George Bosilca wrote:
> Btw, can you run the Netpipe benchmark on this configuration please ?
> Once compiled with MPI support and once with TCP. This might give us
> more equitable details (same benchmark).
NPmpi and NPtcp belong to Netpipe but this doesn't mean that they do the same ;-). Anywa
Steve Wise wrote:
Are you using Chelsio's TOE drivers? Or just a driver from the distro?
We use the Chelsio TOE drivers.
Steve Wise wrote:
Ok. Did you run their perftune.sh script?
Yes, if not we wouldn't get the 1.15 GB/s on the TCP level. We had ~800 MB/s
before
primarily because of t
Hello again ;),
after getting acceptable latency on our Chelsio S320E-CXA adapters we now want
to check if we can
also tune the bandwidth. On TCP level (measured via iperf) we get 1.15 GB/s, on
MPI level (measured
via MPI-Ping-Pong) just 930 MB/s. We already set btl_tcp_sndbuf and
btl_tcp_rcvb
Thanks again for all the answers. It seems that were was a bug in the driver in
combination with
Suse Linux Enterprise Server 10. It was fixed with version 1.0.146. Now we have
12us with NPtcp and
22us with NPmpi. This is still not fast enough but for the time acceptable. I
will check the
alter
Zitat von Jeff Squyres :
On Jul 28, 2008, at 2:53 PM, Andy Georgi wrote:
we use Chelsio S320E-CXA adapters
(http://www.chelsio.com/assetlibrary/products/S320E%20Product%20Brief%20080424.pdf)
in one of our clusters. After tuning the kernel i measured the ping
pong latency via NetPIPE and
Hi all,
we use Chelsio S320E-CXA adapters
(http://www.chelsio.com/assetlibrary/products/S320E%20Product%20Brief%20080424.pdf) in one of our
clusters. After tuning the kernel i measured the ping pong latency via NetPIPE and got ~12us which
is pretty good for TCP i think. So i wrote a simple pin
Dear all,
to explain the behavior of MPI_Reduce on our cluster i ran through the
source of Open MPI 1.2.6. On line 357 i found a mistake (maybe ;-)). It
should be:
return ompi_coll_tuned_reduce_intra_binary(sendbuf, recvbuf, count,
datatype, op, root, comm, segsize);
instead of
return ompi_col
Thanks a lot, i got it.
Andy
George Bosilca schrieb:
You need access to the source in order to look at the decision maps.
They are in ompi/mca/coll/tuned/coll_tuned_decision_fixed.c file.
george.
On Apr 19, 2008, at 3:41 AM, Andy Georgi wrote:
hello all,
i want to measure the time
hello all,
i want to measure the time needed for collective operations on a cluster
by using ompi 1.2.5. i read that the algorithm which is used depends on
msgsize and/or commsize and ompi constructs a decision map for the
collective on the system. now my question:
where can i find these decisio
hi everybody,
i'm new on this list and started using OpenMPI for my parallel jobs. first step
was to measure the latency for blocking communication functions. now my first
question: is it possible that ordained communication pairs will be optimized?
background:
latency for special processnumbe
10 matches
Mail list logo
