AnnaLena77
AnnaLena77
Oh yes, that works fine, also with the btl uct. Sorry, I've been searching for ages but haven't found the known issue.
I have to correct myself. Unfortunately, the approach does not work. I had first tried the whole thing with the same data type as receive-type (“blocktype_resize”). This no longer caused...
> I will try to reproduce the issue. > > Meanwhile, you can try the following workaround and see how it goes > > ``` > mpirun --mca coll ^tuned...
I know the collective modules a little, as I have been working with the code for some time. Tuned selects an algorithm from “base” based on criteria and executes it....
I am working on an open MPI fork that measures internal performance data (similar to Peruse) at runtime and writes it to a database. This performance data is then visualized...
> @AnnaLena77 which UCX version are you using? 1.16.0
> @AnnaLena77 can you try to > > ``` > mpirun -np 81 --mca btl ob1 --mca btl_uct_eager_limit 8178 ./test > ``` No, does not work... I will have a...
> FWIW, I do not observe a deadlock but a crash (assertion error, I configure'd with `--enable-debug` > > I use this trimmed reproducer based on your test case, and...
Strange... very very strange: The problem is caused by the combination of uct and sm. ```c mpirun -n 81 --mca pml ob1 --mca btl ^uct ./cannon_test ``` --> works! ```c...
> I edited my reproducer and changed the allocated size of the `buf` buffer. > > If Open MPI is configure'd with `--disable-debug`, then the program does not hang nor...