[Beowulf] SC13 wrapup, please post your own

Prentice Bisbal prentice.bisbal at rutgers.edu
Wed Nov 27 11:23:22 PST 2013


The problem with that measuring citations is that it  can take several 
years (sometimes decades) for a research paper to gain traction, which 
might be a longer time horizon that we have to show the ROI on an HPC 
investment.

While we're at it, we could through in someone's Erdos number in see how 
having access to HPC resources affects that. That's typically only 
relevant to mathematicians, though.

Regarding Andrew Wiles, I saw him on campus on a regular basis when I 
worked at IAS. One day, while driving home from work, decided to cross 
the street right as I got the green light (which meant he should have 
stayed on the sidewalk - a VERY common problem in Princeton), and I had 
to slam on my breaks to avoid hitting him. I later told a friend I 
almost ran over Andrew Wiles. Of course, the response was "who's that?"

--
Prentice

On 11/26/2013 09:54 AM, Peter St. John wrote:
> Oh there is another metric besides number of papers published; 
> Citation Indexing and Impact Factor, the predecessors of Google "Page" 
> Ranking.
> Instead of counting papers, or counting citations to papers, you count 
> citations weighted by their own citations, recursively.
> So one year Andrew Wiles publishes two papers. Those two papers are 
> read by maybe six specialists in arithmetic algebraic geometry. But 
> those six guys are read by many more, etc (the  recursion converges 
> rapidly, which is why Page Ranking is so effective), so Wiles' Impact 
> Factor is extravagant.
> Peter
>
>
> On Tue, Nov 26, 2013 at 9:39 AM, Tim Cutts <tjrc at sanger.ac.uk 
> <mailto:tjrc at sanger.ac.uk>> wrote:
>
>     On 25 Nov 2013, at 23:03, Prentice Bisbal
>     <prentice.bisbal at rutgers.edu <mailto:prentice.bisbal at rutgers.edu>>
>     wrote:
>
>>     4. I went to a BoF on ROI on HPC investment. All the presentations in
>>     the BoF frustrated me. Not because they were poorly done, but because
>>     they tried to measure the value of a cluster by number of papers
>>     published that used that HPC resource. I think that's a crappy,
>>     crappy
>>     metric, but haven't been able to come up with a better one myself
>>     yet. I
>>     was very vocal with my comments and criticisms of the
>>     presentations, so
>>     if any of the presenters are reading this now, I apologize for
>>     hi-jacking your BoF. Getting good ROI on a cluster is close to my
>>     heart,
>>     but is also difficult to quantify and measure. I hope I can be
>>     part of
>>     the discussion next year.
>
>
>     I can't think of another metric either.  At the top of my
>     organisation, publications are *the* key metric that all
>     scientists are judged on.  Publications are *the* product of any
>     scientific institution.  We don't sell anything, so we can't
>     measure revenue.  All we can measure are papers published per unit
>     time.  The problem is that the publication of the paper is very
>     distant from the building of your compute infrastructure, so it's
>     very hard to put a sensible number on ROI for this stuff.
>
>>     7. The cover band 'London Calling' played the IBM Platform
>>     Computing/Intel party again. Despite calling themselves 'London
>>     Calling'
>>     they still do not play any Clash songs. They are a good cover
>>     band, but
>>     it's starting to get boring seeing the same band play the same
>>     set year
>>     after year.
>
>     It did at least have more atmosphere than either of the parties I
>     went to earlier in the week, which were pretty much like drinking
>     in a morgue.  I won't name them, but we probably all know which
>     ones.  Both were at establishments on the 16th street Mall.  I had
>     a lot more fun and useful conversation in bars after abandoning ship…
>
>     Tim
>
>
>     -- The Wellcome Trust Sanger Institute is operated by Genome
>     Research Limited, a charity registered in England with number
>     1021457 and a company registered in England with number 2742969,
>     whose registered office is 215 Euston Road, London, NW1 2BE.
>
>     _______________________________________________
>     Beowulf mailing list, Beowulf at beowulf.org
>     <mailto:Beowulf at beowulf.org> sponsored by Penguin Computing
>     To change your subscription (digest mode or unsubscribe) visit
>     http://www.beowulf.org/mailman/listinfo/beowulf
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.beowulf.org/pipermail/beowulf/attachments/20131127/e8d3b4e5/attachment-0001.html 


More information about the Beowulf mailing list