My main system is down!

Message boards : Number crunching : My main system is down!
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · Next

AuthorMessage
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1908688 - Posted: 23 Dec 2017, 22:58:04 UTC - in response to Message 1908646.  

Sorry, I misunderstood you. Lots of issues with the OLED/Aura software from all the posts. One new version breaks something and next version fixes it only to break something else. I am at my limit with suggestions as I don't have an actual TR system. You will have to post questions to the Zenith Extreme forum threads for specific help.


Hi Keith, Definitely appreciate your input on this thread. Now that I have the problem understood, I plan to write up a bug report and post on the Asus forum. My only concern is that those forums are dominated by memory overclockers. I have done work to convince myself memory OC'ing is a waste of time which is why I get 2400 ECC memory recently. The one bug I found on C6H concerning X1/X4 BIOS settings being forgotten has been validated by Asus, but still not fixed. I suspect that very few people will see this new Zenith bug, as what we do here is not the market they are targeting. I think mining makes what we do a little more mainstream, but I doubt anyone is mining on TR4 platform. But I will write it up and submit the bug report and see what happens.
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1908688 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1908720 - Posted: 24 Dec 2017, 1:16:25 UTC - in response to Message 1908688.  

Sorry, I misunderstood you. Lots of issues with the OLED/Aura software from all the posts. One new version breaks something and next version fixes it only to break something else. I am at my limit with suggestions as I don't have an actual TR system. You will have to post questions to the Zenith Extreme forum threads for specific help.


Hi Keith, Definitely appreciate your input on this thread. Now that I have the problem understood, I plan to write up a bug report and post on the Asus forum. My only concern is that those forums are dominated by memory overclockers. I have done work to convince myself memory OC'ing is a waste of time which is why I get 2400 ECC memory recently. The one bug I found on C6H concerning X1/X4 BIOS settings being forgotten has been validated by Asus, but still not fixed. I suspect that very few people will see this new Zenith bug, as what we do here is not the market they are targeting. I think mining makes what we do a little more mainstream, but I doubt anyone is mining on TR4 platform. But I will write it up and submit the bug report and see what happens.

Yes, I agree that most posts are about getting memory to overclock, but there are also posts about peripherals and drives too. Been lots of posts about RAID configurations.

Yes, mining could care less about the support platform as long as it supports the GPU's they mine on.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1908720 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13722
Credit: 208,696,464
RAC: 304
Australia
Message 1908824 - Posted: 24 Dec 2017, 21:00:48 UTC - in response to Message 1908581.  

... First is Zenith Extreme

$1,000 for a single socket motherboard? Ouch!
Grant
Darwin NT
ID: 1908824 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1908828 - Posted: 24 Dec 2017, 21:59:17 UTC - in response to Message 1908824.  

... First is Zenith Extreme

$1,000 for a single socket motherboard? Ouch!

??? ASUS ROG Zenith Extreme X399 mobo sells for around $500.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1908828 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13722
Credit: 208,696,464
RAC: 304
Australia
Message 1908832 - Posted: 24 Dec 2017, 22:39:30 UTC - in response to Message 1908828.  

... First is Zenith Extreme

$1,000 for a single socket motherboard? Ouch!

??? ASUS ROG Zenith Extreme X399 mobo sells for around $500.

Not here in Australia. Around $970.
Grant
Darwin NT
ID: 1908832 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1908833 - Posted: 24 Dec 2017, 22:41:55 UTC - in response to Message 1908828.  

... First is Zenith Extreme

$1,000 for a single socket motherboard? Ouch!

??? ASUS ROG Zenith Extreme X399 mobo sells for around $500.

The cheapest that I could get 1 of them here for is $869.00aud on pre order, but that's way out of my league. ;-)

Cheers.
ID: 1908833 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1908834 - Posted: 24 Dec 2017, 22:45:33 UTC - in response to Message 1908832.  
Last modified: 24 Dec 2017, 22:46:33 UTC

... First is Zenith Extreme

$1,000 for a single socket motherboard? Ouch!

??? ASUS ROG Zenith Extreme X399 mobo sells for around $500.

Not here in Australia. Around $970.

Oh, I just saw the $ and expected US dollars, not Aussie $.

Are the prices of the Threadripper CPUs just as expensive in Aussie dollars?
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1908834 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13722
Credit: 208,696,464
RAC: 304
Australia
Message 1908844 - Posted: 24 Dec 2017, 23:48:09 UTC - in response to Message 1908834.  

Are the prices of the Threadripper CPUs just as expensive in Aussie dollars?

$1,230 & $1,500.
Top of the line i9, close enough to $3,000 to make no difference ($2,960).
Grant
Darwin NT
ID: 1908844 · Report as offensive
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1909801 - Posted: 31 Dec 2017, 10:54:01 UTC

My system is now up and running. Let it run a few days with the similar OC conditions as before: 3.75GHz@1.206v running at 60C and measured Vcore of 1.233v with LLC set to auto and fully loaded with BOINC. So it looks like the OC stability issue was the problem of the previous MB. The problems I had getting the new MB to boot seemed to be related to BIOS and OLED/Aura firmware. Now that I have loaded BIOS 902, everything is working fine. I just took the system down to tune fan profiles and tweak the OC. I know it takes more voltage to be stable at the next CPU multiplier of 37.75, so I decide to tweak BCLK also. I am using BCLK of 101 and CPU ratio of 37.25 for a CPU speed of 3.762GHz. So far it is stable, only problem is lscpu and dmidecode report CPU speed of 3725 MHz, which assumes BLCK of 100. Memory speed is reported as 2424 instead of 2400 which takes into account the change in BCLK, so I assume it is running at the conditions I set in BIOS. Anyone seeing similar issues when OC'ing in Linux?
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1909801 · Report as offensive
Profile Karsten Vinding
Volunteer tester

Send message
Joined: 18 May 99
Posts: 239
Credit: 25,201,931
RAC: 11
Denmark
Message 1909817 - Posted: 31 Dec 2017, 13:35:59 UTC

Nice to hear that your system is back in working order.

I guess your kind of setup really separates the lambs from the goats :)

I am strugling with my setup too.
Off course things break during the holidays, so I'll have to wait even longer for replacements.

Good thing I had a few alternative parts lying around, so that it is at least working, even though not at fast as a few days ago :)
ID: 1909817 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1909854 - Posted: 31 Dec 2017, 17:48:58 UTC - in response to Message 1909801.  
Last modified: 31 Dec 2017, 17:55:58 UTC

Anyone seeing similar issues when OC'ing in Linux?

Don't know. Not running a motherboard with a BCLK generator. lspci and lscpu report what you would expect for nominal clocks. Haven't seen any mention of your issue in the threads owing to the low number of Linux users.

I just used the ratio setting for 39.25 and 39.50 for my two boards. You won't get the real exact readout as the chip has no actual internal way to measure actual clocks. The clock speeds are interpreted from the various internal bits. That is why you see your bus clock reported as 99.8 Mhz in all the test apps. Intel chips measure their actual internal clocks. Ryzen can't.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1909854 · Report as offensive
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1910010 - Posted: 1 Jan 2018, 12:01:53 UTC - in response to Message 1909854.  

Anyone seeing similar issues when OC'ing in Linux?

Don't know. Not running a motherboard with a BCLK generator. lspci and lscpu report what you would expect for nominal clocks. Haven't seen any mention of your issue in the threads owing to the low number of Linux users.

I just used the ratio setting for 39.25 and 39.50 for my two boards. You won't get the real exact readout as the chip has no actual internal way to measure actual clocks. The clock speeds are interpreted from the various internal bits. That is why you see your bus clock reported as 99.8 Mhz in all the test apps. Intel chips measure their actual internal clocks. Ryzen can't.


I found a bug report from ~2009 about this problem. Looks like it is still an issue. I thought it was strange the MCLK was correct even though it is also impacted by BLCK. I had to backoff my OC to just 37.5 mulitplier. Anything more needs more voltage and temps are as high as I am comfortable with.

A side affect of having this issue is that when I changed the OC back to the original settings, the reboot must have happend without memory tuning, so I was actually able to see error correction in action. I would get about 1 correctable error per minute. These can be found in /sys/devices/system/edac/mc/mc*

After rebooting again, the system is back to normal with no ecc corrections all day.
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1910010 · Report as offensive
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1910011 - Posted: 1 Jan 2018, 12:03:52 UTC - in response to Message 1909817.  

Nice to hear that your system is back in working order.

I guess your kind of setup really separates the lambs from the goats :)

I am strugling with my setup too.
Off course things break during the holidays, so I'll have to wait even longer for replacements.

Good thing I had a few alternative parts lying around, so that it is at least working, even though not at fast as a few days ago :)


Yes, I can imagine stock at EK is running low. I think you will be happy with their pump reservoir, but I am kind of an EK fan.
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1910011 · Report as offensive
Profile Karsten Vinding
Volunteer tester

Send message
Joined: 18 May 99
Posts: 239
Credit: 25,201,931
RAC: 11
Denmark
Message 1910015 - Posted: 1 Jan 2018, 12:37:49 UTC - in response to Message 1910011.  
Last modified: 1 Jan 2018, 12:41:32 UTC

I actually ordered the new reservoir/pump at a local retailer, and havent received any info on the expected delivery date yet.

I wont call myself an EK fan as such, but the stuff I use from them is very well made, and has run fine. I have had no complaints.
Their TR waterblock is sadly a little lacking compared to others (https://www.youtube.com/watch?v=jjvqPT2CmqI), but off course only when you really push the system.

I just checked my mails, and the XSPC pump/reservoir that broke down, was ordered just under 4 years ago (46 months). I don't think I can actually complain about that as it was not an expensive kit, but breakdowns are irritating nonetheless.

With the new EK pumps MTBF, I expect a longer lifespan. Its also a much more expensive kit.
ID: 1910015 · Report as offensive
Profile Karsten Vinding
Volunteer tester

Send message
Joined: 18 May 99
Posts: 239
Credit: 25,201,931
RAC: 11
Denmark
Message 1910027 - Posted: 1 Jan 2018, 15:00:43 UTC - in response to Message 1910015.  

I have actually added a cooling kit for the pump to my order, in order to try to increase its longlivety.
ID: 1910027 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1910030 - Posted: 1 Jan 2018, 15:23:37 UTC

I am running an Alphacool VPP755 pump.
Seems to be doing a good job.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1910030 · Report as offensive
Profile Karsten Vinding
Volunteer tester

Send message
Joined: 18 May 99
Posts: 239
Credit: 25,201,931
RAC: 11
Denmark
Message 1910044 - Posted: 1 Jan 2018, 16:33:50 UTC - in response to Message 1910030.  
Last modified: 1 Jan 2018, 16:34:07 UTC

I was actually looking at the Alphacool Eisfach, that uses that exact pump.

It does seem solid enough.

Maybe for the next build :)
ID: 1910044 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1910045 - Posted: 1 Jan 2018, 16:37:09 UTC - in response to Message 1910044.  

I am pleased with it. Full PWM speed control, and a variable control built in if you are not using PWM.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1910045 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1910062 - Posted: 1 Jan 2018, 18:22:14 UTC - in response to Message 1910011.  

Except that the consensus of everyone building and testing Threadripppr, is that the EK TR4 waterblock is crap. Worst performing TR4 waterblock on the market. Microfin structure is too small and doesn't even cover 50% of the active dies of TR. The best TR waterblock is the XSPC RayStorm Neo.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1910062 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1910063 - Posted: 1 Jan 2018, 18:30:53 UTC - in response to Message 1910015.  

I actually ordered the new reservoir/pump at a local retailer, and havent received any info on the expected delivery date yet.

I wont call myself an EK fan as such, but the stuff I use from them is very well made, and has run fine. I have had no complaints.
Their TR waterblock is sadly a little lacking compared to others (https://www.youtube.com/watch?v=jjvqPT2CmqI), but off course only when you really push the system.

I just checked my mails, and the XSPC pump/reservoir that broke down, was ordered just under 4 years ago (46 months). I don't think I can actually complain about that as it was not an expensive kit, but breakdowns are irritating nonetheless.

With the new EK pumps MTBF, I expect a longer lifespan. Its also a much more expensive kit.

If I get 4 years out of my XSPC X4 Photon AX360 AM4 kit I have on Numbskull I will be happy. Appears to be well made and is cooling very well. Paid $220. Which isn't very much more than the AIO solutions I have purchased before. Which last less than 3 months on average.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1910063 · Report as offensive
Previous · 1 · 2 · 3 · Next

Message boards : Number crunching : My main system is down!


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.