Re: [Beowulf] 1.2 us IB latency?

Hi Peter, Peter St. John wrote: I was wondering if Peter K's remark generalized: if there are multiple ports, the node has a choice, which may be application dependent. One port for MPI and the other to a disk farm seems clear, but it still isn't obvious to me that a star topology with few lon

Re: [Beowulf] Performance characterising a HPC application

Hi Richard, Richard Walsh wrote: It is useful in a message context in the same way the vector instructions are useful in a memory reference context. Automatic "message" aggregation is an optimization technique being used in UPC compiler development. Message aggregation would be mu

Re: [Beowulf] Performance characterising a HPC application

Gilad, Gilad Shainer wrote: So now we can discuss technical terms and not marketing terms such as price/performance. InfiniBand uses 10Gb/s and 20Gb/s link signaling rate. The coding of the data into the link signaling is 8/10. When someone refer to 10 and 20Gb/s, it is for the link speed and t

[Beowulf] The obligatory PS3 cluster computer

http://www.elecdesign.com/Articles/Index.cfm?AD=1&ArticleID=15229&bypass=1 James Lux, P.E. Spacecraft Radio Frequency Subsystems Group Flight Communications Systems Section Jet Propulsion Laboratory, Mail Stop 161-213 4800 Oak Grove Drive Pasadena CA 91109 tel: (818)354-2075 fax: (818)393-6875

Re: [Beowulf] 1.2 us IB latency?

Peter Kjellstrom wrote: On Wednesday 28 March 2007, Mark Hahn wrote: start timer send(other,small-message) recv(first,small-message) recv(other,small-message) send(first,small-message) stop timer I'll actually see 2.4 us between the timer calls? if I understand, agg

[Beowulf] wiring methods, transient suppression, etc.

For those of you doing things like specifying UPSes, wiring racks (and EPOs), etc. I've just finished reading over the revised IEEE Emerald Book (IEEE Std 1100-2005) and it is definitely the reference book you need. All the stuff you've tried to discern by reading marketing white papers from

Re: [Beowulf] 1.2 us IB latency?

On Wednesday 28 March 2007, Mark Hahn wrote: > >> start timer > >> send(other,small-message) recv(first,small-message) > >> recv(other,small-message) send(first,small-message) > >> stop timer > >> > >> I'll actually see 2.4 us between the timer calls? if I understand, > >> aggreg

Re: [Beowulf] 1.2 us IB latency?

>> also, I'm sorta amazed people keep selling (and presumably buying) >> dual-port IB cards. doesn't that get quite expensive, switch-wise? > > Not defending them but, It could possibly maybe be useful if you have a > stand-alone IB net for, say, storage or something else not mpi. Also, it's > n

[Beowulf] Overcoming processor/accelerator multiplicity and heterogeniety with virtual machines ... ??

All, The nice-ities of commodity, ILP and clock-driven performance improvements have given way to the complexity of multi-core (with two flavors, hetero and homo) and connected accelerators/processors (GPUs, FPGAs, HPC Specific Accelerators). And while the hardware mentioned is various, a large p

Re: [Beowulf] 1.2 us IB latency?

start timer send(other,small-message) recv(first,small-message) recv(other,small-message) send(first,small-message) stop timer I'll actually see 2.4 us between the timer calls? if I understand, aggregation would only help on a streaming test. in fact, this kind of is