• 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.

Ryzen 3900X thread

AgedIdiot said:
Thanks to both LtMatt and Brizzles :)

Been playing Far Cry New dawn - max temp 73c -No problem I trust ?

I am just a little concerned that the board does seem to overdo it ref voltage but I do not know how to settle things down :confused: or if it is possible!!

Am coming from a 5820k system which was much easier to control (then again it only had 6 cores!) and I appreciate that Ryzen is a little different.

Will try with Performance enhancer - level 3 and the power profile you refer to - am not too bothered about all core performance as principally the system is used for gaming and I am not looking for the fastest since ..... but anything which would reduce volts/temps would be appreciated.
Click to expand...

Throttling occurs around 90 degrees+, so your temps are fine.

The voltage looks high when you first start using this chip, it did for me also and i started to panic at first.

Buy once you realise that it is normal behaviour it does bother me now, just an understanding curve.

If you load the cpu you will see that the constant load voltage will drop to a steady ~ 1.20v - 1.25v

I came from an i5 4670K so Ryzen was totally diffirent to anything i was used to.
 
Ryzen is a different beast and completely different, so anything you thought you knew under Intel you can forget with AMD. Everything looks fine, don't worry.

The CPU will use as much voltage as it requires to reach the desired frequency.

My Arctic 360 just arrived but i will put off installing it until tomorrow.

Been a hard day at work so can't be bothered to spend an hour or more installing it when i just want to sit down and have a few beers. Something for first thing tomorrow morning. :D
 
Installed, took just under an hour in total.

Decided not to use Liquid metal and used the provided MX4. Took me a while to clean the Liquid metal off the CPU that i used with the stock cooler.

It looks like i have the Rev 2 Arctic II 360 which has a slightly easier installation (i think). It wasn't too difficult, although i did use the wrong screws initially to attach the mounting brackets to the CPU block.

Only realised when i was tightening it up and it was not making very good contact with the CPU heat spreader that i realised something was wrong and started over.

Fans were already attached to the radiator which was nice and in the correct position i wanted for push.

No instructions included in the box, but was able to scan a barcode off a booklet provided and that pointed me to installation instructions or a video installation.

Wasn't as bad as i thought it would be overall.

Picture
rgoIZDs.jpg

I am currently running the fans at full whack (just over 1700rpm) and considering its running at 100% they are not very loud at all. I will have to turn them down in the BIOS though for silence.

Currently idling at 28-32c with a browser open with half a dozen tabs and two youtube tabs. It's at least 15c+ lower than the stock cooler in the same situation.

Now to test some Football Manager...
 
LtMatt said:
Installed, took just under an hour in total.

Decided not to use Liquid metal and used the provided MX4. Took me a while to clean the Liquid metal off the CPU that i used with the stock cooler.

It looks like i have the Rev 2 Arctic II 360 which has a slightly easier installation (i think). It wasn't too difficult, although i did use the wrong screws initially to attach the mounting brackets to the CPU block.

Only realised when i was tightening it up and it was not making very good contact with the CPU heat spreader that i realised something was wrong and started over.

Fans were already attached to the radiator which was nice and in the correct position i wanted for push.

No instructions included in the box, but was able to scan a barcode off a booklet provided and that pointed me to installation instructions or a video installation.

Wasn't as bad as i thought it would be overall.

Picture
rgoIZDs.jpg

I am currently running the fans at full whack (just over 1700rpm) and considering its running at 100% they are not very loud at all. I will have to turn them down in the BIOS though for silence.

Currently idling at 28-32c with a browser open with half a dozen tabs and two youtube tabs. It's at least 15c+ lower than the stock cooler in the same situation.

Now to test some Football Manager...
Click to expand...

Tested some football manager, saw a peak of 72C under 100% load with all cores at 100% utilisation, clock speed of 4.075Ghz-4.125Ghz. Temp mainly locked at 70c during processing.

Stock cooler with Liquid metal saw a peak of 92C under 100% load with all cores at 100% utilisation, clock speed of 3975Ghz-4.025Ghz. Temp mainly locked at 90C during processing.

The biggest difference is coming under lighter loads where 2-4 cores are lightly stressed. For example when playing Football Manager and watching back the highlights of the game. Now the CPU frequency sits at 4.5Ghz-4.525Ghz solid during this time, whereas previously it would be at 4.325Ghz-4.4Ghz and fluctuate a lot.
 
LtMatt said:
Tested some football manager, saw a peak of 72C under 100% load with all cores at 100% utilisation, clock speed of 4.075Ghz-4.125Ghz. Temp mainly locked at 70c during processing.

Stock cooler with Liquid metal saw a peak of 92C under 100% load with all cores at 100% utilisation, clock speed of 3975Ghz-4.025Ghz. Temp mainly locked at 90C during processing.

The biggest difference is coming under lighter loads where 2-4 cores are lightly stressed. For example when playing Football Manager and watching back the highlights of the game. Now the CPU frequency sits at 4.5Ghz-4.525Ghz solid during this time, whereas previously it would be at 4.325Ghz-4.4Ghz and fluctuate a lot.
Click to expand...
That sounds more like it.

I dont mind my gpu being air cooled, as when I game I mostly wear headphones. But the CPU has to be silent, and the stock one just wasnt good enough noise wise with decent temps.. Very close, but not enough for me.
 
Last edited:
Posted this in another thread, but may save some time if you are banging your head against a wall expecting magic from PBO.

Did some testing of PBO today.

Various loads : Gaming / Super Pi / Prime Blend / CB20 Single Core

Peak/Boost speed measured by HWInfo

Within margin of error on my part - It does nothing.

TVWH11P.png

25mhz i conside margin of error / 0.5% in mhz
Biggest difference of 50mhz in gaming could be down to the inconsistancy of load in a non consistant environment, still only approx 1% variance in mhz
 
Brizzles said:
Posted this in another thread, but may save some time if you are banging your head against a wall expecting magic from PBO.

Did some testing of PBO today.

Various loads : Gaming / Super Pi / Prime Blend / CB20 Single Core

Peak/Boost speed measured by HWInfo

Within margin of error on my part - It does nothing.

TVWH11P.png

25mhz i conside margin of error / 0.5% in mhz
Biggest difference of 50mhz in gaming could be down to the inconsistancy of load in a non consistant environment, still only approx 1% variance in mhz
Click to expand...
Yup. All pointless. Just install and cool it as much as you can, that makes a difference.
 
Am coming to the conclusion that it is probably best to simply ignore the temperatures (and their fluctuations) and the volts that the Asus board sends to the cpu and leave everything on auto (apart from docp).

It is running quietly and not , apparently, overheating. The performance is more than acceptable and I am not looking to get 'the best' when it comes to either speed or temperatures - So , for now at least , I believe it best to leave well alone. Agreed?
 
AgedIdiot said:
Am coming to the conclusion that it is probably best to simply ignore the temperatures (and their fluctuations) and the volts that the Asus board sends to the cpu and leave everything on auto (apart from docp).

It is running quietly and not , apparently, overheating. The performance is more than acceptable and I am not looking to get 'the best' when it comes to either speed or temperatures - So , for now at least , I believe it best to leave well alone. Agreed?
Click to expand...
Definitely. Let the board and chip do its thing.
 
I have mine set up as per the 1usmus guide and am using that power plan. Fully loaded in CB20 it sits at around 4.1 but uses 1.36 volts to do that and goes to 85c fairly quickly but no higher.

I've been experimenting with undervolting and using a - 0.1 offset the temps stay nearer 70c, the multicore score is within a margin of error and the single core is about 2% down. The only difference seems to be that my best core boosts to 4625 rather than 4675 (3 cores go to at least 4.6 with the other 3 good cores at least 4.5 still).

Is 1.36v too high for a multicore load even though I'm not setting it? I'm guessing I'm not seeing clock stretching with my undervolt as the performance is virtually identical, unless anyone has an idea for a better test than CB to prove it?
 
r22snapper said:
I have mine set up as per the 1usmus guide and am using that power plan. Fully loaded in CB20 it sits at around 4.1 but uses 1.36 volts to do that and goes to 85c fairly quickly but no higher.

I've been experimenting with undervolting and using a - 0.1 offset the temps stay nearer 70c, the multicore score is within a margin of error and the single core is about 2% down. The only difference seems to be that my best core boosts to 4625 rather than 4675 (3 cores go to at least 4.6 with the other 3 good cores at least 4.5 still).

Is 1.36v too high for a multicore load even though I'm not setting it? I'm guessing I'm not seeing clock stretching with my undervolt as the performance is virtually identical, unless anyone has an idea for a better test than CB to prove it?
Click to expand...

It does seem high. I don't think mine broke above 1.275v in CB20 MC at stock/PBO
 
r22snapper said:
Is 1.36v too high for a multicore load even though I'm not setting it? I'm guessing I'm not seeing clock stretching with my undervolt as the performance is virtually identical, unless anyone has an idea for a better test than CB to prove it?
Click to expand...

1Usmus / Auto voltage

5 runs of CB20 - This is the highest voltage i saw; 1.231v spike / Average 1.212v

tT9EsDP.png
 
LtMatt said:
Tested some football manager, saw a peak of 72C under 100% load with all cores at 100% load, clock speed of 4.075Ghz-4.125Ghz. Temp mainly locked at 70c during processing.

Stock cooler with Liquid metal saw a peak of 92C under 100% load with all cores at 100% utilisation, clock speed of 3975Ghz-4.025Ghz. Temp mainly locked at 90C during processing.

The biggest difference is coming under lighter loads where 2-4 cores are lightly stressed. For example when playing Football Manager and watching back the highlights of the game. Now the CPU frequency sits at 4.5Ghz-4.525Ghz solid during this time, whereas previously it would be at 4.325Ghz-4.4Ghz and fluctuate a lot.
Click to expand...
Decided to enable PBO and crank it as high as it will go and am now seeing 4.150Ghz - 4.200Ghz under 100% load with temp locked at 75c.
 
Last edited:
Guys, I've joined the club!

Haven't done any form of optimisation yet, apart from setting RAM to XMP profile so it runs at 3600 with 16 timings.

Initial quick and dirty CinebenchR20 yields results of 7268 multicore test. All cores seem to sit at about 4150-4160Mhz, which seems to be ok at first glance. Sincle core seems to hit about 4530Mhz, which seemed on the low side.

Its fully under water, loop consists of a 360 rad, 240 rad and a 1080Ti.

Temp wise when running Cinebench with duration settings (so it loops), it hovers 58/59oC with the occasional touch of 60oC.

Looking forward to more of a play tomorrow evening!
 
Tee Hee Johnson said:
Guys, I've joined the club!

Haven't done any form of optimisation yet, apart from setting RAM to XMP profile so it runs at 3600 with 16 timings.

Initial quick and dirty CinebenchR20 yields results of 7268 multicore test. All cores seem to sit at about 4150-4160Mhz, which seems to be ok at first glance. Sincle core seems to hit about 4530Mhz, which seemed on the low side.

Its fully under water, loop consists of a 360 rad, 240 rad and a 1080Ti.

Temp wise when running Cinebench with duration settings (so it loops), it hovers 58/59oC with the occasional touch of 60oC.

Looking forward to more of a play tomorrow evening!
Click to expand...
Looks pretty good that score.

Don't forget to use the 1usmus power profile and set the Performance enhancer option to 3 (if you have an Asus motherboard).
 
Tee Hee Johnson said:
Guys, I've joined the club!

Haven't done any form of optimisation yet, apart from setting RAM to XMP profile so it runs at 3600 with 16 timings.

Initial quick and dirty CinebenchR20 yields results of 7268 multicore test. All cores seem to sit at about 4150-4160Mhz, which seems to be ok at first glance. Sincle core seems to hit about 4530Mhz, which seemed on the low side.

Its fully under water, loop consists of a 360 rad, 240 rad and a 1080Ti.

Temp wise when running Cinebench with duration settings (so it loops), it hovers 58/59oC with the occasional touch of 60oC.

Looking forward to more of a play tomorrow evening!
Click to expand...

What CPU Block?
 
You must log in or register to reply here.
Back
Top Bottom