• 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 ClockTuner

Tee Hee Johnson said:
Your spec is very similar to mine, and it ran just fine. Although my ram is 3600 cl16 and my 3900x is apparently 'silver'.

I wonder what the differences are with people who get it to run and people who don't.
Click to expand...

Yeah, it's odd.

The diag part runs fine and gives me the possible settings.
When hitting start CB20 runs fine and logs a score but it's when it goes on to do the testing that at some point it will end up just restarting the PC.

I've messed about with all sorts of settings and all I can seem to get it to do is change the length of time before the PC restarts and it's always just a hard power trip restart.

Even with the mem, my manual timings don't work, nor xmp, even setting the mem to 3200 and leaving timings on auto (where it booted at something daft like cl22 !) don't help.

Might try just setting the 4375MHz/1250mV or 4200MHz/1150mV the diag part recommends and see if that is stable.

edit: what did you go for for LLC in the end - assuming you have the same mobo?

He did tweet to ignore the LLC warnings in the log but the documentation seems to contradict itself.
I tried pretty much all the settings but medium seemed to be the one where it didn't complain about it in the diag run.
 
chaosophy said:
Was playing with this last night and didn't ever complete a run with out it rebooting the PC. :/

So it looks like I should have a fairly decent chip, so no idea why the tool never completes for me with out it rebooting the PC. :(
Click to expand...

Same here unfortunately - Asus Strix X570-I + 3900X. Bios is fully updated and have changed all necessary settings. Diagnostic runs fine, but hitting Start always results in a shutdown/restart eventually.

I saw this on reddit:

Which CPU are you using, and did you run the "Diagnostic" beforehand?

I ran into the same issue with my "Silver Sample" 3900X, while testing out the recommended 4300 MHz / 1.25 V Target Frequency/Voltage parameters (assuming this is specific to silver sample 3900X as per the diagnostics results).

PC restarted at one of the stages when it was tryna gauge if 4.45 on CCX 1&2 was doable at 1.25 V. Solved this by upping the target voltage from 1250 mV to 1255 mV.
Click to expand...

Mine also shows as a "Silver Sample", so I'll try adjusting the voltage slightly after he releases an update tonight.
 
Last edited:
I put memory on XMP and changed my LLC to normal/middle. Now the tool keeps screwing up the reference frequency so I think its bugged. If you check his twitter feed out he is releasing a hotfix tonight.
 
Diplodocus said:
Same here unfortunately - Asus Strix X570-I + 3900X. Bios is fully updated and have changed all necessary settings. Diagnostic runs fine, but hitting Start always results in a shutdown/restart eventually.

I saw this on reddit:



Mine also shows as a "Silver Sample", so I'll try adjusting the voltage slightly after he releases an update tonight.
Click to expand...


Interesting, may give that a go. Probably also going to wait for an update or two to the tool before retrying.

Thanks.
 
so far soo good great app. i have R5 3600 silver plate all cores 4225mhz at 1.275v good for gaming and general pc use. cpu temps went down maybe 7- 10dregrees happy with it overall. Will see in long run how it goes
 
Would I be right in assuming this is for an "all core boost". So... setting this up would remove the ability for one of the cores in my good CCX (CCX2) to boost to the... 4.65 I do sometimes get on highly single threaded stuff?
I get that the all core makes it great for CB20 but... with gaming performance as an objective, isn't letting PBO do it's thing a better option?

U2PQKCV.png
 
Last edited:
Mercutio said:
Would I be right in assuming this is for an "all core boost". So... setting this up would remove the ability for one of the cores in my good CCX (CCX2) to boost to the... 4.65 I do sometimes get on highly single threaded stuff?
I get that the all core makes it great for CB20 but... with gaming performance as an objective, isn't letting PBO do it's thing a better option?

U2PQKCV.png
Click to expand...
dont know but my 3Dmark Port Royal score went up with this all core boost from 11181 to 11431. Previous score is stock boost . I know its synthetic benchmark not real game
 
Last edited:
Well... running this between completely untuned yesterday and today on timespy (yes, still synthetic but...) my CPU dropped from 10222 to 9806.
Frequency is clearly lower too 4.549mhz down to 4.341 (which... makes sense with the above tuning).

It's obviously going to change a good bit from game to game (or game benchmark to game benchmark) depending how well the engine uses it multithreading but... yeah, I think I'm headed back to defaults with PBO.

Awesome tool for those wanting all core though. No doubt on that.

As a tip. If you hit start or diag and the progress bar isn't moving, something's not quite right. Stop it, reboot, run CB20 and Ryzen master manually, close them both, open it and try again.

Same thing if you hit diag and no...checking seems to occur before it decides what quality of chip you've got.
 
chaosophy said:
Yeah, it's odd.

The diag part runs fine and gives me the possible settings.
When hitting start CB20 runs fine and logs a score but it's when it goes on to do the testing that at some point it will end up just restarting the PC.

I've messed about with all sorts of settings and all I can seem to get it to do is change the length of time before the PC restarts and it's always just a hard power trip restart.

Even with the mem, my manual timings don't work, nor xmp, even setting the mem to 3200 and leaving timings on auto (where it booted at something daft like cl22 !) don't help.

Might try just setting the 4375MHz/1250mV or 4200MHz/1150mV the diag part recommends and see if that is stable.

edit: what did you go for for LLC in the end - assuming you have the same mobo?

He did tweet to ignore the LLC warnings in the log but the documentation seems to contradict itself.
I tried pretty much all the settings but medium seemed to be the one where it didn't complain about it in the diag run.
Click to expand...

The guru3d guide suggested Turbo LLC on Gigabyte boards, so I stuck to that.

The only other thing I done was reflash with the 30 bios and restore defaults then set xmp and fan curves.
 
No success here with 3900X / Giga Aorus Elite X570 / F30 Bios

1 diagnosis shows a Platinum chip, then another changes to a Silver chip.

CB20 score gives 7050 as the baseline CTR test ( i run a slight undervolt and get 7300 normally )

Hard shutdown during 'tuning'

Will wait for the software to mature before trying again.
 
Running mine now, diagnostics suggesting I have a silver sample :D so not expecting much.

That's despite four of my cores being able to hit 4.65 during regular usage.
 
Tee Hee Johnson said:
The guru3d guide suggested Turbo LLC on Gigabyte boards, so I stuck to that.

The only other thing I done was reflash with the 30 bios and restore defaults then set xmp and fan curves.
Click to expand...

tbh, that guide is a bit of a mess. I know English is not the guys first language, so not having a go but it's clearly written by an engineer that is too close to the project imo.
It could do with being rewritten by someone one step back that comes to it with fresh eyes.

pg 2. - Gigabyte - in most cases Turbo, but it can also be Auto
pg 3. - Gigabyte was slightly behind because in Auto mode Vdroop was far from the recommended range, which is 1-2.3%

And on top of that the program itself reports on vdroop/llc when the diag runs.

For me 'Auto' always had too much vdroop, program reported like 3/4%+ and recommended upping llc.
'Turbo' was around 0% Vdroop rec was lower llc.

I ended up trying all llc settings - all still hard power reset the pc - but for me 'Medium' llc was reported as in the 1-2.x% range with out the program saying to change it.

On top of all that he is now saying on Twitter to just ignore the vdroop messages in the log.
And that it will be fixed in the hotfix.

If it can be ignored why mention it so much and have the program log it?

As above, it seems to me he is too close to the project and is making assumptions while not explaining them properly to effect laymans like us end users.
There are other issues wtih the doc imo, layout, the flow between subjects, changing terminology etc.

Again, not having a go at the guy, I think he's doing great things - his ram calc boosted my perf a lot and this looks to be another great tool - I just think maybe some help on the documentation side could help a bit.
 
chaosophy said:
tbh, that guide is a bit of a mess. I know English is not the guys first language, so not having a go but it's clearly written by an engineer that is too close to the project imo.
It could do with being rewritten by someone one step back that comes to it with fresh eyes.

pg 2. - Gigabyte - in most cases Turbo, but it can also be Auto
pg 3. - Gigabyte was slightly behind because in Auto mode Vdroop was far from the recommended range, which is 1-2.3%

And on top of that the program itself reports on vdroop/llc when the diag runs.

For me 'Auto' always had too much vdroop, program reported like 3/4%+ and recommended upping llc.
'Turbo' was around 0% Vdroop rec was lower llc.

I ended up trying all llc settings - all still hard power reset the pc - but for me 'Medium' llc was reported as in the 1-2.x% range with out the program saying to change it.

On top of all that he is now saying on Twitter to just ignore the vdroop messages in the log.
And that it will be fixed in the hotfix.

If it can be ignored why mention it so much and have the program log it?

As above, it seems to me he is too close to the project and is making assumptions while not explaining them properly to effect laymans like us end users.
There are other issues wtih the doc imo, layout, the flow between subjects, changing terminology etc.

Again, not having a go at the guy, I think he's doing great things - his ram calc boosted my perf a lot and this looks to be another great tool - I just think maybe some help on the documentation side could help a bit.
Click to expand...
I thought the guide was pretty clear and easy to understand.

Check pre requisites, extract cinebench, click run.
 
Brizzles said:
No success here with 3900X / Giga Aorus Elite X570 / F30 Bios

1 diagnosis shows a Platinum chip, then another changes to a Silver chip.

CB20 score gives 7050 as the baseline CTR test ( i run a slight undervolt and get 7300 normally )

Hard shutdown during 'tuning'

Will wait for the software to mature before trying again.
Click to expand...

If hitting diag ends immediately with that info, try and reboot, run master manually and close it, run cb20 manually and close it. Then try CTR.
 
OK, results are in.

With voltage at 1250mV, it wouldn't complete a cinebench run, tried reducing each ccx starting with the weakest but no dice.

Increased voltage to 1275mv and ran again.
Stock score = 7020
With undervolt and pbo off = 7300
Ryzen clock tuner = 7798.

Just running so cpu stability test to ensure this is sound.
 
Just to add I needed to bump the voltage to hold the ccx speeds from the original test. Just running the test again, now my first two ccx are holding 4.5 instead of 4450, so should get a slightly higher cinebench once tested.
 
nutsnut said:
I thought the guide was pretty clear and easy to understand.

Check pre requisites, extract cinebench, click run.
Click to expand...

The information is there to get what you need out of it to run the program but honestly it's a terribly written piece of documentation.
The flow of it is awful, badly laid out, jumps all over the place, changes terminology.

It decent enough considering it's one guy doing it all and english not being their first language but it'd never pass to being published most places imo.
 
You must log in or register to reply here.
Back
Top Bottom