Forums first comp: CPUz Benchmark

[Mr. Fox]

8 hours ago, ArchStanton said:

I am also sorry to hear that.  Forgive my crass attempt at levity earlier.

I did not take offense.  Someone a lot smarter than me once said, "A merry heart does good, like medicine, But a broken spirit dries the bones." I agree with that.

 

Edited February 20, 2022 by Mr. Fox

[J7SC_Orion]

1 hour ago, Mr. Fox said:

Thank you.

I was hoping the CPU might be OK. The circumstances under which it stopped working certainly would give the impression the CPU wouldn't be harmed. But, the fault follows the CPU to another computer. It won't boot in either system, and any CPU boots on the system where it failed.

 

I tried all three BIOS positions. Swapped out RAM sticks. Cleared CMOS. I recovery flashed it as you would with no RAM or CPU. Finally, I moved the CPU to the work setup. It doesn't boot in the other computer either. In the Strix (no LCD codes, just lights) it fails POST with the VGA light on. I have a spare 10850K and 10900KF and both boot fine on the Dark mobo. (Well, I HAD as spare 10900KF, but now it is in the Strix board and the CPU from that one is back in the Dark again.)

 

The highest voltage I tried was 1.550V (which I have used with all of my CPUs on chilled water) but I could not do 5.7 stable for anything other than CPU-Z benchmark and gave up. VCCSA and VCCIO are manually set and I try to avoid "auto" for voltage, amp and power settings. I also avoid using adaptive voltage. I run core voltage static as well. I generally only leave boot voltage values on Auto.

 

 

...

 

...fortunately, I got immunity from the Ravens, at least until the meteors have wiped out the Raptors 

[Storm-Chaser]

Updates, slo mo first

[Storm-Chaser]

On 19/02/2022 at 14:01, Mr. Fox said:

Here is a new one when you get to it. 9454.4

So the chiller is back on. Tried to go for 5722MHz in the middle of the night after getting everything squared away and it crashed half way through the bench. Im going to blame less than ideal temps but we should have a cold snap later in the week. If I can get it down to 0*F I think I might have a fighting chance. 

[bonami2]

I think i got a good chip!   Only 1.3v 4.0 ring 4.1 e-core 5.1 p-core

Need to tune my ram now.

 

I wonder if 1.35-1.4v would improve my max clock. But my cooling cant handle any avx stress test already for some reason..

 

 

[Storm-Chaser]

[Avacado]

9 hours ago, bonami2 said:

I think i got a good chip!   Only 1.3v 4.0 ring 4.1 e-core 5.1 p-core

Need to tune my ram now.

 

I wonder if 1.35-1.4v would improve my max clock. But my cooling cant handle any avx stress test already for some reason..

 

 

Are you on an asus main board? If so, what is the SP rating of your chip?

[bonami2]

On 28/02/2022 at 15:13, Avacado said:

Are you on an asus main board? If so, what is the SP rating of your chip?

 

No it an msi z690a Pro ddr4

I dont think msi offer an SP rating?

 

Edit: Cpu stable at 1.3v

 

Edited March 12, 2022 by bonami2

[Avacado]

1 hour ago, bonami2 said:

 

No it an msi z690a Pro ddr4

I dont think msi offer an SP rating?

 

Ran 12h folding@home on 20 thread and when i used the computer about 5 min after it rebooted so maybe not super stable

 

I run an all P core at 5.3Ghz and an all E core at 4.0Ghz at 1.43v. Can still hit 70-78c on full load under water. 

[bonami2]

52 minutes ago, Avacado said:

I run an all P core at 5.3Ghz and an all E core at 4.0Ghz at 1.43v. Can still hit 70-78c on full load under water. 

 

2x420mm ?

 

am using a inwin 360mm i dont even know how to stress test those cpu because it temp throttle on most of them except occt large data.

 

Wondering if i should go custom loop one days. Always dreamed about it but am lazy

 

My phantek enthoo 719 can handle about any build.

 

 

[Mr. Fox]

OK, here is a first try with the new hardware.

[J7SC_Orion]

21 minutes ago, Mr. Fox said:

OK, here is a first try with the new hardware.

 

...Nice ! First time I have seen CPUz single > 900 

[ArchStanton]

1 hour ago, Mr. Fox said:

OK, here is a first try with the new hardware.

What kind of full load PPT are we talking at those settings?

[Mr. Fox]

1 hour ago, J7SC_Orion said:

 

...Nice ! First time I have seen CPUz single > 900 

Thank you. Seems like a solid silicon sample.

26 minutes ago, ArchStanton said:

What kind of full load PPT are we talking at those settings?

I did not check in HWINFO64. In Cinebench R23 at 5.4GHz on P-cores, about 375W. Probably not quite that high in CPU-Z at 5.5GHz. 

 

https://hwbot.org/submission/4943498_mr._fox_cinebench___r23_multi_core_with_benchmate_core_i9_12900k_30908_cb/

[ArchStanton]

7 minutes ago, Mr. Fox said:

about 375W

Woof!

[Mr. Fox]

Looks like I got a good lottery ticket on the RAM and CPU IMC as well. I am only just now starting to tweak things. I have a ton of stuff to learn about Z690 overclocking and DDR5. Like going from Intel to Ryzen... or speaking a different language. These read/write/copy speeds remind me of my monster 7980XE benching days.

 

https://hwbot.org/submission/4943746_mr._fox_wprime___32m_with_benchmate_core_i9_12900k_1sec_532ms

https://hwbot.org/submission/4943767_mr._fox_wprime___1024m_with_benchmate_core_i9_12900k_37sec_90ms

Edited March 4, 2022 by Mr. Fox

[Avacado]

14 hours ago, Mr. Fox said:

Looks like I got a good lottery ticket on the RAM and CPU IMC as well. I am only just now starting to tweak things. I have a ton of stuff to learn about Z690 overclocking and DDR5. Like going from Intel to Ryzen... or speaking a different language. These read/write/copy speeds remind me of my monster 7980XE benching days.

 

https://hwbot.org/submission/4943746_mr._fox_wprime___32m_with_benchmate_core_i9_12900k_1sec_532ms

https://hwbot.org/submission/4943767_mr._fox_wprime___1024m_with_benchmate_core_i9_12900k_37sec_90ms

At 1.51v, be careful before you degrade that chip before it's lived any of it's life. Believe me, it screams out of the box, but you keep running more than 1.45v on her and those scores won't last long. Show me a 1b Y-Cruncher at those speeds, bet it will BSOD. You did score a good chip though. Was that Cinebench a 5.4GHz all core?

 

Lastly, please, please disable your E cores if you are going to run the voltage that high. You might not see an immediate degredation on your P cores, but the E cores CAN NOT handle that much voltage. 

 

I have played with the 12900KF for awhile now. If I could do it all over again, I would set 1.43v Core. Sync all core 53/40 Uncore 47/50 Max Auto. I'm really trying to save you from going the route I went while "Pushing it to the limit" 

Edited March 4, 2022 by Avacado

[Mr. Fox]

5 hours ago, Avacado said:

At 1.51v, be careful before you degrade that chip before it's lived any of it's life. Believe me, it screams out of the box, but you keep running more than 1.45v on her and those scores won't last long. Show me a 1b Y-Cruncher at those speeds, bet it will BSOD. You did score a good chip though. Was that Cinebench a 5.4GHz all core?

 

Lastly, please, please disable your E cores if you are going to run the voltage that high. You might not see an immediate degredation on your P cores, but the E cores CAN NOT handle that much voltage. 

 

I have played with the 12900KF for awhile now. If I could do it all over again, I would set 1.43v Core. Sync all core 53/40 Uncore 47/50 Max Auto. I'm really trying to save you from going the route I went while "Pushing it to the limit" 

Thank you for your words of wisdom and concern. I sincerely appreciate them. I am not sure why Benchmate showed 1.51V unless it was an overshoot issue, and I will monitor that closer. I had the BIOS set at 1.475V override. I definitely don't want to kill this excellent CPU sample. It runs 52/41 on cores and 43x cache at 1.275V and 53/41 at 1.295V.

 

My bigger concern right now relating to my daily driver clock speeds and not benching is the PCH temperature. Doing nothing (just idle) the PCH hovers around 80-85°C and I have seen it at 90°C. I have all five NVMe slots filled. If I remove a couple of them the temperatures drop about 5-8°C. I guess I am going to have to do some kind of hardware modification to keep the PCH from committing suicide. The PCH voltage is set to 0.850V (BIOS default value) but instead of "Auto" I switched it to override in an effort to try to control the high temps, since I don't know that "Auto" is not higher than default.

 

Edited March 4, 2022 by Mr. Fox
comments about PCH temperature

[Avacado]

39 minutes ago, Mr. Fox said:

Thank you for your words of wisdom and concern. I sincerely appreciate them. I am not sure why Benchmate showed 1.51V unless it was an overshoot issue, and I will monitor that closer. I had the BIOS set at 1.475V override. I definitely don't want to kill this excellent CPU sample. It runs 52/41 on cores and 43x cache at 1.275V and 53/41 at 1.295V.

 

My bigger concern right now relating to my daily driver clock speeds and not benching is the PCH temperature. Doing nothing (just idle) the PCH hovers around 80-85°C and I have seen it at 90°C. I have all five NVMe slots filled. If I remove a couple of them the temperatures drop about 5-8°C. I guess I am going to have to do some kind of hardware modification to keep the PCH from committing suicide. The PCH voltage is set to 0.850V (BIOS default value) but instead of "Auto" I switched it to override in an effort to try to control the high temps, since I don't know that "Auto" is not higher than default.

 

I would cream my pants If I could run my chip at 53/41 @1.295. Is that fully stable at 100% load with AVX (0 Offset)? If so, keep that setting and feel proud. You might even be able to inch your Uncore to 45. Even with Manual voltage set I would recommend setting an Auto voltage limit for the same for added protection. What is your LLC?

 

You are going to need active cooling on your PCH it would seem. That is a shitload of drives. Is your GPU running at 8x with all of those in play?

[Mr. Fox]

3 hours ago, Avacado said:

I would cream my pants If I could run my chip at 53/41 @1.295. Is that fully stable at 100% load with AVX (0 Offset)? If so, keep that setting and feel proud. You might even be able to inch your Uncore to 45. Even with Manual voltage set I would recommend setting an Auto voltage limit for the same for added protection. What is your LLC?

 

You are going to need active cooling on your PCH it would seem. That is a shitload of drives. Is your GPU running at 8x with all of those in play?

Yes that is Cinebench stable with 0 AVX offset. This is my first MSI motherboard and I've not become intimately familiar with all of the settings as I have with Asus and EVGA firmware. The LLC is set on 6 but that's not the same as 6 on an Asus board. It's probably equal to about 5 on Asus. But I'm just guessing LOL. EVGA does LLC settings the best by expressing the values as a percentage rather than a meaningless number. 

 

This firmware seems like it is geared more toward tickling the aesthetic focus of a gamer than a no-nonsense overclocker the way things are laid out. And of course, as usual, some of the terminology is the MSI version of something Asus, EVGA, or another brand call something else. I wish all brands would adhere to stricter nomenclature standards.

 

My video card still runs at x16 even with all of the m.2 slots filled and even with a sixth NVMe SSD inserted into the open X4 slot on an add-in card. I do have the onboard Wi-Fi/BT and second NIC disabled in the BIOS since I have no need for any of them and that may be allowing it to run at x16. 

 

It is unfortunate that all of the enthusiast motherboards are using UWP trash control center software. MSI Center is not as bloated and life-sucking as Armory Crate, but it is still rubbish. The UWP version of Mystic Light is every bit as trashy as Armory Create.

[Storm-Chaser]

@Mr. Fox Good call on the MSI motherboard.

 

[bonami2]

How are you cooling those chip?

 

My 12700k

 

I cant even go higher than 1.3v because it already hit 98c on some core while folding.

Using an inwin sr36 360mm

 

And voltage over 1.4v?

 

Back when i had my 4790k over 1.4 would cause degradation pretty fast if i remember well.

[Avacado]

3 hours ago, bonami2 said:

How are you cooling those chip?

 

My 12700k

 

I cant even go higher than 1.3v because it already hit 98c on some core while folding.

Using an inwin sr36 360mm

 

And voltage over 1.4v?

 

Back when i had my 4790k over 1.4 would cause degradation pretty fast if i remember well.

Well, Mines under an Optimus Sig v2 block with 960mm of rad space. Still hits high 70's/80's in certain benches. I do have the rockitcool delid kit that I plan on using at some point. 

Edited March 5, 2022 by Avacado

[Mr. Fox]

6 hours ago, bonami2 said:

How are you cooling those chip?

 

My 12700k

 

I cant even go higher than 1.3v because it already hit 98c on some core while folding.

Using an inwin sr36 360mm

 

And voltage over 1.4v?

 

Back when i had my 4790k over 1.4 would cause degradation pretty fast if i remember well.

I also use an OptimusPC water block. I think they are by far the best. The Signature and Foundation water blocks are built extremely well in addition to working superbly. My radiator is a MORA 360, but I don't use any kind of radiator when I do extensive benching because the radiator conflicts with the efforts of the chiller and prevents the water from getting cold enough. I simply reroute the lines with QDC fittings and bypass the radiator.

 

I'm also planning to delid and go bare die. I have been running bare die since 7th Gen and I wish Intel would offer a no IHS option for their processors. Everything would be so much easier that way. Some people say bare die is dangerous, but laptops have almost always been bare die. Even the mobile Extreme CPUs were.

 

 

Edited March 5, 2022 by Mr. Fox

[J7SC_Orion]

4 hours ago, Mr. Fox said:

I also use an OptimusPC water block. I think they are by far the best. The Signature and Foundation water blocks are built extremely well in addition to working superbly. My radiator is a MORA 360, but I don't use any kind of radiator when I do extensive benching because the radiator conflicts with the efforts of the chiller and prevents the water from getting cold enough. I simply reroute the lines with QDC fittings and bypass the radiator.

 

I'm also planning to delid and go bare die. I have been running bare die since 7th Gen and I wish Intel would offer a no IHS option for their processors. Everything would be so much easier that way. Some people say bare die is dangerous, but laptops have almost always been bare die. Even the mobile Extreme CPUs were.

 

 

 

How are you cooling the DDR5 modules ? I've been following several DDR5 threads and it seems that DDR5 is quite sensitive to heat, including the onboard PMIC.