[Beowulf] Broadwell HPL performance

[John Hearns]

Doug,

I read the same report and that what was confusing me too.
I don't get quite those numbers, but what I get is impressive.

Predicting the theoretical peak performance of a CPU socket these days is not an exact science, as I suppose it once was when mechanical relays were clicking on and off (*)
When AVX is running on a core it will clock down, and there seems to be a 'fudge factor' of 2 applied to reflect then gains which the FMA instructions give you.  Interesting stuff anyway.


(*) I bet it was nt an exact science then either!


________________________________________
From: Beowulf [beowulf-bounces at beowulf.org] on behalf of Douglas Eadline [deadline at eadline.org]
Sent: 21 April 2016 13:56
To: Christopher Samuel
Cc: beowulf at beowulf.org
Subject: Re: [Beowulf] Broadwell HPL performance

> On 20/04/16 16:52, John Hearns wrote:
>
>> I would be grateful for pointers towards HPL performance figures for
>> Broadwell (v4) processors.
>>
>> I ask as I am getting some very good values and I want to do a sanity
>> check!
>
> Dell has some HPL and STREAM benchmark numbers here (found with a quick
> google):
>
> http://en.community.dell.com/techcenter/high-performance-computing/b/general_hpc/archive/2016/03/31/measuring-performance-of-intel-broadwell-processors-with-high-performance-computing-benchmarks

Expect the author is a bit confused about calculating % of peak
and in one case (E2650) seems be reporting a result
of 878 GFLOPS vs a theoretical peak of 844.

He  states in Figure 1 that the E2650 result is 127% efficient
which I assume is based on his bogus 356 GFLOP theoretical
maximum, thus the actual number he reports in
Figure 1 is 878 GFLOPS, (1.27x346x2 processors)

Basically I don't trust these numbers. I assume the rest of the
data are equally wrong. Being an old school
kind of dude a link to the raw output is always nice
and some run-time data like HT (on off) and number of threads
etc. is helpful.

>
> Others from Boston in the UK:
>
> https://www.boston.co.uk/blog/2016/04/06/intel-xeon-e5-2600-v4-codename-broadwell-launch-and-preliminary-bench.aspx

They report 859 GFLOPS for the E5-2650, again above peak, but they
seem to state that HT is on. How many threads do they use for the test?

I assumed that enabling HT hurt HPL numbers (at least in my MPI tests it
does). Is it possible that for these tests, HT helps performance (a bit),
but in the case of the Dell blog, including the HT "cores" means doubling
the peak number which would make the result look bad?

Sigh.

--
Doug


>
> All the best,
> Chris
> --
>  Christopher Samuel        Senior Systems Administrator
>  VLSCI - Victorian Life Sciences Computation Initiative
>  Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
>  http://www.vlsci.org.au/      http://twitter.com/vlsci
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf
>
> --
> Mailscanner: Clean
>


--
Doug

--
Mailscanner: Clean

_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf

#####################################################################################
Scanned by MailMarshal - M86 Security's comprehensive email content security solution.
#####################################################################################
Any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the company. Employees of XMA Ltd are expressly required not to make defamatory statements and not to infringe or authorise any infringement of copyright or any other legal right by email communications. Any such communication is contrary to company policy and outside the scope of the employment of the individual concerned. The company will not accept any liability in respect of such communication, and the employee responsible will be personally liable for any damages or other liability arising. XMA Limited is registered in England and Wales (registered no. 2051703). Registered Office: Wilford Industrial Estate, Ruddington Lane, Wilford, Nottingham, NG11 7EP