Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ptp-dev] CDT error parsers for Cray, PGI compilers

I think I agree with Doug (include them with PTP) partially because it seems the simplest from our point of view.

We had similar questions about some UPC (Unified Parallel C) tools in the past.
The parsers went into CDT (Mike Kucera, the author, was a CDT committer)
The PLDT assistance tools went into PTP (I'm the author and a PTP committer)
The Berkeley UPC toolchain went into CDT (Max Billingsley, the author, was a committer to neither project and we discussed it; he chose CDT for the toolchain but also contributed performance tools eclipse integration for  "PPW" to PTP)

So yes, our projects are closely related, Doug. Synchronized projects (code mirrored on local and remote systems) are not specific to HPC projects.  They are starting to be the tipping point to convincing entrenched HPC groups to adopt HPC.  Greg has proposed to describe them in his EclipseCon proposal [1]l and Wyatt Spear and I have proposed to discuss how the HPC/scientific community uses Eclipse [2].
Remote tools are also being investigated by the LInuxTools project - in fact they have a talk proposal too [3].  They have used some of the parts of PTP as I've seen in their mailing list.
It would be great to get us all together at EclipseCon.   (Hope our talks get accepted so we have a hope of getting travel funding!)

[1] http://www.eclipsecon.org/2012/sessions/automatic-remote-project-synchronization-using-git
[2] http://www.eclipsecon.org/2012/sessions/building-platform-computational-science
[3] http://www.eclipsecon.org/2012/sessions/remote-linux-tooling

...Beth

Beth Tibbitts
Eclipse Parallel Tools Platform  http://eclipse.org/ptp
IBM STG - High Performance Computing Tools
Mailing Address:  IBM Corp., 745 West New Circle Road, Lexington, KY 40511


Inactive hide details for "Schaefer, Doug" ---12/01/2011 03:24:30 PM---My quick thought would be to go with #3. As you mention "Schaefer, Doug" ---12/01/2011 03:24:30 PM---My quick thought would be to go with #3. As you mention these are more likely for the HPC crowd that


    From:

"Schaefer, Doug" <Doug.Schaefer@xxxxxxxxxxxxx>

    To:

Parallel Tools Platform general developers <ptp-dev@xxxxxxxxxxx>,

    Date:

12/01/2011 03:24 PM

    Subject:

Re: [ptp-dev] CDT error parsers for Cray, PGI compilers

    Sent by:

ptp-dev-bounces@xxxxxxxxxxx




My quick thought would be to go with #3. As you mention these are more likely for the HPC crowd that has clearly rallied around PTP. CDT has mainly been focused on host and embedded development.
 
But it is a great question and a sign at how closely our projects are related (and why I follow this list and the things y’all do J). A lot of the problems that the CDT community is starting to face are very similar to the HPC world. I’d even say the worlds are converging. Working together is going to continue to become more and more important.
 
Cheers,
Doug.
 
 
From: ptp-dev-bounces@xxxxxxxxxxx [mailto:ptp-dev-bounces@xxxxxxxxxxx] On Behalf Of Jeffrey Overbey
Sent:
 Thursday, December 01, 2011 3:16 PM
To:
 ptp-dev
Subject:
 [ptp-dev] CDT error parsers for Cray, PGI compilers
 
This is probably a question for Doug and Greg.
 
I wrote error parsers for the C/C++ and Fortran compilers from Cray and the Portland Group.  For the time being, a single error parser recognizes errors from both the C/C++ and Fortran compilers since the error message formats are almost identical, although I could easily split them into separate C/C++ and Fortran error parsers if necessary.
 
Obviously, it makes sense to include the Fortran compiler error parsers in Photran.  I'm not sure what to do with the C/C++ error parsers.  I think there are 3 options.
 
    1. Put the combined C/C++-and-Fortran error parsers into Photran.  I could name them, e.g., "Photran Error Parser for Cray Compilers."  However, having to install Photran to get an error parser for your Cray C/C++ compiler is weird.
    2. I could contribute the C/C++ error parsers to CDT.  I'm not sure CDT would want them, though... AFAIK, the base CDT installation only contains error parsers for the GNU toolchains and Visual C++.
    3. The C/C++ error parsers could be provided by PTP, since AFAIK those compilers are mainly used in HPC.
 
What are your thoughts?
 
Jeff_______________________________________________
ptp-dev mailing list
ptp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-dev


GIF image

GIF image


Back to the top