Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Folding at home (http://folding.stanford.edu/) is a drisributed computing system that is used to study protein folding. Each user signs up and installs a client on their PC. Then when their PC is inactive the client crunchs through downloaded datasets. They have just released a client that now uses the extra power of the graphics card to crunch the data.
Cheers
Cam
________________________________
From: Confocal Microscopy List on behalf of Guy Cox
Sent: Sat 16/08/2008 5:09 AM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
OK, maybe I'm iggerant, but what on earth is 'folding at home'?
Guy
Optical Imaging Techniques in Cell Biology
by Guy Cox CRC Press / Taylor & Francis
http://www.guycox.com/optical.htm
______________________________________________
Associate Professor Guy Cox, MA, DPhil(Oxon)
Electron Microscope Unit, Madsen Building F09,
University of Sydney, NSW 2006
______________________________________________
Phone +61 2 9351 3176 Fax +61 2 9351 7682
Mobile 0413 281 861
______________________________________________
http://www.guycox.net <http://www.guycox.net/>
________________________________
From: Confocal Microscopy List [mailto:[log in to unmask]] On Behalf Of aki hoji
Sent: Saturday, 16 August 2008 12:45 AM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal An alternative CUDA is OpenCL. Interestingly, OpenCL will be implemented in the next major upgrade of MacOS X, Snow Leopard, and how easily imaging apps can utilize OpenCL depends on how it is implemented. Hopefully, Snow Leopard has an OpenCL frame work which makes it so much easier for any apps to take advantage of GPGPU. There is one thing though. I wonder how hot my MBP will be when it is fully utilizing GPU.
Aki Hoji
University of Pittsburgh
On Aug 14, 2008, at 6:26 PM, Nowell, Cameron wrote:
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
You are spot on. The only software i know of that uses a GPU for analysis is the GPU client for the folding at home project. Now this raises a very interesting issue. Folding at home uses the Nvidia CUDA language to run applications on the GPU. The GPU has a much faster floating point unit so therefore can crunch raw data more efficently. For example when running the CPU version of the folding at home client on my dual core 2.8Ghz 12MB cache CPU i get around 100 iterations per second, but when i change over to the GPU version runnign through 2 512MB 8800GTX nvidia cards i get around 2000 iterations per second. So GPU are great for crunching data if the program is written for them. Sadly though there are no imaging apps yet that take advantage of this, in the future i am sure there will be.
Cheers
Cam
________________________________
From: Confocal Microscopy List [mailto:[log in to unmask]] On Behalf Of Craig Brideau
Sent: Friday, 15 August 2008 5:46 AM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
CPU speeds is something that could vary a lot even with the same application. It will depend on what you are doing. If you are just throwing around a pretty 3D model, the GPU will be doing the majority of the work. But if you start doing some analysis of that 3D model you will start to ramp up your CPU usage. The speed of your CPU will only affect how fast a set of data can be crunched, the amount of RAM in your machine will govern how big that data set can be. So if you are happy to wait a little, a low end CPU with lots of RAM would be a better combination.
I was thinking mainly of analysis when I made my earlier comment that the GPU might not be used as much. So in terms of displaying the data, the GPU is king, but in performing actual statistics or the like on an image it's mainly the CPU that's doing the work, correct? Are you aware of software that uses the GPU for the actual analysis?
Craig
I am sure you can hit a bottleneck (CPU or GPU becoming the limit) with analysis software if you give it enough data. But as long as you have a decent mid to high range CPU, and while i have been an AMD fan boy for years that CPU should be a dual core Intel chip (something from the 9000 series with large cache memory would be good). Couple that with a good GPU board and you will be fine. Oh i guess one other thing is make sure the mother board is using the newer Intel P35 chipset as well.
Really it comes down to how much money you have to spend and what you are trying to model/analyse.
Cheers
Cam
From: Confocal Microscopy List [mailto:[log in to unmask]] On Behalf Of Craig Brideau
Sent: Thursday, 14 August 2008 1:03 PM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
I wasn't aware of that the Olympus and Leica systems actually took advantage of GPUs. Any thoughts as to what sort of crunching they use them for?
Given the tremendous variety of video cards, how would one go about matching optimal card to optimal software? At what point do you have to worry about basic CPU speeds as well? (i.e. GPU vs CPU mismatch...?)
Craig
On Wed, Aug 13, 2008 at 8:59 PM, Michael C. Wussow <[log in to unmask]> wrote:
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hello,
I have to completely agree with Cam. His comments are spot on and
what we at Bitplane recommend to our customers. However, it is recommend
that you check with the company that you are getting the software from just
to make sure there are no incompatibilities between the software and the
graphics card.
If you want specific recommendations don't hesitate to contact us offline.
Mike
Bitplane Inc.
Michael C. Wussow
Vice President and General Manager Bitplane Inc.
Cell Phone: 651-336-4600
Fax: 866-691-9112
Toll Free: 1-888-3D-BITPX (332-4879)
Visit Our Web Site At: www.bitplane.com <http://www.bitplane.com/>
-----Original Message-----
From: Confocal Microscopy List [mailto:[log in to unmask]] On
Behalf Of Nowell, Cameron
Sent: Wednesday, August 13, 2008 3:45 PM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi List,
I owuld like to disagree with Craig. There are now quite a few
programs that use the graphics card to process 3D data in real time now.
MetaMorph, Imaris, Volocity, 3D Studio, Olympus FV1000, Leica LAS AF just to
name a few.
There is an enourmous ammount of choice out there in the graphics card
market, but my recomendation (and the recomendation of many other people) is
to buy the best GAMING card you can afford. While each graphics card
manuafacturer has "Workstation" class cards that cost a heap of money, these
cards oare not really any better than a good gaming card. So avoid things
like the Nvidia Quadro, ATI FirePro or Wildcat Studio type of cards.
So you are going to need something like an Nvidia Geforce 280GTX or an ATI
Radeon HD 4870. You will need at least 512MB of RAM on the card as well.
Also these cards cost in the order of $500-600 US dollars.
Now slapping a highend card in a lower end computer is not going to fix
things either. To match the card you will need to have a good CPU (dual coe
is fine, there are very, very few programs coded for more than 2 cores
currently) and the more system RAM the better. If you are playing with large
data sets (eg 2000 images per 3D stack) you will need to go beyond the
maximum 4GB of RAM in a 32 bit system and go for a 64Bit system and put 16
or 32 GB RAM into it.
Fast hardrives will always help a little bit but not too much.
Also don't worry about going for a two graphics cards running togther
(Nvidia SLI or ATI CrossFire), as far as i know there are no modeling
programs out there yet that can leaverage both cards.
Hope that helps you out, if you need any more help just let me know.
Cheers
Cam
Cameron Nowell
Research and Microscopy Imaging Core
Peter MacCallum Cancer Centre
Melbourne, Australia
________________________________
From: Confocal Microscopy List on behalf of Craig Brideau
Sent: Thu 14/08/2008 6:16 AM
To: [log in to unmask]
Subject: Re: graphic cards
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Actually most image display and manipulation software uses your CPU rather
than your GPU. The CPU does all the number crunching then hands the
resulting data off to the GPU (video card) to display it. There are a
handful of programs that will actually use the GPU instead, but they are few
and far between as they require special programming.
What sort of processor, video card, and how much RAM do you have in your
computer?
Craig
On Wed, Aug 13, 2008 at 2:04 PM, Marc Thibault <[log in to unmask]>
wrote:
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi all,
I would like to know what are the best graphic cards to add in your
PC for microscopic 3D image analysis and manipulation.
Mine is currently not up to the task.
Thanks
Marc
This email (including any attachments or links) may contain
confidential and/or legally privileged information and is
intended only to be read or used by the addressee. If you
are not the intended addressee, any use, distribution,
disclosure or copying of this email is strictly
prohibited.
Confidentiality and legal privilege attached to this email
(including any attachments) are not waived or lost by
reason of its mistaken delivery to you.
If you have received this email in error, please delete it
and notify us immediately by telephone or email. Peter
MacCallum Cancer Centre provides no guarantee that this
transmission is free of virus or that it has not been
intercepted or altered and will not be liable for any delay
in its receipt.
No virus found in this incoming message.
Checked by AVG - http://www.avg.com <http://www.avg.com/>
Version: 8.0.138 / Virus Database: 270.6.2/1609 - Release Date: 13/08/2008 6:43 AM
This email (including any attachments or links) may contain
confidential and/or legally privileged information and is
intended only to be read or used by the addressee. If you
are not the intended addressee, any use, distribution,
disclosure or copying of this email is strictly
prohibited.
Confidentiality and legal privilege attached to this email
(including any attachments) are not waived or lost by
reason of its mistaken delivery to you.
If you have received this email in error, please delete it
and notify us immediately by telephone or email. Peter
MacCallum Cancer Centre provides no guarantee that this
transmission is free of virus or that it has not been
intercepted or altered and will not be liable for any delay
in its receipt.
This email (including any attachments or links) may contain
confidential and/or legally privileged information and is
intended only to be read or used by the addressee. If you
are not the intended addressee, any use, distribution,
disclosure or copying of this email is strictly
prohibited.
Confidentiality and legal privilege attached to this email
(including any attachments) are not waived or lost by
reason of its mistaken delivery to you.
If you have received this email in error, please delete it
and notify us immediately by telephone or email. Peter
MacCallum Cancer Centre provides no guarantee that this
transmission is free of virus or that it has not been
intercepted or altered and will not be liable for any delay
in its receipt.
No virus found in this incoming message.
Checked by AVG.
Version: 7.5.526 / Virus Database: 270.6.3/1612 - Release Date: 14/08/2008 6:03 PM
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.526 / Virus Database: 270.6.3/1612 - Release Date: 14/08/2008 6:03 PM
This email (including any attachments or links) may contain
confidential and/or legally privileged information and is
intended only to be read or used by the addressee. If you
are not the intended addressee, any use, distribution,
disclosure or copying of this email is strictly
prohibited.
Confidentiality and legal privilege attached to this email
(including any attachments) are not waived or lost by
reason of its mistaken delivery to you.
If you have received this email in error, please delete it
and notify us immediately by telephone or email. Peter
MacCallum Cancer Centre provides no guarantee that this
transmission is free of virus or that it has not been
intercepted or altered and will not be liable for any delay
in its receipt.
|
