cancel
Showing results for 
Search instead for 
Did you mean: 

Drivers & Software

bradc
Adept III

Re: gcc segmentation faults on Ryzen / Linux

bsp2020 wrote:

It looks like I'm running out of disk space. I cleaned up all the temporary files and tried again. But kill-ryzen script fill up 20GB of free space in about 40 minutes. How are you running the script for days when it generates so much files?

I never used the kill-ryzen script. I wrote my own.

[Bash] Brads Ryzen test - Pastebin.com

This just creates a folder in /tmp, untars the files and builds them. Logs get left in /tmp with *-build-fail-date if the build was a failure, otherwise it removes them.

/tmp/result is a file that counts the builds. Every build adds a '.' to the line vs an 'f'  for failure.

I let that run for 4 days without an issue.

I got the impression kill-ryzen was designed to make it easy for people to use livecd and basic builds. I already have established systems so I wanted something that wouldn't keep trying to re-download the source and was simple as possible.

oberbutze
Adept II

Re: gcc segmentation faults on Ryzen / Linux

How much ram do you have? It's recommended to run the kill-ryzen script witz 4 loops with 4 threads.

Copy from readme

Requirements

To successfully finish a round of 16 builds more than 16Gb of RAM are required because GCC is a large software package. However, very often, one of the parallel build fails long before 16Gb of RAM are actually used (in the example, e.g, at about 2.2Gb total RAM usage). You can try it, but it will fail if you run out of memory (should it if you have sufficient swap space?). The use of a compressed ramdisk can also be switched off by replacing "USE_RAMDISK=true" with "USE_RAMDISK=false", but seems to increase the "TIME TO FAIL". Alternatively, smaller software packages could be tried (suggestions welcome).

You can also try to run fewer loops but allow them to use more threads, e.g., 8 loops with 2 threads each.

./kill-ryzen.sh 8 2

However, with only 16Gb RAM, this configuration might still run out of memory. 4 loops with 4 threads each is a save choice on machines with 16Gb RAM.

./kill-ryzen.sh 4 4

bsp2020
Elite

Re: gcc segmentation faults on Ryzen / Linux

I used "USE_RAMDISK=false" because I only have 16GB of RAM in my machine. I thought 20GB of empty disk space would allow me to run it longer... I'm dual booting and did not allocate large disk space to Linux partition. 

I'll try your suggestion.

Thank you.

0 Likes
Anonymous28
Journeyman III

Re: gcc segmentation faults on Ryzen / Linux

Hello.
How long does it take to wait for a return shipment of the processor replacement?
I sent my processor September 12 to the Netherlands, was delivered September 13.
Since then, I have not received either a new processor or any news from AMD. It's been over a week.

0 Likes
oberbutze
Adept II

Re: gcc segmentation faults on Ryzen / Linux

@gutek@

I send my CPU just on day before yours and my replacement arrived today in the Netherlands. I hope they will ship it to me today. Depending if its arriving on saturday it was nearly two weeks.

0 Likes

Re: gcc segmentation faults on Ryzen / Linux

Dammit, thats long....

I hope in your 4.5kg package there is also my replacement

0 Likes
oberbutze
Adept II

Re: gcc segmentation faults on Ryzen / Linux

Mine was only 2.27kg

0 Likes
scorpio810
Adept III

Re: gcc segmentation faults on Ryzen / Linux

_blackbishop_ a écrit:

Dammit, thats long....

I hope in your 4.5kg package there is also my replacement

No, you can wait 10 days or more after your shippement is arrived on AMD warehouse I think they try it, and if you not lucky you can never see the RMAed CPU !

Edit : Don't try RMA at the risk of losing a lot of money if your package gets stuck ... by example if you not a Linux software programmer or a Gentoo guy this problem not concern you. 😉

The gcc segmentation fault is independent from the other problems. If the whole system hangs, this is not necessarily a CPU bug, but could also indicate problems elsewhere.

0 Likes
ttbek
Adept III

Re: gcc segmentation faults on Ryzen / Linux

Guys, you need to chill out. Even if the ticket is closed you can open a

new one and reference the old one. They will check their shipping records

and you will get your RMA CPU there just may be some delays because of

that. In my case I was asked to send them a mail with the tracking number

when I sending the CPU, they did ship mine around that time, total

turnaround time was 6 or 7 days Netherlands <---> Qatar. The package I

sent them was charged as 0.5 kilos to them and the one they sent me was 1

kilo. Anything about a 4.5 kilo package and huge (~1000 for a 1700x, lol,

it is like 3x retail price) border valuation that people have been saying

FedEx is doing just mean FedEx and/or government officials are being

totally corrupt. That isn't AMDs fault and you should appeal that through

whatever the appropriate local channels are. And be cautious, corrupt

people don't like having their corruption pointed out.

0 Likes
scorpio810
Adept III

Re: gcc segmentation faults on Ryzen / Linux

ttbek

Your replacement part has been shipped through Fedex 5650 8494 1671 to our AMD EU warehouse in Netherlands.

www.fedex.com/apps/fedextrack/?action=track&tracknumbers=565084941671&cntry_code=us&wsch=true

Tracking number 565084941671

Weight 9.8 lbs / 4.45 kgs

Delivered To Shipping/Receiving

Total shipment weight 9.8 lbs / 4.45 kgs

0 Likes