• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** The Official Alder Lake owners thread ***

Vimes said:
@Kelt - have you checked your vcore voltage when it is running cinebench R23...? Not what you have dialled in, but actually running at when at full load overclock. The reason I ask is that when I had the MSI 690 Pro I needed to use adaptive voltage with a negative offset, or it would scale up quite high with just a manual vcore setting. I noted your Cinebench R23 screenshot in that other thread, and you are not far off thermal throttling, at 95c.

Just noting what you have written about entering 1.26v for your vcore.
Click to expand...

Just an update, did some delving, and found that the AIO and case fans (they're all connected to the Corsair Core Commander) had been set to quiet by default in Icue.

So set them to the next one up (balanced), down volted a tad more to 1.24V, ran 20 minutes of R23 and the CPU package temp peaked at 91c, so a little cooler, core temps never went above 90c.

Peak CPU watts was 212W, so happy with that.

Also, you might like to know, the latest BIOS from MSI for this board, which apparently helps greatly with DRAM overclocking, has also fixed the 99.9 bus clock, it now reads 100, so the CPU clocks now read 5100 and 4000 within HWINFO. :)
 
My biggest headache is the huge differences in temperature between cores.

When I push to 5.1GHz I feel my old D14 Noctua should be fine even in C23.

Temps though are a massive 10 °C difference.

Core 5 at 100°C with core 3 at around 97°C
Then two cores about 5 degrees less and 2 cores 10 degrees less!

EDIT: Also been playing with ram and not convinced (as always) that it is worth the effort.
XMP 3200C14 vs 4000C17 (only run 66% memtest to far as didn't realise how long that test took and needed the computer)
 
Last edited:
Kelt said:
Just an update, did some delving, and found that the AIO and case fans (they're all connected to the Corsair Core Commander) had been set to quiet by default in Icue.

So set them to the next one up (balanced), down volted a tad more to 1.24V, ran 20 minutes of R23 and the CPU package temp peaked at 91c, so a little cooler, core temps never went above 90c.

Peak CPU watts was 212W, so happy with that.

Also, you might like to know, the latest BIOS from MSI for this board, which apparently helps greatly with DRAM overclocking, has also fixed the 99.9 bus clock, it now reads 100, so the CPU clocks now read 5100 and 4000 within HWINFO. :)
Click to expand...


That's good :)

You might want to think about trying out y-cruncher and perhaps OCCT as mentioned previously. Using those for stability testing and then C R23 for benchmarking seems reasonable to me. I have noted that my vcore voltage, obviously linked to temps, is no where near the same, meaning low, as it is when stability testing using y-cruncher.
chances are, going off your 90c with 1.24v you will find that you will throttle when using it - good to have HWinfo open whilst running. I would be very surprised if you could keep the 1.24v vcore as well.

What memory are you using and have you seen our posts in regards to the Gear 1 and 2 settings...?

Great result with the latest BIOS - I know that when I was emailing them in regards to the sleep / wake issues I also made them aware of the "issues" with the BCLK reporting of a 99.951Mhz bus speed rather than a true 100Mhz. Good for them in resolving that. Most excellent board you have got there.
 
=XDC=FluphyBunny said:
My biggest headache is the huge differences in temperature between cores.

When I push to 5.1GHz I feel my old D14 Noctua should be fine even in C23.

Temps though are a massive 10 °C difference.

Core 5 at 100°C with core 3 at around 97°C
Then two cores about 5 degrees less and 2 cores 10 degrees less!

EDIT: Also been playing with ram and not convinced (as always) that it is worth the effort.
XMP 3200C14 vs 4000C17 (only run 66% memtest to far as didn't realise how long that test took and needed the computer)
Click to expand...

I think there is a difference you have noted from should to is..! Assuming that you have a decent paste and fan settings perhaps it illustrates when the cores are at almost full load (R23 isn't that demanding) perhaps a AIO or other exotic cooling might well be needed. But maybe running such benchmarks or tests don't typify your day to day use and so do not matter as much..? As it is you understand you would be throttling for one cores use. IIRC core 5 seems the hottest for me also when running Cinebench................



but at a max of 84c. This is an older screenshot tho.

Do you need the vcore you are using at full load...?
 
Last edited:
Managed to get around to doing some gaming and run some CPU benchmarks on my system and checking temps.

The 12600k with no manual overclocking, just allowing the Tomahawk to do its own thing results in HWInfo reporting 5Ghz all Pcores and 4Ghz on the Ecores with 1.3v and temps whilst gaming are around 50 degrees and pulling around 70w whilst gaming.

On Cinebench r23, it hit a very respectable 17774 score which puts it faster than a 16 core/32 thread 1950x Threadripper and on HWInfo it was hitting around 70 degrees on the same clocks and pulling around 150w. Plenty of headroom left in this chip running the Arctic Cooling Freezer ii 240mm.

The exhaust from the radiator in the roof of the Lian Li Lancool ii 205 mesh barely feels warm during cinebench runs and the system fans ramp up to a gentle whirr. More or less silent under gaming. Massively impressed.
 
Surprising at time what a 12700k can pull at max draw, well max so far for me.....

max-draw.png


who needs to do star jumps, lol.
 
Vimes said:
I think there is a difference you have noted from should to is..! Assuming that you have a decent paste and fan settings perhaps it illustrates when the cores are at almost full load (R23 isn't that demanding) perhaps a AIO or other exotic cooling might well be needed. But maybe running such benchmarks or tests don't typify your day to day use and so do not matter as much..? As it is you understand you would be throttling for one cores use. IIRC core 5 seems the hottest for me also when running Cinebench................



but at a max of 84c. This is an older screenshot tho.

Do you need the vcore you are using at full load...?
Click to expand...


Only use C23 as a means to an end. c23 is FAR more demanding than any game will ever be. In this case just to show the terrible core to core temperature gap.
For stability I have run heaven and other than that just play games as that is what the system is for. Will try ycruncher.

Memory is g-skill F4-3200C14-8GFX 14-14-14-34.
OC: Gear 1, 4000MHz 17-17-17-34 1.380 volts. ERFC 350, EREFI 50000 (new to these and still playing)
Rough AIDA64 runs to see where I was:

3200C14
Read: 50301
Write: 47676
Copy: 48539
Latency: 62ns

4000C17
Read: 65212
Write: 61551
Copy: 60976
Latency: 54.5

Need to run full memtest on the 4000c17 although gaming day to day it was fine.

The issue was, I didn't see any difference from just clicking XMP, which is what I am now running.
 
Last edited:
Ayahuasca said:
I've got a 12700k arriving tomorrow but not sure I'll get the system built.

@mrk do you still use a 4000D case? If so do you have an AIO mounted on top?
Click to expand...

I have the 5000D Airflow not 4000D but they are basically the same just different on size scaling. My AIO is front mounted not at the top. Best placement for cooling performance and two birds with 1 stone killed this way, intake for case, and cold air food through the radiator.

pc_RTX3080Ti_backangle_.jpg
 
mrk said:
I have the 5000D Airflow not 4000D but they are basically the same just different on size scaling. My AIO is front mounted not at the top. Best placement for cooling performance and two birds with 1 stone killed this way, intake for case, and cold air food through the radiator.
Click to expand...

Cheers, I think I'll have to do the same in terms of front mounting. I bought two 140mm fans for the front intakes though, might be able to use them on the top as exhausts but that would likely mean negative air pressure.
 
Should not be a problem really. That sort of config is fairly standard. I have 1x rear exhaust 120mm and 1x top exhaust 140mm. Idle temps are 27 dgerees.
 
Tubes bottom or top no right or wrong just make sure that if your pump is on the cold plate then pump is not the highest point of the loop so that any air bubbles don't travel up and sit at the pump.

J2C did a great video on it:

 
Just to state that the reliability history tool in Windows has been looking much better since those SA and RAM BIOS changes. The odd one here and there are nVidia driver updates and stuff but everything is on the upward now hah.

A little known but very useful tool to check up on every now and then.

40LCKnr.jpg
 
You must log in or register to reply here.
Back
Top Bottom