dmitrygx
dmitrygx
@jsofri did you have a chance to check v1.12 or master branch?
> @yosefe @dmitrygx happy to update that you are correct, using v1.12.0 solved it > > ``` > root@bff1c4b09bf0:/# ucx_info -v > # UCT version=1.12.0 revision 6ab55c4 > ``` @jsofri...
> maybe there is some system security or limits config that prevents xpmem from attaching peer process memory? @yosefe can it be a problem trying to do `xpmem_attach()` for 0-size...
> What did I do wrong? It is not allowed to mix "^" with desired transports. pls, set `UCX_TLS=dc,self,posix,sysv,cma` - according to you output you have `cma`/`posix`/`sysv` - SHM transports...
> > @dmitrygx good catch! i guess we should just ignore 0-size regions and create some kind of dummy rkey for them > > As a hack I can allocate...
> > > > @dmitrygx good catch! i guess we should just ignore 0-size regions and create some kind of dummy rkey for them > > > > > >...
@shamisp @thomasgillis I'd start checking if taking the latest UCX helps to fix @thomasgillis's issue. is it possible to download UCX by following the instruction (see below)? ``` 1. git...
@bureddy @Akshay-Venkatesh could you review pls?
> We should disallow passing GPU memory when UCX_TLS explicitly disable Cuda. What if users disable memtype cache, or pass smaller buffer that would use eager/bcopy? does it mean that...
> I think we should do it when ENABLE_PARAMS_CHECKS is turned on (debug build) 👍 > We can document "ucp_request_param_t::memory_type" that it must be one of the types returned by...