Re: Advice for offline image analysis computer

Posted by Hu Xian on
URL: http://confocal-microscopy-list.275.s1.nabble.com/Advice-for-offline-image-analysis-computer-tp1109084p1109390.html

Hi there,

Well, generally, I would prefer 64 bits over 32 bits operation system.

But the details of operation system greatly depends on the software you want to use.

For instance, we know the ImageJ works much better on Leopard than Vista 64 bits. If I'm not wrong, you can allocate around 1.5G memory for ImageJ on Vista before it starts to crash, whereas on leopard it can take upto 10Gb of memory.

And some of the software might not be stable on certain operation, velocity isn't very stable on vista 64 bits yet. Imaris is better in stability, but it does not supply multicore system yet(hence our 2.8G 8-core doesnot work nearly as well as 3.2G single core)

I'm not very familiar with matlab and metamorph.
But a computer with more ram and computing power will definitely helps accelerate rendering process a lot.

$5000 is a lot of money. Not sure about European market, but should be too far off. You should be able to get a 64bits operation system with 64GB, 8 core. Disc space is not that critical as long as you are not using the computer as data storage.

We were trying to purchase an image processing workstation around the same price from Dell, I can send you the quote for your reference(no commercial interest).

Thanks,

Edna, HU Xian

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Jean-Christophe Olaya
Sent: Monday, September 22, 2008 5:57 PM
To: [hidden email]
Subject: Re: Advice for offline image analysis computer

[hidden email] schrieb:

>> Dear listers,
>>
>> I have to buy a new computer for our team that will be used as for offline
>> image analysis. The software required (ImageJ, Metamorph, Matlab) require
>> to
>> build a Windows/Vista machine. The cost would be around $5,000 / 5000€.
>> The
>> job would be processing (a lot) and rendering. My questions are :
>> - Should I go for a "traditionnal" computer, I mean a 32-bits, dual core,
>> XP
>> computer with maxed RAM (I guess it is 3GB or so) ? Is it usefull to go to
>> more fancy stuff like 64 bits, 4 or 8 core, 8 to 16GB RAM machine ? Would
>> ImageJ/Metamorph/Matlab really benefit from it (I would love to have an
>> answer from the Metamorph/Matlab people) ?
>> - What about graphic cards and GPU ? What is the best choice ? I've heard
>> about new strategies to speed up processing by making the GPU churn data
>> as
>> well as graphics, but I don't think it is really commercially available
>> now
>> or implemented in commercial software yet.
>>
>>    
>
> Salut Christophe, je te fais une reponse off-liste histoire de causer
> frenchie. Le software que je connais le mieux est Matlab, voila ce que je
> peux t'en dire.
>
>
> Sous windows, le support par Matlab des architectures 64 bits est pour
> tres bientot (soit la fin de l'annee soit debut 2009 au pire). Il y a
> beaucoup de benefices a y trouver:
> - Matlab bosse presque uniquement avec des nombres 'double' qui prennent
> 64 bits. Avec une architecture 64 bit, chaque operation sur eux prend un
> tic et non pas deux comme maintenant.
> - Matlab est tres gourmand en memoire (il la gere assez mal), avec un
> system 64 bit tu peux te payer 16Go de RAM.
> - Certaines operations de Matlab prennent avantages du multi coeur
> nativement (multiplication de matrices etc...), mais autrement et pour la
> grande majorite de ce qui va t'interesser, un simple coeur suffit. Aller
> au dela de 4 coeurs est inutile dans ton cas je pense.
> - Il existe des version d'ImageJ qui travaille en 64bit. Dans mon groupe
> un type bosse dessus (va voir sur la page notre projet
> http://pacific.mpi-cbg.de/wiki/index.php/Main_Page, c'est encore le
> bordel, ne t'attend a rien de magique). C'est tres utile, parce que la
> aussi on a vite fait de bouffer toute la memoire.
>
>
> Je sais que Imaris et Volocity utilisent les GPU a fond, et en dependent
> meme pour la visualisation 3D/4D (acceleration materielle et tout et
> tout). Pour Matlab et Metamorph (ce dernier etant quand meme bien en
> retard..) j'en doute.
>
> J'espere que c'est utile
> A bientot
> jy
>
>
>
>
> --
> Jean-Yves Tinevez, PhD
> Image Processing Facility head
> Max Planck Institute of Molecular Cell Biology and Genetics
> Pfotenhauerstr. 108, 01307 Dresden
> Germany
> tel [phone]
> fax [phone]
> e-mail: [hidden email]
>  

...pas si "horts liste", ta réponse ;-)

JC


--
Jean-Christophe Olaya
Biotechnologisches Zentrum der TU Dresden
Light Microscopy Facility
Tatzberg 47-51
01307 Dresden
Germany
email: [hidden email]
Tel: [phone]
Fax: [phone]