That's not how it works with the voltage. You want to be aiming closer to 1.325v under full load. Although that's no official number either. Up to 1.5v is ok for lightly loaded tasks.
-
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.
3900x PBO.
- Thread starter Scottx125
- Start date
That's not how it works with the voltage. You want to be aiming closer to 1.325v under full load. Although that's no official number either. Up to 1.5v is ok for lightly loaded tasks.
I know. Moderate load line 1.4v, aggressive load line is 1.35v. Either way as XFR is temperature dependant I'm going to improve the cooling to see if it improves my results with what I tested earlier.
this is true, there is a reason why even pbo/xfr voids the warranty. The only gauranteed operation mode is having the cpu perfectly stock.
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.
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
That's normal. Have a read of reddit you'll see tons of people saying pbo does nothing, tons of people saying pbo is slower than stock and another ton of people like pianos who says it's the best thing since alices bread. The only conclusion is that everyone has different quality cpu and the only gauranteed performance is the one you get when the cpu is stock out of the box
I think a lot of it is down to motherboards too. I wouldn't have thought I had a bad chip with it being 4.4ghz at 1.3v all cores. I've had it on 4.5ghz but that requires closer to 1.4v and I'm not happy to run that high.
That's a good chip if you can hit 4.4Ghz at 1.3v. I'd probably be able to hit 4.25ghz. I've noticed on mine at 4.4Ghz it takes a bit of extra juice to become stable.
Yeah it's definitely not a bad one. It's under a custom loop with two 360 rads so cooling is definitely not an issue either.
360mm AIO / 70 degrees max load
Lol yeah that's def a good setup then xD. I've just got an Arctic Liquid Freezer II 360 I just installed. Gonna add some extra fans and turn it round (installed it upside down temporarily as it didn't fit right way. I've done some tweaking so it should fit right way up).
Pretty much the same on my test. Annoys me as 70 degrees IMO is on the low side of high strain operating temps. I'd prefer to run mine hotter at 80-85 degrees. Shame you can't tell XFR to only down clock when it starts nearing a set temp.
Sounds correct for that temp then.
Weird. Was experimenting with PBO/XFR after flashing the bios. Computer started running into general stability issues after i disabled LN2 mode in the bios. Re-enabled my old manual OC profile and the system shut down after momentarily freezing and then rebooted 0.o. I've reset everything to stock now and it seems fine. Seems like either updating the bios or messing with the various settings has caused some instability. I'm going to stop messing for the time being and let my computer get back to normal, after which I might try my manual OC again.
If you flash a bios to new version you need to apply all settings again, otherwise if you load from previous saved settings you will end up with a lot of issues.
Will do, I'll wipe the bios and redo the settings.
Update:
Yup seemed to do the job. Thanks. Decided to spend the entire afternoon trying PBO/XFR various combinations on cinebench R20 comparing scores etc. From what I've found, PBO/XFR working together gives the best results as PBO keeps the all core clocks on average 50-100mhz higher. In terms of lowering the off-set to improve temps and consequently performance. I went down to -0.1v, performance was sort-of the same, there was a slight downward trend and the performance ranges increased but all in all the temps dropped by maybe 2 or 3 degrees so it really isn't worth it.
My summary of PBO/XFR is that it's good if you got a good chip and regularly boost up to high numbers. However for me running Cinebench R20 closing most apps down on my computer and running 3 tests on each power mode with various other settings and memory OC to 3733mhz, the best score I could get was around 7150 and on single core 512. I think the issue is the settings are too sensitive to temperature, for me 70 degrees is moderate and I tend to cap my temps at 80-85. So I'll be moving back to my manual OC until I can modify my cooler to see if it improves the PBO/XFR OC.
Last edited:
Just wondering anyone know what the Ryzen 3900x's max safe Amps/Watts are? Been reading into how LLC effects Vdroop. I'm running mine with LLC lvl 5 with no noticeable increase in voltage under heavy load. What do you guys think is a safe wattage and current? I've read that generally MOBO's for 3900x should be able to supply 140W CPU power, 140A EDC and 95a TDC as default. Meaning there's probably some headroom above these numbers?
Last edited:
Dont exceed 1.27v on all core overclock
How are you reading voltages ?
LLC should show an effect on the actual chip voltage.
You wont be able to remove the PPT / EDC / TDC 'limits' beyond what your motherboard states, even then the chip will limit itself ;
Last edited:
Was running 1.3625v when i got my score of around ~7700 CBr20(currently around 7200 on auto). I'm now on auto messing with XFR/PBO. I've managed to eek out some more performance. Now I'm boosting to 4.6Ghz on single core and 4.075/4.1ghz on CBr20. Gonna keep messing. Want to hit all core of 4.2Ghz as my goal :3.
Lol just took a second look at that. You mean -0.0625 xD?!
