SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
what is XLOC_000xxx system? feng Bioinformatics 1 01-06-2012 07:32 AM
PubMed: The JCVI standard operating procedure for annotating prokaryotic metagenomic Newsbot! Literature Watch 0 02-10-2011 06:10 AM
What Operating System Do You Use For Analysis? john_mu Bioinformatics 7 06-02-2010 02:24 AM
454 and cluster operating systems - patching Fedora rglover 454 Pyrosequencing 3 05-20-2010 06:23 AM
standard operating procedure 454 analysis neoanderson 454 Pyrosequencing 1 04-15-2010 03:00 AM

Reply
 
Thread Tools
Old 01-04-2011, 05:25 PM   #1
grlazo
Junior Member
 
Location: Berkeley, CA

Join Date: Apr 2010
Posts: 1
Lightbulb Tools Optimized for Operating System Environments

It seems that almost all well used NGS software used for assemblies and etc... are well adapted for the general Linux community, but in most cases they're adapted for Debian, Ubuntu, or Red Hat builds. We've become well acquainted with many of the software tools under Ubuntu, but unfortunately our machines are limited in memory or CPU capabilities. We have some machines of decent capability operating under the Solaris and SLES OS environments, but many of the packages fail to compile and are plagued with incompatibilities. It would be great to have some hints toward overcoming some of the compiling, or library hurdles taken to get some of these software packages working on the less used OS environments. Some compile issues have also been with the chipsets in use such as ia64 vs x86, and others with OS releases. In general, it would be great if prerequisites and success stories can be set for many of these packages regarding chipsets, OS, compile libraries and packages.
grlazo is offline   Reply With Quote
Old 01-05-2011, 04:08 AM   #2
dawe
Senior Member
 
Location: 4530'25.22"N / 915'53.00"E

Join Date: Apr 2009
Posts: 258
Default

This is a problem you'll have with most of the software available (not only for bionformatics)... BTW, I have some packages built on Solaris, Nexenta and FreeBSD, I may help in porting, if you like.

d
dawe is offline   Reply With Quote
Old 01-05-2011, 06:23 AM   #3
drio
Senior Member
 
Location: 4117'49"N / 24'42"E

Join Date: Oct 2008
Posts: 323
Default

Sometimes developers focus mainly in getting the tool up and running on the traditional OS and hardware, that's linux (The windows of this field ) and x86.

I have to say there are many tools out there that follow very well the typical open source building process and benefit from this approach. When the tool presents issues with a particular environment and the user contacts the community, authors, developers and users work together to make the tool run on that particular environment.
__________________
-drd
drio is offline   Reply With Quote
Old 01-05-2011, 06:49 AM   #4
nickloman
Senior Member
 
Location: Birmingham, UK

Join Date: Jul 2009
Posts: 356
Default

Quote:
Originally Posted by grlazo View Post
It seems that almost all well used NGS software used for assemblies and etc... are well adapted for the general Linux community, but in most cases they're adapted for Debian, Ubuntu, or Red Hat builds. We've become well acquainted with many of the software tools under Ubuntu, but unfortunately our machines are limited in memory or CPU capabilities. We have some machines of decent capability operating under the Solaris and SLES OS environments, but many of the packages fail to compile and are plagued with incompatibilities. It would be great to have some hints toward overcoming some of the compiling, or library hurdles taken to get some of these software packages working on the less used OS environments. Some compile issues have also been with the chipsets in use such as ia64 vs x86, and others with OS releases. In general, it would be great if prerequisites and success stories can be set for many of these packages regarding chipsets, OS, compile libraries and packages.
Running Ubuntu as a virtual machine on your higher spec'd machines might be an answer.

Another is to rent machines via Amazon or equivalent if you really can't afford to invest in hardware for your NGS analysis.

However, you might find you need less CPU and memory than you think. 8Gb of RAM and 8 cores is adequate for many tasks, and machines like this can be built for under $1000.
nickloman is offline   Reply With Quote
Old 01-05-2011, 07:15 AM   #5
jmw86069
Member
 
Location: RTP, NC, USA

Join Date: Jun 2009
Posts: 28
Default

I agree with the OP that this issue is valuable enough to capture the tricks of the trade somewhere. A few examples could be:
- compiling R optimized for the cpu and OS, whether on linux or OSX -- the speed can be improved substantially over the general-purpose default Ubuntu R package if compiled with the proper, and/or math libraries (ATLAS-tuned BLAS, Lapack, etc.)
- compiling tools with one of various flavors of MPI.
- use of different GCC versions for compiles.
- use of OpenCL/CUDA/ATI libraries in place of standard libraries -- I'm still curious whether their math libraries could substitute, for example, in R compiles above. And of course, I don't know the improvement/impact.

Are you proposing something specific? I.e. a wiki, or a Github?
jmw86069 is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 06:16 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO