Hello all, 

 

I have an unRaid system with an i9-10900K, Gigabyte Z490 Vision G motherboard, 64gb ram and several HDD's. I run several dockers, mainly for media, including Emby, along with a Win11 VM which has a dedicated 2TB Sabrent nvme drive. This VM is off whenever I'm not using it and so it's allocated cores (12) are redundant the rest of the time (if there is a way to have unraid use these cores when the VM is off, that would be a bonus to my more pressing question below). 

 

I don't run it hard and everything usually ticks along fine, except for the temperature levels my CPU and motherboard reach, which when under load, can exceed 100+ degrees. I have ample air cooling, located in the loft, which is like an icebox, however to be extra cautious I have an AIO Kraken X53 arriving tomorrow to replace my Noctua CPU cooler. 

 

I have attached below how I have assigned cpu cores, however I am leaning towards there being an issue there that is causing the excess heat issue due to a banner always appearing at the top of unRaid when I enter the "cpu pinning" page that states "CPU Isolation: A reboot is required to apply changes". I have attached a screenshot of my Syslinux Configuration page as I read elsewhere that this may be the culprit, but I don't see the issue. 

 

Any advice as to where I may be going wrong would be greatly appreciated as I'm getting to the point of not chancing leaving the server running when I'm not at home for fear of the house burning down. 😂

 

 

 

Edited November 29, 20214 yr by BeardedNoir

Can't tell what is running without the diagnostics

6 minutes ago, Squid said:

Can't tell what is running without the diagnostics

 

I knew I had forgotten something. 😁

 

tower-diagnostics-20211128-1525.zip

Based upon the diagnostics, I'd say that the "cpu load" is actually I/O wait where the processor is waiting for data from some device.  The dashboard takes that into consideration, even though it's not "actual" cycles being consumed

 

Quote

Z490 VISION G

Quote

Intel(R) Core(TM) i9-10900K

 

Let me ask you a question.  Have you adjusted the BIOS to actually reflect reality on the PL1 / PL2 states or are you running it on the default settings which is an overclock used by most motherboard manufacturer's in order to artificially inflate their benchmarks for reviews?

 

My Z490 vision G (and most motherboards) shipped with PL settings that dictated that the CPU was allowed to operate at up to 4096 Watts of TDP for an infinite period of time and simply let the CPU thermally throttle down when it's temperature controls went into effect.

 

IE: On the default motherboard settings with my i9-10900 and the stock Intel cooler, the CPU would constantly go into thermal overload.  Set the PL1 / PL2 TDPs and time allowed in the BIOS as specified by Intel.  Otherwise you are actually running an overclock and if your cooling isn't sufficient you're going to see exactly what you're seeing.

 

https://forums.unraid.net/topic/100360-z490-multi-core-enhancement/?tab=comments#comment-926134

 

https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

 

https://www.gamersnexus.net/guides/3389-intel-tdp-investigation-9900k-violating-turbo-duration-z390

 

 

Any many more 

7 minutes ago, Squid said:

Based upon the diagnostics, I'd say that the "cpu load" is actually I/O wait where the processor is waiting for data from some device.  The dashboard takes that into consideration, even though it's not "actual" cycles being consumed

 

 

Let me ask you a question.  Have you adjusted the BIOS to actually reflect reality on the PL1 / PL2 states or are you running it on the default settings which is an overclock used by most motherboard manufacturer's in order to artificially inflate their benchmarks for reviews?

 

My Z490 vision G (and most motherboards) shipped with PL settings that dictated that the CPU was allowed to operate at up to 4096 Watts of TDP for an infinite period of time and simply let the CPU thermally throttle down when it's temperature controls went into effect.

 

IE: On the default motherboard settings with my i9-10900 and the stock Intel cooler, the CPU would constantly go into thermal overload.  Set the PL1 / PL2 TDPs and time allowed in the BIOS as specified by Intel.  Otherwise you are actually running an overclock and if your cooling isn't sufficient you're going to see exactly what you're seeing.

 

https://forums.unraid.net/topic/100360-z490-multi-core-enhancement/?tab=comments#comment-926134

 

https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

 

https://www.gamersnexus.net/guides/3389-intel-tdp-investigation-9900k-violating-turbo-duration-z390

 

 

Any many more 

 

Top man, thank you! 

 

I just have the motherboard on default settings, I've certainly never made any changes to the PL1 / PL2 TDPs. I'll go and watch that video now and make those changes. Much appreciated, thanks. 👍

 

Whilst I've got this thread open, can I ask you (or anyone else passing through) about;

1) Whether the allocated cpu pinning can be released back to the unraid system when the VM is not in use. 

2) Secondly, regarding the banner that always pops up when I enter the "cpu pinning" page, is that ok to ignore, or have I made an error somewhere? 

1 hour ago, BeardedNoir said:

1) Whether the allocated cpu pinning can be released back to the unraid system when the VM is not in use. 

 

Isolated cores can't be returned without a reboot.  Non-isolated cores are free o use by any process

 

34 minutes ago, Squid said:

Isolated cores can't be returned without a reboot.  Non-isolated cores are free o use by any process

 

 

Cheers buddy. 

On 11/28/2021 at 3:41 PM, Squid said:

Based upon the diagnostics, I'd say that the "cpu load" is actually I/O wait where the processor is waiting for data from some device.  The dashboard takes that into consideration, even though it's not "actual" cycles being consumed

 

 

Let me ask you a question.  Have you adjusted the BIOS to actually reflect reality on the PL1 / PL2 states or are you running it on the default settings which is an overclock used by most motherboard manufacturer's in order to artificially inflate their benchmarks for reviews?

 

My Z490 vision G (and most motherboards) shipped with PL settings that dictated that the CPU was allowed to operate at up to 4096 Watts of TDP for an infinite period of time and simply let the CPU thermally throttle down when it's temperature controls went into effect.

 

IE: On the default motherboard settings with my i9-10900 and the stock Intel cooler, the CPU would constantly go into thermal overload.  Set the PL1 / PL2 TDPs and time allowed in the BIOS as specified by Intel.  Otherwise you are actually running an overclock and if your cooling isn't sufficient you're going to see exactly what you're seeing.

 

https://forums.unraid.net/topic/100360-z490-multi-core-enhancement/?tab=comments#comment-926134

 

https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

 

https://www.gamersnexus.net/guides/3389-intel-tdp-investigation-9900k-violating-turbo-duration-z390

 

 

Any many more 

 

Just finished making the switch in the bios mate, changed both from AUTO to 65Watts. System runs great so far, and the highest temp reached is 28 degrees. 👍

TDP for the k version listed as 125W (with some other TDP listed as 95).  I'd try both those settings and see what your temps are

2 hours ago, Squid said:

TDP for the k version listed as 125W (with some other TDP listed as 95).  I'd try both those settings and see what your temps are

 

Cheers mate, I'll have a try with both settings. 👍