g+
g+ Communities
Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
<== Date ==> <== Thread ==>

Subject: RE: gpu integration into epics
From: <james.rowland@diamond.ac.uk>
To: <rodrigo.castro@ciemat.es>, <tech-talk@aps.anl.gov>
Date: Wed, 17 Oct 2012 12:48:18 +0000
Dear Rodrigo

Do you have any measurements of latency from the CPU to the GPU? I am interested in using GPUs for feedback control with ethernet as the transport layer, and a latency budget of 100us for the feedback cycle.

James

________________________________________
From: tech-talk-bounces@aps.anl.gov [tech-talk-bounces@aps.anl.gov] on behalf of Rodrigo Castro [rodrigo.castro@ciemat.es]
Sent: 16 October 2012 11:55
To: EPICS Tech
Subject: Re: gpu integration into epics

Hello Sean,

First of all, I am sorry for the delay in the answer but I just received a reference to this question (through an email from a colleague) and I wasn't subscribed. Now, I am.

About the work, I am one of the authors. I am working for the Spanish National Fusion Laboratory in Madrid and I have been working with some colleagues implementing EPICS based Fast Control technology for ITER.

As we showed in the presentation, we have done all the development based on asynDrivers. In fact, we have developed an infrastructure based on asynDriver modules, which are instantiated at IOC start time, and a set of configurable links with user defined data structures which are also configurable at start time. The result has been a modular system which has been able to ingrate two (but can be more) GPUs, with a steady state internal hight data rate movement (256 MBytes/sec - 32 acquired channels to 1 Msamples/sec), and with total deviation times about 1% (soft real-time system).

Currently we have modules for:
        - FPGA (FlexRIO) data acquisition
        - GPUs
        - Waveform injection
        - Remote data streaming for archiving
        - INPUT and OUTPUT shared memory connectors (for connecting with other IOCs or processes)
        - Monitoring
        - Data streaming (for archiving)

Additionally, the system supports input data synchronization (for processing).

Based on achieved good results, and the modularity of the solution, we are extending its applicability to soft real-time systems demanded on other fusion machines.

Sorry for the extension of the message but I tried to give a quick view of the EPICS technology involved.

Best regards,
Rodrigo

--
Dr. Rodrigo Castro

e-mail:         rodrigo.castro@ciemat.es

Tel: (+34) 91 3466419
Fax: (+34) 913466124

CIEMAT (National Research Agency for Energy and Environment)
Avda. Complutense, 22
28040 Madrid
----------------------------------------------------

> Hi
>
> I came across a talk from the "7th Workshop on Fusion Data Processing Validation and Analysis" on the use of
> gpu's with epics.
> http://www.fusione.enea.it/EVENTS/eventifiles/validation7-2012/workshop-slides/1stday/GPUs.pptx
>
>
> Does anybody have details or knowledge of the progress of gpu integration into epics? This was from March 2012, and given the title of the conf, obviously
> related to ITER.
>
> I assume at least 1 of the 6 authors is on this list maybe ?
>
> Thanks
> Sean


-- 
This e-mail and any attachments may contain confidential, copyright and or privileged material, and are for the use of the intended addressee only. If you are not the intended addressee or an authorised recipient of the addressee please notify us of receipt by returning the e-mail and do not use, copy, retain, distribute or disclose the information in or attached to the e-mail.
Any opinions expressed within this e-mail are those of the individual and not necessarily of Diamond Light Source Ltd. 
Diamond Light Source Ltd. cannot guarantee that this e-mail or any attachments are free from viruses and we cannot accept liability for any damage which you may sustain as a result of software viruses which may be transmitted in or with the message.
Diamond Light Source Limited (company no. 4375679). Registered in England and Wales with its registered office at Diamond House, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0DE, United Kingdom
 





Replies:
RE: gpu integration into epics Dalesio, Leo
References:
Re: gpu integration into epics Rodrigo Castro

Navigate by Date:
Prev: Re: Immediately reprocess a record [SEC=UNCLASSIFIED] Tim Mooney
Next: RE: gpu integration into epics Dalesio, Leo
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
Navigate by Thread:
Prev: Re: gpu integration into epics Rodrigo Castro
Next: RE: gpu integration into epics Dalesio, Leo
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
ANJ, 18 Nov 2013 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· EPICSv4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·