Intel® Manycore Testing Lab (Archived)

Performance of MTL login node

Newcomer to MTL here. My program runs fine on the Manycore Testing
Lab as long as it's singlethreaded. Whenever I enable multithreading the
performance gets worse on the same input. Starting more threads only
makes it slower.

At first I thought it was just poor design on my
part - too much contention for mutexes or something. However, now I'm
testing a program with NO synchronization between threads and
it's still happening. The performance is what one might expect from a single
core processor.

An Intel Microgrant application can be as simple as one paragraph!

Hello Team Intel Manycore Testing Lab,

Im starting a whole second thread on this topic because it is totally worth making sure the point is well understood in the Manycore Testing Lab community.

Applying for an Intel Microgrant is a very quick, very simple process. In fact, it can be as short as writing a single paragraph and submitting it. Folks, this is for money in your pocket to help you do work you are probably planning to do anyway.

2011 Intel Microgrant Awards: apply today!

Hello Intel Manycore Testing Lab forum readers!

Whether you are using the Manycore Testing Lab now, or you have in the past, or youre just deciding whether (or not) its for you, I wanted you to know about a great opportunity that Intel has running right now for its Academic Community.

If you are including parallel programming in your current curriculum, or, if you hope to modify your materials to include parallelism, then you need to check out the Intel Microgrant Award page, here:

Compilation problem with ICPC + OpenMP...


I am currently testing a parallel code on the MTL and I have some problems with ICPC+OpenMP compilation.

Note : the only command I launch on the MTL is "source /opt/intel/composerxe/bin/compilervars.shintel64" in order to compile with ICPC.

With GMP+OpenMP, I can compile my code without any problems. The makefile I use is this one :

Having trouble getting VNC to work

I am trying to remote into the login node using VNC (specifically Real VNC and Ultra VNC).I am using Putty for ssh tunneling.With Real VNC I get "read: connection reset by peer (10054)" after entering my password.With Ulra VNC I get "Write extract: Socket error while writing"I am using "L5901 localhost:5900" as my tunneling parameters for putty. My understanding is that this means that if I VNC into localhost:5901 I will get forwarded to 5900 on the login node.Are there any detailed instructions available for this.Thanks.

How to test these parameters

Hi, everyone
I want to collect the latency of hierarchy memory, would you please give me some advice about the testing sofware. I'm working on the Intel i3/i5/i7 processor; it seems that the Calibrator(0.9e) and Memory analyzer 3.8 didn't work well. They can't give thelatency of memory accessing, nor can work on the i series processors. Besides, how to measure the latency of L1I miss and branch misprediction?
Are there any other softwares or methods that can meet these requrements?
thanks a lot!
Best regards

Apply for Microgrant Awards- First deadline coming up soon!

Have you used the Intel Manycore Testing Lab to develop new course materials? Intel is sponsoring cash awards to encourage the creation of teaching content, including tools, games, labs, demonstrations, and other examples that can be used in the classroom to introduce parallel programming concepts.

Measurements of bandwidth, latencies, and NUMA effects

As part of some other research I was doing (for the memory system of our parallel dialect of ML, Manticore), I gathered a bunch of information on the memory bandwidth and latencies on the MTL machines and a 48-core AMD machine using enhanced versions of the STREAM benchmark. It ended up not fitting into the paper, but I thought the results might be useful/interesting to people here:

S’abonner à Intel® Manycore Testing Lab (Archived)