There may actually be API changes in the UCX library that are causing this 
issue.

Could someone from the UCX community comment on this issue?


On Oct 3, 2019, at 1:15 PM, Llolsten Kaonga via users 
<users@lists.open-mpi.org<mailto:users@lists.open-mpi.org>> wrote:

Hello all,

It looks like this build error could be connected to the OS:

On CentOS 7.5, we can build OpenMPI 3.*.* and 4.*.* versions
On CentOS 7.7, we can build OpenMPI 3.*.* but not 4.*.*

I hope this additional bit of information is useful and helps to resolve the 
problem.

I thank you.
--
Llolsten

From: users 
<users-boun...@lists.open-mpi.org<mailto:users-boun...@lists.open-mpi.org>> On 
Behalf Of Llolsten Kaonga via users
Sent: Wednesday, October 02, 2019 12:46 PM
To: 'Open MPI Users' <users@lists.open-mpi.org<mailto:users@lists.open-mpi.org>>
Cc: Llolsten Kaonga <l...@soft-forge.com<mailto:l...@soft-forge.com>>
Subject: [OMPI users] openmpi-4.0.1 build error

Hello all,

OS: CentOS 7.7
OFED: MLNX_OFED_LINUX-4.7-1.0.0.1

Running the command “make all install” returns:

In file included from btl_uct_device_context.h:16:0,
                           from btl_uct_component.c:40:
btl_uct_rdma.h: In function ‘mca_btl_uct_get_rkey’:
btl_uct_rdma.h:58:5 error: too few arguments to function  ‘uct_rkey_unpack’
        uct_status = uct_rkey_unpack ((void *) remote_handle, rkey);
        ^

…

In file included from btl_uct.h:41:0,
                           from btl_uct_device_context.h:15,
                           from btl_uct_component.c:40
/usr/include/uct/api/uct.h:1377:14 note: expected ‘const struct uct_md_config_t 
*’ but argument is of type ‘struct uct_md **’
  Ucs_status_t uct_md_open(uct_component_h component, const char *md_name,
btl_uct_component.c:348:5: error: too many arguments to function ‘uct_md_open’
       uct_md_open (md_desc->md_name, uct_config, &md->uct_md);
       ^

I will be happy to send the whole log if that would be more useful/helpful

I thank you.
--
Llolsten


--
Jeff Squyres
jsquy...@cisco.com<mailto:jsquy...@cisco.com>

Reply via email to