[NLPL Task Force (A)] [NLPL Board] Statistics NLPL
Stephan Oepen
oe at ifi.uio.no
Tue Apr 17 06:12:39 UTC 2018
hi martin,
in trying to gauge actual cpu and gpu usage by NLPL partners, i found your
summary below.
can you clarify for which period those ‘used’ values are? it almost looks
as if the values are over the project duration up to the day of reporting?
for example, uppsala was granted what looks like a generous quota of 645k
in january 2017, of which 14 months later 29k have been used?
overall, these usage figures look suspiciously small to me, even when
including the projects that are not formally under the NLPL umbrella: the
largest usage so far is CrossNLP, at 196k since may 2016 (at an average of
9k per month)?
also, what are the units? is there a CSC web page explaining the fine
points of accounting?
we are promised three million hours per year by CSC, and a total of three
million for 2017–19 by Sigma2. if the above reading is correct, we should
do everything we can to increase actual NLPL usage.
—regarding allocation decisions, no need to involve the steering group
(unless there are general questions of strategy): the infrastructure task
force has been given the mandate to control our allocations.
i would suggest allocating DIKU (anders), say, 200k initially. that is a
tiny fraction of what we have available, and i would find it difficult to
deny denmark something that sweden has already been given, or?
best wishes, oe
On Mon, 26 Feb 2018 at 12:26 Martin Matthiesen <martin.matthiesen at csc.fi>
wrote:
> Hello,
>
> Here the current stats of 26.2.18 with Jörg's and Filips projects. For
> technical reasons the two lists (Projects/Quota Details) overlap a bit in
> information.
> Let me know if only one is enough.
>
> Regards,
> Martin
>
> Projects:
>
> Joakim Nivre: 2000509(project_2000509, 04.01.2017-31.12.2019): Deep
> learning for Natural Language Processing (NeIC-NLPL)
> Quota:644901 members:10 discipline:Kielitieteet(Languages)
> Stephan Oepen: 2000582(project_2000582, 22.03.2017-22.03.2019): NeIC-NLPL
> Quota:10000 members:11 discipline:Tietojenkäsittely ja
> informaatiotieteet(Computer and information sciences)
> Jörg Tiedemann: 2000661(project_2000661, 09.06.2017-09.06.2018): NLPL-OPUS
> Quota:280901 members:2 discipline:Kielitieteet(Languages)
> Bjorn Lindi: 2000760(project_2000760, 09.10.2017-09.10.2019): Nordic
> Language Processing Lab (NLPL)
> Quota:10000 members:1 discipline:Kielitieteet(Languages)
> Jörg Tiedemann: 2000288(project_2000288, 29.03.2016-29.03.2017): BAULT
> Quota:9008 members:8 discipline:Kielitieteet(Languages)
> Jörg Tiedemann: 2000309(project_2000309, 28.04.2016-28.04.2019): CrossNLP
> Quota:470087 members:14 discipline:Muu tekniikka(Other
> engineering and technologies)
> Filip Ginter: tuy4622(textdat, 07.04.2005-15.03.2019): Textual Data Mining
> for Bioinformation Management
> Quota:291203 members:10 discipline:Biotieteet(Biological
> sciences)
> Filip Ginter: 2000391(project_2000391, 23.08.2016-10.01.2019): TurkuNLP EDU
> Quota:101324 members:2 discipline:Tietojenkäsittely ja
> informaatiotieteet(Computer and information sciences)
>
> Quota details:
>
> Project 2000509 Deep learning for Natural Language Processing (NeIC-NLPL)
> Joakim Nivre
> start 04.01.2017 end: 31.12.2019 budget: 30.08.2017
> CSC budget: 644901 used: 28978 remain: 615923
> -----------------------------------------------------------------
> Project 2000582 NeIC-NLPL Stephan Oepen
> start 22.03.2017 end: 22.03.2019 budget: 22.03.2017
> CSC budget: 10000 used: 0 remain: 10000
> -----------------------------------------------------------------
> Project 2000661 NLPL-OPUS Jörg Tiedemann
> start 09.06.2017 end: 09.06.2018 budget: 14.12.2017
> CSC budget: 280901 used: 35946 remain: 244955
> -----------------------------------------------------------------
> Project 2000760 Nordic Language Processing Lab (NLPL) Bjorn Lindi
> start 09.10.2017 end: 09.10.2019 budget: 09.10.2017
> CSC budget: 10000 used: 0 remain: 10000
> -----------------------------------------------------------------
> Project 2000288 BAULT Jörg Tiedemann
> start 29.03.2016 end: 29.03.2017 budget: 29.03.2016
> CSC budget: 9008 used: 7089 remain: 1919
> -----------------------------------------------------------------
> Project 2000309 CrossNLP Jörg Tiedemann
> start 28.04.2016 end: 28.04.2019 budget: 21.02.2018
> CSC budget: 470087 used: 196167 remain: 273920
> -----------------------------------------------------------------
> Project 2000391 TurkuNLP EDU Filip Ginter
> start 23.08.2016 end: 10.01.2019 budget: 10.01.2018
> CSC budget: 101324 used: 26961 remain: 74363
> -----------------------------------------------------------------
> Project tuy4622 Textual Data Mining for Bioinformation Management Filip
> Ginter
> start 07.04.2005 end: 15.03.2019 budget: 19.02.2018
> CSC budget: 291203 used: 99657 remain: 191546
> -----------------------------------------------------------------
>
>
> --
> Martin Matthiesen
> CSC - Tieteen tietotekniikan keskus
> CSC - IT Center for Science
> PL 405, 02101 Espoo, Finland
> +358 9 457 2376, martin.matthiesen at csc.fi
> Public key :
> https://pgp.mit.edu/pks/lookup?op=get&search=0x74B12876FD890704
> Fingerprint: AA25 6F56 5C9A 8B42 009F BA70 74B1 2876 FD89 0704
>
> ----- Original Message -----
> > From: "Jörg Tiedemann" <jorg.tiedemann at helsinki.fi>
> > To: "Stephan Oepen" <oe at ifi.uio.no>
> > Cc: "Martin Matthiesen" <martin.matthiesen at csc.fi>, "board" <
> board at nlpl.eu>
> > Sent: Friday, 23 February, 2018 21:58:25
> > Subject: Re: [NLPL Board] Statistics NLPL
>
> > Just a note: I already had to extend the NLPL-OPUS project once last year
> > because I was running out of the initial hours.
> >
> > Jörg
> >
> >> On 23 Feb 2018, at 20.48, Stephan Oepen <oe at ifi.uio.no> wrote:
> >>
> >> many thanks, martin!
> >>
> >> to complement the NLPL-specific numbers, i think it would also be
> >> valid to consider the usage of joerg and filip on their non-NLPL
> >> projects (and likewise for our group in oslo on Abel). unless joerg
> >> or filip object, could you pull out those additional numbers too?
> >>
> >> seeing as we are far from using up our NLPL allocations, arguably we
> >> should all charge all our computing to NLPL accounts?
> >>
> >> best, oe
> >>
> >>
> >> On Fri, Feb 23, 2018 at 8:42 PM, Martin Matthiesen
> >> <martin.matthiesen at csc.fi> wrote:
> >>> Hello,
> >>>
> >>> It seems these statistics were never sent and kept in my drafts
> folder. This
> >>> reflects the situation in January of the NLPL Project at CSC.
> >>>
> >>> Have a nice weekend!
> >>> Martin
> >>>
> >>>
> >>>
> >>> "Projects:"
> >>>
> >>> Joakim Nivre: 2000509(project_2000509, 04.01.2017-31.12.2019): Deep
> learning for
> >>> Natural Language Processing (NeIC-NLPL)
> >>> Quota:644901 members:10
> discipline:Kielitieteet(Languages)
> >>> Stephan Oepen: 2000582(project_2000582, 22.03.2017-22.03.2019):
> NeIC-NLPL
> >>> Quota:10000 members:11 discipline:Tietojenkäsittely ja
> >>> informaatiotieteet(Computer and information sciences)
> >>> Jörg Tiedemann: 2000661(project_2000661, 09.06.2017-09.06.2018):
> NLPL-OPUS
> >>> Quota:280901 members:2
> discipline:Kielitieteet(Languages)
> >>> Bjorn Lindi: 2000760(project_2000760, 09.10.2017-09.10.2019): Nordic
> Language
> >>> Processing Lab (NLPL)
> >>> Quota:10000 members:1
> discipline:Kielitieteet(Languages)
> >>>
> >>> These are stats from 1.1.2018:
> >>>
> >>> Project 2000509 Deep learning for Natural Language Processing
> (NeIC-NLPL)
> >>> Joakim Nivre
> >>> start 04.01.2017 end: 31.12.2019 budget: 30.08.2017
> >>> CSC budget: 644901 used: 14379 remain: 630522
> >>> -----------------------------------------------------------------
> >>> Project 2000582 NeIC-NLPL Stephan Oepen
> >>> start 22.03.2017 end: 22.03.2019 budget: 22.03.2017
> >>> CSC budget: 10000 used: 0 remain: 10000
> >>> -----------------------------------------------------------------
> >>> Project 2000661 NLPL-OPUS Jörg Tiedemann
> >>> start 09.06.2017 end: 09.06.2018 budget: 14.12.2017
> >>> CSC budget: 280901 used: 18624 remain: 262277
> >>> -----------------------------------------------------------------
> >>> Project 2000760 Nordic Language Processing Lab (NLPL) Bjorn Lindi
> >>> start 09.10.2017 end: 09.10.2019 budget: 09.10.2017
> >>> CSC budget: 10000 used: 0 remain: 10000
> >>> -----------------------------------------------------------------
> >>>
> >>>
> >>>
> >>>
> >>> --
> >>> Martin Matthiesen
> >>> CSC - Tieteen tietotekniikan keskus
> >>> CSC - IT Center for Science
> >>> PL 405, 02101 Espoo, Finland
> >>> +358 9 457 2376, martin.matthiesen at csc.fi
> >>> Public key :
> https://pgp.mit.edu/pks/lookup?op=get&search=0x74B12876FD890704
> >>> Fingerprint: AA25 6F56 5C9A 8B42 009F BA70 74B1 2876 FD89 0704
> >>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nlpl.eu/archives/infrastructure/attachments/20180417/5d5861ee/attachment.htm>
More information about the infrastructure
mailing list