[Beowulf] Intel buys QLogic InfiniBand business

Vincent Diepeveen diep at xs4all.nl
Sat Jan 28 16:12:59 PST 2012


On Jan 28, 2012, at 10:28 PM, Mark Hahn wrote:
[snip]

> out of curiosity, has anyone set up a head-to-head comparison
> (two or more identical machines, both with a Qlogic and a Mellanox  
> card of
> the same vintage)?
>
> regards, mark hahn.

Mark, i stumbled upon the same problem a few months ago when i  
googled for 4x infiniband you can find something,
when moving up to QDR it becomes more sporadic.
Not to mention that the interesting test is where the cards are bad -  
latency.
If you find anything, usually it's manufacturer side statements  
without clear testsetup and usually doing 0 byte tests.

This is exactly why i intend to write a benchmark.

What i personally believe is not important whether FDR,  pci-e 3.0  
and a considerable higher claimed bandwidth than pci-e 2.0 QDR.

What i do believe is that one must measure objectively.

That's why i'm posting for a while now that as soon as the cluster  
works here i'm gonna
write a benchmark to measure latencies moving up the read length  
slowly so that it more and more gets a bandwidth game and simply  
present the
graph for the interested readers.

We're not interested in theoretic tests of 1 core busy that is  
measuring a latency of another core at the other side busy.

A test really requires all cores busy and hammering onto the network  
card.

In the end always everything is a measure of bandwidth of course, but  
even then the lack of scientists online who tested objectively QDR,
no matter *what manufacturer*, such tests really are there in short  
supply and some of them either just tested 1 tiny thing or a  
theoretic thing,
or just lacked all realism when i read the rest of the article.

All with all, after some days of googling,

I found 1 tester who toyed something using the same switch (good  
idea) but the graphs drawn presenting the results are tough to interpret
and basically was interested in something else than what's fast now  
for the network cards.

Running the same oldie tests, whereas all manufacturers have way  
faster alternatives now, such as RDMA reads, is just not interesting.

To be continued in some months...

> _______________________________________________
> 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




More information about the Beowulf mailing list