[Wrf-users] current Xeon processors

[Κώστας Μυρωνάκης]

Hi,
it seems that the Intel continues its old tactics of marketing.
To see how we come out of this trap!
We go back some years.
Look here:

#Cores #Chips #Processor   #Result #year #price
4      1    Intel i7-860   121     2010  274,08 €
4      1    Intel i7-870   127     2010  538,46 €
4      1    Intel i7-920   109     2010  266,78 €
6      1    Intel i7-980X  168     2010  1.013,47 €
8      2    Intel E5520    200     2010  369,88 € x2
8      2    Intel E5420    109     2010  295,84 € x2

Lets say that the i7-860 run WRF test case in 1 hour.
there is not difference in time with i7-870 which cost double price !!! and
it has a biggest cache .. bla bla bla ...

Another trap
i7-980X (X = 6 core) is faster 50% than i7-860 but its cost 4x !!!
if a have a cluster with four i7-860 its give 4x121=484 ~ 150% faster than
i7-980X !!!

Another trap
Lets see Xeon, the E5520 has a double performance than E5420.
The configuration of two E5420 has a double price than i7-860 and its
slower.
Βe careful.

Another trap
To buy or rent a PC?
It depends, if i am a government  or a person !!!
Difficult decision.

So open an excel file, find the possessors with same results and take or
rent the cheapest !!!

http://www.cpubenchmark.net/high_end_cpus.html


Efstathios Konstantinos Mironakis
http://www.forecastmaps.eu



2016-02-26 21:47 GMT+02:00 Zulauf, Michael <Michael.Zulauf at iberdrolaren.com>
:

> Hi all...
>
>
>
> We’re looking to upgrade our computational cluster, whose primary purpose
> is to run WRF, and I’m in the process of writing up the specs.  We’ll
> almost definitely stay with our current type of architecture (HP blades,
> infiniband interconnect, etc).
>
>
>
> Our current blades are dual Xeon x5670 (each CPU has 6 cores, 2.93 GHz, 12
> MB cache),  and we’ll probably go with dual Xeon E5s of some flavor.  Here
> are a few of the candidates:
>
>
>
> Xeon E5-2640v3
>
> 8 cores
>
>                 2.6GHz
>
>                 20 MB cache
>
>
>
> Xeon E5-2660v3
>
> 10 cores
>
>                 2.6GHz
>
>                 25 MB cache
>
>
>
> Xeon E5-2680v3
>
> 12 cores
>
>                 2.5GHz
>
>                 30 MB cache
>
>
>
> Leaving aside cost (for now), anyone have recommendations on these, or
> other similar processors?  When we went thru this several years ago, there
> didn’t appear to be too much benefit with going with the processors that
> had the highest number of cores.  I think the additional cores seemed to
> run significantly less efficiently (at least in the case of WRF).
>
>
>
> Has that changed?  The processors with more cores are substantially more
> expensive (and may be clocked slightly lower), but if WRF runs well on them
> then it might be worth shelling out the money (not my money) to get more
> cores.  I probably can’t justify buying more blades, even if I go with
> cheaper CPUs, so if the E5-2680v3 (for example) runs significantly better
> than the others, I could probably justify them.
>
>
>
> Thoughts?
>
>
>
> Thanks,
>
> Mike
>
>
>
> --
>
> Mike Zulauf
>
> Meteorologist, Lead Senior
>
> Operational Meteorology
>
> Iberdrola Renewables
>
>
>
> ==============================================================
>
> Please consider the environment before printing this email.
>
> If you have received this message in error, please notify the sender and immediately delete this message and any attachment hereto and/or copy hereof, as such message contains confidential information intended solely for the individual or entity to whom it is addressed. The use or disclosure of such information to third parties is prohibited by law and may give rise to civil or criminal liability.
>
> The views presented in this message are solely those of the author(s) and do not necessarily represent the opinion of Iberdrola Renewables, LLC. or any company of its group. Neither Iberdrola Renewables, LLC. nor any company of its group guarantees the integrity, security or proper receipt of this message. Likewise, neither Iberdrola Renewables, LLC. nor any company of its group accepts any liability whatsoever for any possible damages arising from, or in connection with, data interception, software viruses or manipulation by third parties.
>
>  ==============================================================
>
>
> _______________________________________________
> Wrf-users mailing list
> Wrf-users at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/wrf-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20160329/ecddc51e/attachment-0001.html