12, varname=0x6735ff=b1, varsize=8, SWcachesize=0 )
Function handle is 0xabbd50
__pgi_cu_module_function( name=0x6735e0=picalc_10_gpu_red, lineno=7, argname=(nil)=, argsize=0, varname=(nil)=, varsize=0, SWcachesize=0 )
Function handle is 0x9234c0
__pgi_cu_alloc(size=31256,lineno=7,name=)
__pgi_cu_alloc(31256) returns 0x400240000
__pgi_cu_uploadc( "b1", size=8, offset=0, lineno=7 )
constant data b1 at address 0x400140000 devsize=8, size=8, offset=0
First arguments are:
0 0

0x00000000 0x00000000
__pgi_cu_launch_a(func=0xabbd50, grid=3907x1x1, block=256x1x1, lineno=7)
__pgi_cu_launch_a(func=0xabbd50, params=0x7fff13e3d5fc, bytes=8, sharedbytes=2048)
First arguments are:
2359296 4

0x00240000 0x00000004
__pgi_cu_launch_a(func=0x9234c0, grid=1x1x1, block=256x1x1, lineno=10)
__pgi_cu_launch_a(func=0x9234c0, params=0x7fff13e3d5fc, bytes=12, sharedbytes=2048)
First arguments are:
2359296 4 3907

0x00240000 0x00000004 0x00000f43
__pgi_cu_downloadc( "b1", size=8, offset=0, lineno=7 )
call to cuMemcpyDtoH returned error 700: Launch failed
CUDA driver version: 5050

#################################################

I think my graphic card is with problems...
What do you think ?
Do you know some way i can test my card, to see if all is ok with it ?
Thanks a lot.
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6119
Location: The Portland Group Inc.

PostPosted: Wed May 15, 2013 11:46 am    Post subject: Reply with quote

Quote:
I followed your suggestion and installed a 64 bits version of ubuntu 11.10 and all upgrades available.
Thanks. At least we know it's not the OS.

Quote:
Do you know some way i can test my card, to see if all is ok with it ?
Please run a few of the NVIDIA CUDA C SDK examples. https://developer.nvidia.com/cuda-downloads

- Mat
Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling All times are GMT - 7 Hours
Goto page Previous  1, 2, 3, 4, 5  Next
Page 4 of 5

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © phpBB Group