Difference between revisions of "Talk:BioC Requirements Information Gathering"

From UFRC
Jump to navigation Jump to search
 
Line 3: Line 3:
 
* The perceived direction of the discussions at the BioC Investor meetings as of now - early November - seems to give more and more priority to the storage as the primary need and the raw computing capacity and the memory configuration as a somewhat secondary although not unimportant consideration. So, we should split the gathering of usage information between these two large areas to be able to distil the most important conclusions more easily.
 
* The perceived direction of the discussions at the BioC Investor meetings as of now - early November - seems to give more and more priority to the storage as the primary need and the raw computing capacity and the memory configuration as a somewhat secondary although not unimportant consideration. So, we should split the gathering of usage information between these two large areas to be able to distil the most important conclusions more easily.
  
'''Note:''' Please use <nowiki>~~~~</nowiki> - for tildas to mark your username and the date of the comment to allow some tracking of information, ideas and proposals.
+
'''Note:''' Please use <nowiki>~~~~</nowiki> - 4 tildas to mark your username and the date of the comment to allow some tracking of information, ideas and proposals. You don't have to fill out every section - just the ones you feel strongly about, those you would like to influence. Click on the "Edit" link above your lab name and start writing. Thank you!
  
 
= Storage =
 
= Storage =

Latest revision as of 21:42, 2 November 2011

General requirements

  • The perceived direction of the discussions at the BioC Investor meetings as of now - early November - seems to give more and more priority to the storage as the primary need and the raw computing capacity and the memory configuration as a somewhat secondary although not unimportant consideration. So, we should split the gathering of usage information between these two large areas to be able to distil the most important conclusions more easily.

Note: Please use ~~~~ - 4 tildas to mark your username and the date of the comment to allow some tracking of information, ideas and proposals. You don't have to fill out every section - just the ones you feel strongly about, those you would like to influence. Click on the "Edit" link above your lab name and start writing. Thank you!

Storage

List your estimates of your storage needs for the next 6, 12 and 18 months here. All labs investing over $3k are listed below alphabetically.

Baer Lab

Barbazuk Lab

Bloom Lab

Brocchieri Lab

ICBR

Kirst Lab

Kolaczkowski Lab

Mai Lab

McIntyre Lab

Paine Lab

Peter Lab

Ponciano Lab

Pulliam Lab

Reed Lab

Renne Lab

Riva Lab

Salemi Lab

Soltis Lab

Triplett Lab

UFGI

Smaller Investors

Please feel free to add additional lab names.

Memory

Baer Lab

Barbazuk Lab

Bloom Lab

Brocchieri Lab

ICBR

Kirst Lab

Kolaczkowski Lab

Mai Lab

McIntyre Lab

Paine Lab

Peter Lab

Ponciano Lab

Pulliam Lab

Reed Lab

Renne Lab

Riva Lab

Salemi Lab

Soltis Lab

Triplett Lab

UFGI

Smaller Investors

Please feel free to add additional lab names.

CPU

Baer Lab

Barbazuk Lab

Bloom Lab

Brocchieri Lab

ICBR

Kirst Lab

Kolaczkowski Lab

Mai Lab

McIntyre Lab

Paine Lab

Peter Lab

Ponciano Lab

Pulliam Lab

Reed Lab

Renne Lab

Riva Lab

Salemi Lab

Soltis Lab

Triplett Lab

UFGI

Smaller Investors

Please feel free to add additional lab names.

Job Duration

Baer Lab

Barbazuk Lab

Bloom Lab

Brocchieri Lab

ICBR

Kirst Lab

Kolaczkowski Lab

Mai Lab

McIntyre Lab

Paine Lab

Peter Lab

Ponciano Lab

Pulliam Lab

Reed Lab

Renne Lab

Riva Lab

Salemi Lab

Soltis Lab

Please feel free to add additional lab names.

  • Magitz 09:37, 31 October 2011 (EDT) :

A fair amount of the Soltis Lab usage of the cluster will be for phylogenetic analyses. Many of these last 1-2 weeks, with a few as long as a month. Being able to schedule longer jobs will be important. However, my sense is that things have improved from where we were a few years ago, in that we rearely need jobs that run over a month. Between checkpointing and better software, most of the analyses that I've seen running over 30days on our lab cluster or Fisher were

  1. forgotten about by the user who started them and no longer needed
  2. set up with unrealistic expectations (jobs on bootstrap replicate 3 after 30 days set to run for 1,000 replicates, etc)
  3. could have been set up in way to complete within a month.

Triplett Lab

UFGI

Smaller Investors