very high bandwidth, low latency manner?
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Steffen Persvold sp at scali.comFri Apr 12 06:10:34 PDT 2002
- Previous message: very high bandwidth, low latency manner?
- Next message: very high bandwidth, low latency manner?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Fri, 12 Apr 2002, Iwao Makino wrote: > At 5:29 +0200 12.04.2002, Steffen Persvold wrote: > >On Thu, 11 Apr 2002, Craig Tierney wrote: > > > >> > >> I talked to a guy at SC2002 from Quadrics and he said > >> that list pricing on a Quadrics network was about $3500 > >> per node when you are in the 100s of nodes and up. > >> The price includes the cards, cables, switches, > >> etc. This doesn't include any sort of discount that you > >> might get. Myrinet is about $2000 for an equivelent > >> network at list price. Dolphin/SCI falls around $2245 list > >> per node (if the system is > 144 nodes and you have to get > >> the 3d card). > >> > >> > > > >This is list prices for the cards only, right ? What about the switches > >needed. AFAIK Quadrics and Myrinet both need switches, SCI don't (which > >makes the total system cost a bit lower doesn't it ?). > > As said on above, QsNet is per node price including all. > Do does Myrinet, so SCI/Dolphin and Myrinet is about equal. > Yes, sorry I missed that statement :) > SCI has good idea of not using switches, but on the other hand, it is > little more complex to connect. > True, but if one of your Myrinet switches breaks down you loose 64 nodes in a 256 node system (standard "CLOS" configuration). I don't know the MBTF for Myrinet switches, but I would expect it to be rather high (redundant power supplies ?). Please don't misunderstand me, I find the Myrinet interconnect very interesting and also competitive with SCI both from a technological point of view and wrt. pricing. The only thing this list is lacking is some head to head performance comparisons of the different interconnects e.g some NAS benchmarks and maybe also PMB. Regards, -- Steffen Persvold | Scalable Linux Systems | Try out the world's best mailto:sp at scali.com | http://www.scali.com | performing MPI implementation: Tel: (+47) 2262 8950 | Olaf Helsets vei 6 | - ScaMPI 1.13.8 - Fax: (+47) 2262 8951 | N0621 Oslo, NORWAY | >320MBytes/s and <4uS latency
- Previous message: very high bandwidth, low latency manner?
- Next message: very high bandwidth, low latency manner?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
