Hi Santiago, On Thu, Apr 28, 2016 at 05:31:25PM +0200, Santiago Vila wrote: > Sorry, I can't reproduce this consistently, but when the problem > happens, I see a segfault like this in dmesg: > > [ 5377.752190] julia[21086]: segfault at 20000ffe2 ip 00007f033628b807 sp > 00007ffc819cbdb0 error 4 in libjulia.so[7f03361ed000+eb000] > > which I think it is the result of bad memory handling.
Can you reproduce this sporadic error on multiple machines with different hardware? If not, does memtest86+ pass on that machine? What are the hardware specs of the machine(s)? > I think part of the problem is that parallel builds are enabled no > matter the number of CPUs I have. In the logs, I see this: > > env JULIA_CPU_CORES=2 [...] > > and in fact debian/rules has this: > > # Set number of parallel workers for tests > ifneq (,$(filter parallel=%,$(DEB_BUILD_OPTIONS))) > TESTS_ENV += JULIA_CPU_CORES=$(patsubst parallel=%,%,$(filter > parallel=%,$(DEB_BUILD_OPTIONS))) > else ifeq (,$(filter parallel,$(DEB_BUILD_OPTIONS))) > TESTS_ENV += JULIA_CPU_CORES=2 > endif > > I'm not sure if this is policy compliant or not, but I think the > number of cores to be used should be limited in either case by the number > of actual CPUs available, as there is little point in using parallel > builds (or parallel workers for test suite) when you have only one CPU. The tests are run with a minimum of 2 processes regardless of the number of available CPUs. This is *required* since one of the processes supervises and restarts the other process when the memory limit per process is exceeded. As Graham has written earlier, julia builds fine even on a system with only 1 CPU and 1 GB. Either your hardware is defect, or you have a special setup and we need to find out what is special. Regards, Peter
