I can second this. We use SATA raid5 locally in our fMRI-server, continuously running without probs so far. -----Oorspronkelijk bericht----- Van: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]Namens hao zhang Verzonden: woensdag 30 maart 2005 9:24 Aan: [log in to unmask] Onderwerp: Re: [SPM] SATA vs SCSI and local HD vs NAS Agree with Russ. I used PC/SATA under Linux while ago, felt much faster than Sun/Blade workstation. Additional to adding RAID box to your workstation, attaching a external harddisk via USB2 or Firewire port should be also a convenient option, though I never really compared there two options. Regards, Zhang Hao On Tue, 29 Mar 2005 16:18:47 -0800, Russ Poldrack <[log in to unmask]> wrote: > Renat, > I think few people buy SCSI RAID systems any more because they are so > expensive, and the reliability and speed of SATA RAID is quite good. > we use SATA raid systems hung off of a dedicated server, fed over a > gigabit ethernet, and find that this gives us quite good performance. > Local hard drive would obviously be faster, but that's not a scalable > solution. > cheers > russ > > On Mar 29, 2005, at 4:04 PM, Renat Yakupov wrote: > > > Hello SPM users. > > This quite off topic, but I was hoping you could give your opinion or > > share > > your experiences with me. > > We are thinking about upgrading our processing workstation. And we are > > thinking of buying RAID5. The question is, if the CPU speed and RAM is > > not > > an issue (I think we will get a cluster of a few BLADE servers), > > between > > SATA or SCSI RAID5, would any of them be significantly faster than the > > other, or would any of them be so slow that there would be no point in > > getting a lot of computational power? Also, given it is SATA or SCSI > > RAID5 > > and ethernet bandwidth is not an issue, would setting it up as a local > > hard > > drive or in a network attached storage system be significantly faster? > > This is all related to doing usual fMRI preprocessing and analysis. > > Thank you. > > > > Regards, > > Renat. > > > > > --- > Russell A. Poldrack, Ph.d. > Assistant Professor > UCLA Department of Psychology > Franz Hall, Box 951563 > Los Angeles, CA 90095-1563 > > phone: 310-794-1224 > fax: 310-206-5895 > email: [log in to unmask] > web: www.poldracklab.org > > -- Zhang Hao M.D. M.Sc. Senior Pharmacological Physician UCB Pharma R&D Chemin du Foriest Braine-l'Alleud B-1420 Belgium Tel: +32-2-386 6035 Fax: +32-2-386 3537