desktop workstation configuration
2
0
Entering edit mode
Shi, Tao ▴ 720
@shi-tao-199
Last seen 7.1 years ago
Hi list, We're planning to get a new Linux desktop workstation. I'm wondering what kind of configuration (especially processor, memory, and OS) people are using (no cluster please). I know it's all about how much money you're willing to spend, but I just want to get a feel on what settings people are happy with for their daily bioinformatics jobs (including NGS data). Thank you very much! ...Tao [[alternative HTML version deleted]]
• 954 views
0
Entering edit mode
Thomas Girke ★ 1.7k
@thomas-girke-993
Last seen 18 months ago
United States
This is probably not a topic for this list, but still here are some suggestions based on my end of the year experience with purchasing computer hardware for similar reasons: For NGS analysis I would look into computers with 2 or 4 CPU sockets and as many as 32 DIMM (memory) slots. These systems are relatively inexpensive right now and real workhorses for NGS analysis. Roughly, one machine with 48 cores (4x twelve core CPUs), 64GB of RAM and 6-12TB of disk space is around $10-12K or about$18K with 256GB of RAM, while 2 CPU socket systems with 32GB of RAM are in the range of $5-6K. Systems like this are available from various vendors as workstations (most expensive!) or rack-mountable 1U or 2U servers. If you don't deal with assemblies of NGS data then 32-64GB of RAM will be probably enough for most things, but if you do then you can never have enough memory. The most important hardware component in the NGS field is probably disk space. If you plan to process a lot of data in your lab then you might want to consider an attached/centralized storage solution that can grow over time, but these usually add a lot to the cost (~$8-10K for 24-32TB) and setup time. Also, since powerful computers are noisy, I would consider a rack-mountable solution rather than a workstation that you install in a server room at your university. A server setup is usually much more cost efficient because many users can use it at the same time. Thomas On Fri, Dec 17, 2010 at 10:34:49AM -0800, Shi, Tao wrote: > Hi list, > > We're planning to get a new Linux desktop workstation. I'm wondering what kind > of configuration (especially processor, memory, and OS) people are using (no > cluster please). I know it's all about how much money you're willing to spend, > but I just want to get a feel on what settings people are happy with for their > daily bioinformatics jobs (including NGS data). > > > Thank you very much! > > > ...Tao > > > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >
0
Entering edit mode
Hi Thomas and Steve, Thanks for your helpful suggestions! ...Tao ________________________________ From: Thomas Girke <thomas.girke@ucr.edu> Cc: bioconductor at stat.math.ethz.ch Sent: Fri, December 17, 2010 2:00:51 PM Subject: Re: [BioC] desktop workstation configuration This is probably not a topic for this list, but still here are some suggestions based on my end of the year experience with purchasing computer hardware for similar reasons: For NGS analysis I would look into computers with 2 or 4 CPU sockets and as many as 32 DIMM (memory) slots. These systems are relatively inexpensive right now and real workhorses for NGS analysis. Roughly, one machine with 48 cores (4x twelve core CPUs), 64GB of RAM and 6-12TB of disk space is around $10-12K or about$18K with 256GB of RAM, while 2 CPU socket systems with 32GB of RAM are in the range of $5-6K. Systems like this are available from various vendors as workstations (most expensive!) or rack-mountable 1U or 2U servers. If you don't deal with assemblies of NGS data then 32-64GB of RAM will be probably enough for most things, but if you do then you can never have enough memory. The most important hardware component in the NGS field is probably disk space. If you plan to process a lot of data in your lab then you might want to consider an attached/centralized storage solution that can grow over time, but these usually add a lot to the cost (~$8-10K for 24-32TB) and setup time. Also, since powerful computers are noisy, I would consider a rack-mountable solution rather than a workstation that you install in a server room at your university. A server setup is usually much more cost efficient because many users can use it at the same time. Thomas On Fri, Dec 17, 2010 at 10:34:49AM -0800, Shi, Tao wrote: > Hi list, > > We're planning to get a new Linux desktop workstation. I'm wondering what kind > > > of configuration (especially processor, memory, and OS) people are using (no > cluster please). I know it's all about how much money you're willing to spend, > > > but I just want to get a feel on what settings people are happy with for their > daily bioinformatics jobs (including NGS data). > > [[elided Yahoo spam]] > > > ...Tao > > > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: >http://news.gmane.org/gmane.science.biology.informatics.conductor >
0
Entering edit mode
@steve-lianoglou-2771
Last seen 1 day ago
United States
Hi, On Fri, Dec 17, 2010 at 1:34 PM, Shi, Tao <shidaxia at="" yahoo.com=""> wrote: > Hi list, > > We're planning to get a new Linux desktop workstation. ?I'm wondering what kind > of configuration (especially processor, memory, and OS) people are using (no > cluster please). ?I know it's all about how much money you're willing to spend, > but I just want to get a feel on what settings people are happy with for their > daily bioinformatics jobs (including NGS data). I'm not sure what type of answers you think will be helpful. Whatever you purchase will likely have some of the latest processors w/ several cores, and however you shake that out, it'll probably be OK (each core is probably around ~3 GHz of something (i5, i7, something else?) and I guess you'll get something between 8 and 12 cores). I think the most useful suggestion I could make is to get more RAM than you think you need. When I got my desktop (2008(?)), I thought the 8GB of RAM it had was a lot, but I also wasn't working with much NGS data back then. Now I am, and I'd like to have at least 16GB -- especially if you think you want to do much read-preprocessing within Bioconductor tools. Also .. fast hard drives. -steve -- Steve Lianoglou Graduate Student: Computational Systems Biology ?| Memorial Sloan-Kettering Cancer Center ?| Weill Medical College of Cornell University Contact Info: http://cbio.mskcc.org/~lianos/contact