Hui Zhou

If you don't care about testing particular provider, you can disable them e.g. `--disable-opx`

> > I am eager to move to 1.19.x > > Actually, I think I made a mistake. My PR targets the master branch, but I think I should have...

https://jenkins-pmrs.cels.anl.gov/job/mpich-review-ch4-xpmem/199/jenkins_configure=debug,label=xpmem,netmod=ofi,shmmod=xpmem/console

test:mpich/ch3/most test:mpich/ch4/most

> Is error propagation required by the MPI standard for collectives? No.

Locally verified: ``` [release ~/build/4.3.0a1] ls share/man/man3/*_c.3 |wc 168 168 6239 [release ~/build/4.3.0a1] ls share/doc/mpich/www3/*_c.html |wc 168 168 7751 [release ~/build/4.3.0a1] cat share/doc/mpich/www3/MPI_Allgather_c.html [release ~/build/4.3.0a1] cat share/man/man3/MPI_Allgather_c.3 .so man3/MPI_Allgather.3 [release...

> Actually on a closer look, the HTML docs need the same aliasing that was done for the manpages. Will have to do another PR to fix. I'll work on...

What is the output from `lstopo` (a utility from `hwloc`)?

We call `hwloc_set_cpubind` to bind the cpus, but it appears we did not check the return code. So it seems somehow the function may fail -- which result in no...

I was not able to reproduce it yet: ``` zhouh@amdgpu04:~/pull_requests/mpich-main> HYDRA_TOPO_DEBUG=1 mpirun -n 2 -bind-to user:0-15,16-31 ./t [proxy:0@amdgpu04] created hwloc xml file /tmp/hydra_hwloc_xmlfile_Rvrhg6 process 0 binding: 1111111111111111000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 process 1 binding:...