PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

How to get CUDA fortran beta version installed

 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling
View previous topic :: View next topic  
Author Message
tty103



Joined: 19 Oct 2009
Posts: 8

PostPosted: Thu Oct 22, 2009 10:05 am    Post subject: How to get CUDA fortran beta version installed Reply with quote

I downloaded both the pvf and workstation packages (9.04 updated Sep. 18). during install process, seems no indication cuda fortran is included. I tried compile a test code, it complains about all the CUDA fortran directives. I tried both use a .cuf extension, and use -Mcuda option. Is it a guide on how to setup the cuda fortran beta version? Thanks
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6129
Location: The Portland Group Inc.

PostPosted: Thu Oct 22, 2009 1:47 pm    Post subject: Reply with quote

Hi tty103,

The CUDA Fortran Beta is only available on Linux. Windows will be supported in the next release (10.0).

Thanks for your interest,
Mat
Back to top
View user's profile
ghandurah



Joined: 03 Nov 2009
Posts: 9

PostPosted: Sun Nov 22, 2009 10:23 pm    Post subject: Reply with quote

@mkcolg

can this -Mcuda option be used for C code compiled with -ta=nvidia? or is it for CUDA fortran only?

and is there anyway that I can view the CUDA code generated by the PGI compiler and modify it?

Thank You


EDIT: I found out that I could use -ta=nvidia,keepgpu (for the 2009 relase) to view the GPU code, but this is not mentioned in the PGI user guide 2009, only in the 2010 guide and release notes!!

can I modify that code now and recompile using it?
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6129
Location: The Portland Group Inc.

PostPosted: Mon Nov 23, 2009 10:26 am    Post subject: Reply with quote

Hi ghandurah,

Quote:
can this -Mcuda option be used for C code compiled with -ta=nvidia? or is it for CUDA fortran only?

"-Mcuda" is for CUDA Fortran only. For CUDA C, you will need to use NVIDIA's nvcc.

"-ta=nvidia" can be used with either C or Fortran but only with PGI's Accelerator directive/pragma model. The accelerator model is distinct from CUDA Fortran.

Please see my response to your other post on how to use the "keepgpu" flag.
http://www.pgroup.com/userforum/viewtopic.php?t=1653

- 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
Page 1 of 1

 
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