[Bug 219399] System panics after several hours of 14-threads-compilation orgies using poudriere on AMD Ryzen...

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=219399

--- Comment #67 from Nils Beyer <nbe at renzel.net> ---
Perhaps interesting to you - here is another Ryzen stress test:

    https://github.com/hayamdk/ryzen_segv_test

I had to patch it with:
----------------------------- SNIP -------------------------------------
diff --git a/ryzen_segv_test.c b/ryzen_segv_test.c
index 8d64215..74d8530 100644
--- a/ryzen_segv_test.c
+++ b/ryzen_segv_test.c
@@ -323,11 +323,6 @@ int main(int argc, const char *argv[])
 {
        int64_t loops;
        pthread_t t1, t2, t3;
-#ifdef _MSC_VER
-#else
-       cpu_set_t cpuset;
-       int cpu;
-#endif
        pid_t pid = getpid();

        if(argc > 1) {
@@ -352,15 +347,6 @@ int main(int argc, const char *argv[])
        pthread_create(&t2, NULL, (void*)threadx, (void*)1);
        pthread_create(&t3, NULL, (void*)threadx, NULL);

-#ifdef _MSC_VER
-#else
-       cpu = random() % n_cpus;
-       CPU_ZERO(&cpuset);
-       CPU_SET(cpu, &cpuset);
-       sched_setaffinity(pid, sizeof(cpu_set_t), &cpuset);
-       fprintf(stderr, "PID:%d CPU:%d\n", (int)pid, cpu);
-#endif
-       
        pthread_join(t1, NULL);
        pthread_join(t2, NULL);
        pthread_join(t3, NULL);
----------------------------- SNIP -------------------------------------

compiled it and executed:

    ./run.sh 16 2500000

It generates segmentation faults rather quickly (< 1 min).

As a side note: on my Intel E3-1220v3 system there are no segfaults at all
(using the same command line)...

-- 
You are receiving this mail because:
You are the assignee for the bug.