[Beowulf] Re: Beowulf Digest, Vol 37, Issue 58

On Mon, 26 Mar 2007, H?kon Bugge wrote: > And based on this I did not call it significant > findings, but merely an indication of RDMA being > faster (upto 16 cores) or equal fast as message > passing for _this_ application and dataset. My apologies Hakon, I misunderstood your intention. I'm

RE: [Beowulf] Performance characterising a HPC application

obviously, there are many applications which have absolutely no use for bandwidth greater than even plain old gigabit. equally obvious, there are others which are sensitive to small-packet latency, which is not affected by DDR or dual-rail. Yes, there are application that don't utilize the inter

[Beowulf] 1.2 us IB latency?

http://www.hpcwire.com/hpc/1336731.html well, if the numbers mean what they appear to mean, this is quite a dramatic breakthrough: 1.2 us, 25M msg/s. since we just happend to be talking about "conventional" IB vs Infinipath (1.29 us, 11.3M msg/s). or is this an example of message aggregation

Re: [Beowulf] Performance characterising a HPC application

Greg Lindahl wrote: > On Mon, Mar 26, 2007 at 03:58:59PM -0500, Richard Walsh wrote: > >> but of course aggregation is a legitimate optimization technique >> because not all message patterns are of the Gups variety just as not >> all memory references are absent locality. >> > > This is tru

Re: [Beowulf] Performance characterising a HPC application

On Mon, Mar 26, 2007 at 03:58:59PM -0500, Richard Walsh wrote: > but of course aggregation is a legitimate optimization technique > because not all message patterns are of the Gups variety just as not > all memory references are absent locality. This is true, although I would call it more of an "

Re: [Beowulf] Performance characterising a HPC application

Kozin, I (Igor) wrote: >> real codes that computes a minimum. However, an alltoall on many >> cores/nodes would exercise the same metric (many sends/recvs on the same >> NIC at the same time), but would be harder to cheat and be much more >> meaningful IMHO. >> > > Could not agree more. We

Re: [Beowulf] Performance characterising a HPC application

Greg Lindahl wrote: > Note that the new MVAPICH has message coalescing, which causes its > 2-node streaming bandwidth and message rate to rise. Note that real > apps rarely have that message pattern -- instead, they send a single > message each to lots of other nodes before synchronizing. Message r

Re: [Beowulf] Cell programming

On Mon, Mar 26, 2007 at 10:00:42AM -0500, Brian D. Ropers-Huilman wrote: > >paulo lopes > > http://www.hpcwire.com/hpc/1274729.html I've been unable to find any information to what the virtualization does to the GBit Ethernet latency and throughput. -- Eugen* Leitl http://leitl.org";>leitl htt

Re: [Beowulf] Performance characterising a HPC application

On Mon, 26 Mar 2007, Gilad Shainer wrote: > > > Offload, usually implemented by RDMA offload, or the ability > > for a NIC to autonomously send and/or receive data from/to > > memory is certainly a nice feature to tout. If one considers > > RDMA at an interface level (without looking at the

Re: [Beowulf] Performance characterising a HPC application

On Mon, Mar 26, 2007 at 09:38:43AM -0700, Gilad Shainer wrote: > This is yet another example of "fair" comparison. Unlike Qlogic, > Mellanox offer a family of products for PCIe servers, and there are > multiple MPI versions that support those products. The performance > depends on the hardware you

Re: [Beowulf] Performance characterising a HPC application

On Fri, Mar 23, 2007 at 11:53:14AM -0700, Gilad Shainer wrote: > and not only on pure latency or pure bandwidth. Qlogic till recently (*) > had the lowest latency number but when it comes to application, the CPU > overhead is too high. QLogic's overhead is lower than Mellanox, how low do you want

Re: [Beowulf] Cell programming

I have not yet progammed a Playstation, the documentation says that it is easy, see in particular http://www.cellperformance.com/articles/2006/11/crosscompiling_for_ps3_linux.html as a starting point. Though not officially supported for the IBM compilers, people have gotten advice on the IBM Foru

Re: [Beowulf] Cell programming

On 3/26/07, Paulo Afonso Lopes <[EMAIL PROTECTED]> wrote: Has anybody had any experience on the Sony PS3 has a "Cell node"? IBM's Cell blades are hugely expensive (around 17K euros per blade, here) when compared to PS3 (around 600 euros per box). paulo lopes http://www.hpcwire.com/hpc/1274729.

[Beowulf] Re: Beowulf Digest, Vol 37, Issue 58

On Mon, 26 Mar 2007, H?kon Bugge wrote: > Hi Christian, > > At 01:19 24.03.2007, [EMAIL PROTECTED] wrote: > >I've yet to see a significant number of message-passing applications > >show that an RDMA offload engine, as opposed to any other messaging > >engine, is a stronger performance determinant

Re: [Beowulf] The recently solved Lie Group problem E8

On 3/23/07, Michael Hannon <[EMAIL PROTECTED]> wrote: Peter St. John wrote: . . . > I wish I know more about the SAGE (machine) that hosts the SAGE > (software) that was used for this ... See the appended blurb from one of our sales guys. - Mike I th

Re: [Beowulf] The recently solved Lie Group problem E8

Peter St. John wrote: . . . I wish I know more about the SAGE (machine) that hosts the SAGE (software) that was used for this, but apparently washington.edu 's web server can't handle the CNN exposure as well as their number cruncher can crunch numbers. They are still dow

RE: [Beowulf] Performance characterising a HPC application

If anyone is interested in a 30 day evaluation copy of Scali MPI Connect... please send me an email. [EMAIL PROTECTED] Thanks/Ken -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Eugen Leitl Sent: Friday, March 23, 2007 4:26 PM To: beowulf@beowulf.org S

[Beowulf] Re: Beowulf Digest, Vol 37, Issue 58

Hi Christian, At 01:19 24.03.2007, [EMAIL PROTECTED] wrote: I've yet to see a significant number of message-passing applications show that an RDMA offload engine, as opposed to any other messaging engine, is a stronger performance determinant. That's probably because there are other equally imp

Re: [Beowulf] Cell programming

Has anybody had any experience on the Sony PS3 has a "Cell node"? IBM's Cell blades are hugely expensive (around 17K euros per blade, here) when compared to PS3 (around 600 euros per box). paulo lopes ... > I have been coding on the Cell for just over a month now. ... -- Paulo Afonso Lopes

[Beowulf] May I ask you about memtest test pattern?

Dear David Sorry to bother you. I am a student in Taiwan. I am interested in the coding about MemTest86+ V1.65 memory test pattern. Eventhough I have basic C and Assembly Language Knowledge, I have not able to understand the pattern part. Could you sugguest some books or knowledge for me to under

Re: [Beowulf] A start in Parallel Programming?

On Thu, 15 Mar 2007, Richard Walsh wrote: > Bremerman was a physicist.   My original encounter with the limit/him > was in a fascinating book published in the early 80s called "The Book of > Ignorance", a collection of articles on what we do not and/or cannot know > that we might like to. By some