You are not logged in.
How long did ur succesful test took? Using ram is okay as its load faster than my basic ssd.
Last edited by simplisticways (2017-08-31 17:18:32)
Offline
How long did ur succesful test took? Using ram is okay as its load faster than my basic ssd.
To finish compiling? I don't remember... I think the first thread failed after around 3 hours, and all of the other threads finished (some passed, some failed) less than an hour after that?
I'll run it again now.
EDIT: "[loop-3] TIME TO FAIL: 11414 s" (3.17 hours) I had to pause the test because my kids INSISTED I play Super Metroid... I'll run it some more tonight.
Last edited by drcouzelis (2017-09-01 00:24:56)
Offline
Here's my results, after running it for about 6 hours. I had to disable "USE_RAMDISK". It was my second attempt. The first caused my computer to die a horrible miserable death with error after error of "NMI watchdog: BUG: soft lockup - CPU#6 stuck for 22s!"
Anyway, the results:
[loop-0] Fri Sep 1 23:48:05 EDT 2017 start 0
[loop-0] Fri Sep 1 23:50:32 EDT 2017 build failed
[loop-0] TIME TO FAIL: 147 s
[loop-6] Fri Sep 1 23:48:11 EDT 2017 start 0
[loop-6] Sat Sep 2 02:43:02 EDT 2017 build failed
[loop-6] TIME TO FAIL: 10497 s
[loop-5] Fri Sep 1 23:48:10 EDT 2017 start 0
[loop-5] Sat Sep 2 03:18:32 EDT 2017 build failed
[loop-5] TIME TO FAIL: 12627 s
[loop-10] Fri Sep 1 23:48:15 EDT 2017 start 0
[loop-10] Sat Sep 2 03:25:12 EDT 2017 start 1
[loop-11] Fri Sep 1 23:48:16 EDT 2017 start 0
[loop-11] Sat Sep 2 03:26:56 EDT 2017 start 1
[loop-2] Fri Sep 1 23:48:07 EDT 2017 start 0
[loop-2] Sat Sep 2 03:26:58 EDT 2017 start 1
[loop-1] Fri Sep 1 23:48:06 EDT 2017 start 0
[loop-1] Sat Sep 2 03:25:05 EDT 2017 start 1
[loop-4] Fri Sep 1 23:48:09 EDT 2017 start 0
[loop-4] Sat Sep 2 03:27:05 EDT 2017 start 1
[loop-3] Fri Sep 1 23:48:08 EDT 2017 start 0
[loop-3] Sat Sep 2 03:25:33 EDT 2017 start 1
[loop-9] Fri Sep 1 23:48:14 EDT 2017 start 0
[loop-9] Sat Sep 2 03:27:13 EDT 2017 start 1
[loop-7] Fri Sep 1 23:48:12 EDT 2017 start 0
[loop-7] Sat Sep 2 03:27:39 EDT 2017 start 1
[loop-8] Fri Sep 1 23:48:13 EDT 2017 start 0
[loop-8] Sat Sep 2 03:27:20 EDT 2017 start 1
As you can see, the first compile failed a little after two minutes. Loops 0, 5, and 6 failed. The other 9 appears to each have successfully finished at least one build of GCC.
Offline
Thanks! I should enable my watchdog again.. didn't have time to test my computer that much but this weekend should have time.
So if I get pass them all without failure.. then my ryzen is okay.. well lets find out
Offline
.
Last edited by felipe (2017-09-11 18:04:32)
Offline
[KERN] Sep 16 08:54:12 Archer kernel: sh[21012]: segfault at 7fa2691db670 ip 00007fa268f2d490 sp 00007ffe734dec98 error 4 in libc-2.26.so[7fa268e97000+1ae000]
This doesnt look good.
Last edited by simplisticways (2017-09-16 05:58:27)
Offline
[loop-1] Sat Sep 16 08:51:55 EEST 2017 start 0
[loop-2] Sat Sep 16 08:51:56 EEST 2017 start 0
[loop-3] Sat Sep 16 08:51:57 EEST 2017 start 0
[loop-4] Sat Sep 16 08:51:58 EEST 2017 start 0
[loop-5] Sat Sep 16 08:51:59 EEST 2017 start 0
[loop-6] Sat Sep 16 08:52:00 EEST 2017 start 0
[loop-7] Sat Sep 16 08:52:01 EEST 2017 start 0
[loop-8] Sat Sep 16 08:52:02 EEST 2017 start 0
[loop-9] Sat Sep 16 08:52:03 EEST 2017 start 0
[loop-10] Sat Sep 16 08:52:04 EEST 2017 start 0
[loop-11] Sat Sep 16 08:52:05 EEST 2017 start 0
[loop-11] Sat Sep 16 08:54:13 EEST 2017 build failed
[loop-11] TIME TO FAIL: 139 s
[KERN] Sep 16 08:54:12 Archer kernel: sh[21012]: segfault at 7fa2691db670 ip 00007fa268f2d490 sp 00007ffe734dec98 error 4 in libc-2.26.so[7fa268e97000+1ae000]
[loop-0] Sat Sep 16 09:12:20 EEST 2017 build failed
[loop-0] TIME TO FAIL: 1226 s
[loop-3] Sat Sep 16 09:12:25 EEST 2017 build failed
[loop-3] TIME TO FAIL: 1231 s
[loop-1] Sat Sep 16 09:12:28 EEST 2017 build failed
[loop-1] TIME TO FAIL: 1234 s
[loop-6] Sat Sep 16 09:12:31 EEST 2017 build failed
[loop-6] TIME TO FAIL: 1237 s
[loop-2] Sat Sep 16 09:12:32 EEST 2017 build failed
[loop-2] TIME TO FAIL: 1238 s
[loop-4] Sat Sep 16 09:12:36 EEST 2017 build failed
[loop-4] TIME TO FAIL: 1242 s
[loop-10] Sat Sep 16 09:12:36 EEST 2017 build failed
[loop-10] TIME TO FAIL: 1242 s
[loop-5] Sat Sep 16 09:12:37 EEST 2017 build failed
[loop-5] TIME TO FAIL: 1243 s
[loop-7] Sat Sep 16 09:12:38 EEST 2017 build failed
[loop-7] TIME TO FAIL: 1244 s
[loop-9] Sat Sep 16 09:12:38 EEST 2017 build failed
[loop-9] TIME TO FAIL: 1244 s
[loop-8] Sat Sep 16 09:12:41 EEST 2017 build failed
[loop-8] TIME TO FAIL: 1247 s
Offline
[KERN] Sep 16 08:54:12 Archer kernel: sh[21012]: segfault at 7fa2691db670 ip 00007fa268f2d490 sp 00007ffe734dec98 error 4 in libc-2.26.so[7fa268e97000+1ae000]
This doesnt look good.
By 99% chance OOM - add some or more swap space.
Online
16gb ram and 16b swap file memory so I think I should be okay?
Last edited by simplisticways (2017-09-16 06:44:30)
Offline
Does the benchmark provide more informative logs?
Standard error: you did not forget to swapon, did you? ;-)
Online
I did use swapon BUT now trying the
./kill-ryzen.sh 4 4
to be safe.
Offline
example from loop-0 build.log the end part
make[3]: *** [/mnt/ramdisk/workdir/gcc-7.1.0/libgcc/shared-object.mk:14: unwind-dw2.o] error 1
make[3]: *** waiting for unfinished jobs....
make[3]: leaving directory '/mnt/ramdisk/workdir/buildloop.d/loop-0/x86_64-pc-linux-gnu/libgcc'
make[2]: *** [makefile:21950: all-stage1-target-libgcc] error 2
make[2]: leaving directory '/mnt/ramdisk/workdir/buildloop.d/loop-0'
make[1]: *** [makefile:27079: stage1-bubble] error 2
make[1]: leaving directory '/mnt/ramdisk/workdir/buildloop.d/loop-0'
make: *** [makefile:942: all] error 2
Last edited by simplisticways (2017-09-16 07:04:32)
Offline
https://bbs.archlinux.org/viewtopic.php … 5#p1733865
I had to disable "USE_RAMDISK"
Online
Wow so it really needs to run only by swap? That would be very slow... but cant be helped then..
Offline
I think the problem is that also all sources are placed into RAM what might take a good share of the available RAM before any build process. In doubt, track memory usage to see whether you're running out.
Online
After a while of usage.. my computer still hangs. But this time I somehow managed to logout.. and then screen was full of strange things.
Such as : RIP kmem_cache_alloc
Anyone know what does it mean?
Last edited by simplisticways (2017-10-02 14:05:34)
Offline
No, I'm afraid I don't know what that means.
Unfortunately, when there is a CPU (or RAM or PSU) issue, sometimes the only symptom is "strange things happen on my computer", and the only real way I know of to figure out the problem is by testing the parts with other hardware that is known to be good.
When I upgraded to Ryzen, the only new hardware I got was the CPU, motherboard, and RAM. Everything has been working perfectly for over eight years. So when I saw that my computer was randomly crashing, I knew that it COULD have been because of my PSU, but it was kind of unlikely that it coincidentally crapped out the exact same time I installed the new computer parts.
Ryzen CPUs have known issues under Linux (segmentation faults when compiling, and reboots). That's why people, like me, are doing an RMA with AMD. It's a hassle, but it's free and appears to help.
Have you done a memory stress test, for example with MemTest86?
Offline
That's a generic issue around memory allocation - there should be some call trace nearby, indicating what caused it (where the interesting question is whether this is a deterministic bug, ie. the cause is always the same, or random failure)
Online
I disabled "USE_RAMDISK" but it still seemed to use ram and not only swap.. also every time every single build fails around the same time.. so I find it hard to prove that its really the same problem. This test like never goes on for longer than half hour in my box.
So I started the RMA process.
Just wondering that if the Product Part Number and Product Serial Number are those that are printed in my cpu.. as I have thrown the boxes to trash months ago.
Offline
I had a great experience doing an RMA with AMD to replace a faulty Ryzen 5 1600.
Old CPU
SN: 9R67974070167
PART: YD1600BBM6IAE
When attempting to compile 12 instances of GCC using all twelve cores, a segmentation fault occurred on at least one build before finishing.
When left running, my computer would randomly reboot or crash every 2 to 6 days.
New CPU
SN: 9GY9258U70272
PART: YD1600BBAEBOX
When attempting to compile 12 instances of GCC using all twelve cores, it successfully compiled all 12 instances for 3 loops, for a total of about 12 hours of compiling.
When left running, my computer was up for a complete 10 days before I manually restarted it.
I have no complaints!
...to answer the question above, if you don't have your original box, the only way to get the serial number is by removing the heat sink and the thermal paste. But as soon as I did that and sent the email, AMD asked me to mail them the faulty CPU anyway, so it wasn't a big deal. My computer was down for about two weeks for the whole RMA process.
EDIT: Less than a day after I posted this, my computer froze. Twice. I disabled "C States" in BIOS again, so I'll see if that help. The compile issue is still good, though.
Last edited by drcouzelis (2017-10-15 19:06:06)
Offline
EDIT: Less than a day after I posted this, my computer froze. Twice. big_smile I disabled "C States" in BIOS again, so I'll see if that help. The compile issue is still good, though. smile
Amd Sucks !
Although I haven't see freeze/reboots in the last weeks, fingers cross
Have you tried with a new memory?
Last edited by felipe (2017-10-28 17:34:47)
Offline
Although I haven't see freeze/reboots in the last weeks, fingers cross
Have you tried with a new memory?
No, I don't have other memory to try.
Amd Sucks !
Awww, overall I've been very happy with my new computer! Although I did learn an important lesson about living life on the bleeding edge...
Offline
I'm happy too, but when it freeze I regret of the purchase and makes me want to sell my AMD Ryzen 1600/b350 plus prime and get an Intel Core i7 or i5 6600k, perhaps the best option is to wait for the new generation of AMD cpu, like AMD FX 6100(bulldozer) to AMD FX 6300(piledriver) or something...
Last edited by felipe (2017-11-06 13:48:15)
Offline
Other thing that I noted is freezes usually happen after updating the kernel... no immediately but usually in the same boot... anyways I'm about to get a I7 and sell this processor
Last edited by felipe (2017-11-07 18:46:03)
Offline
Other thing that I noted is freezes usually happen after updating the kernel... no immediately but usually in the same boot... anyways I'm about to get a I7 and sell this processor
That sounds like a fine decision.
...but have you already done an RMA? There are known issues, and if you're going to take out the CPU anyway, consider doing an RMA with AMD. It was a pretty painless process, and you'll get a new CPU in a box (with a new fan, if that's what you had), so I guess you could sell that... I really like and support AMD, but I also think you should have a functional CPU that you paid for.
Offline