Le 08/02/2012 22:33, Paul H. Hargrove a écrit : > > > On 2/8/2012 8:58 AM, Jeff Squyres wrote: >> * Fix conversion from/to Linux libnuma when some NUMA nodes have no >> memory. > > Tests on the virtual node I have access to where that problem report > originated is still not quite right. > There is now a different assertion failing than I had seen before: >> lt-linux-libnuma: >> /users/phh1/OMPI/hwloc-1.3.2rc1-linux-ppc64-gcc//hwloc-1.3.2rc1/tests/linux-libnuma.c:83: >> main: Assertion `!memcmp(&nodemask, &numa_all_nodes, >> sizeof(nodemask_t))' failed. >> /bin/sh: line 5: 19416 Aborted ${dir}$tst >> FAIL: linux-libnuma > > I don't have any clue if that represents forward or backward progress.
For the record, I am debugging this offline and of course this is again libnuma's fault. The old nodemask API seems to be broken on machines with sparse node ids. And the actual behavior in such cases changed in 2.0.6 (Paul has 2.0.3, I have 2.0.8). I think I will disable most nodemask tests when node ids are sparse. I am also complaining on the numa devel list about their crappy interface and doc... If they confirm that the nodemask API is totally deprecated, I'll mark our nodemask helpers as deprecated and we'll remove them as soon as possible. Way too much time wasted already. Brice