[OMPI devel] Use non-cpu frequency for timers

Hi, In amd64, we have rdtsc instruction which uses a timer updated at cpu frequency. The frequency is given by variable, opal_timer_linux_freq which is the cpu frequency. Consider the case where a timer updated at non-cpu frequency is to be used. Question: In this case, how do we specify the fr

Re: [OMPI devel] RFC: Public Test Repo

If OpenHPC folks want to cultivate such a repo I think that would be great. If such a repo existed then I think we should work with them on making sure that it integrates nicely into MTT so we can add it to the nightly regressions. Getting a release of MTT is something I would like to see happen s

[OMPI devel] Datatype flag?

Hey folks Is there some flag by which the datatype code can know what transport is being used? For example, suppose a transport can handle certain datatype configurations itself, without the converting dealing with them (e.g., contiguous vs non-contiguous). Essentially, it’s an “offload” capabi

Re: [OMPI devel] Datatype flag?

Ralph, We do not have such infrastructure as there are too many possibilities to handle efficiently. However, different projects (Mellanox and NVidia) are using techniques that look similar to what you propose. The main idea is to replace the default behavior of the convertor with a BTL-specializ

Re: [OMPI devel] Use non-cpu frequency for timers

Abhishek, That frequency variable doesn't need to be the CPU frequency. If you look at opal/mca/timer/linux/timer_linux_component.c:opal_timer_linux_find_freq() you will see the initialization for opal_timer_linux_freq. In the case of PPC the time stamp counter runs at a frequency, distinct from

[OMPI devel] v1.10.3rc2 ready for final testing

Hi folks I have posted rc2 of 1.10.3 - this contains all the changes that have been provided, and should be the final release candidate. Barring anything you find, I expect to just make a final pass on the NEWS items and release by end of May. https://www.open-mpi.org/software/ompi/v1.10/