Overclock.net banner

1 - 5 of 5 Posts

·
Registered
Joined
·
1 Posts
Discussion Starter #1
Sup, I oc-ed mine R3 1200 to 3.7GHz at 1.25V (set in bios) on ASRock AB350M-HDV. I started stress testing with Aida64 and RealBench and everything was fine. After that I started a Prime95 test (small fft's) my vcore started to raise to 1.28V and go back to 1.264V *which i guess is normal* but the strange thing is that some of the cores went to 3794MHz *even I set the multiplier to 37 at bios*. Is that normal cuz I didnt saw things like that before when I was at 1.275V (I dropped the voltage to 1.25 my lowest stable)
 

Attachments

·
⤷ αC
Joined
·
11,240 Posts
It might be simply your base clock (typical 100MHz) fluctuating if you manually set 37x multiplier.

Voltage overshooting to 1.28V instead of 1.25V is likely due to your load line calibration settings (LLC). The drop in a stress test is due to V_Droop (a voltage drop under load).

Per Robert Hallock at AMD, up to 128W on all AM4 boards is the design criteria they gave all the motherboard vendors. Your package power there is 60W so I doubt it is due to the motherboard.
 

·
Registered
Joined
·
39 Posts
Hi,

I just testing my overclock R5 1600 3.85 GHz @1.3V with OCCT large data set for 3 hours, running good
however it was crash on @1.2875V (only small different 0.0125V)

my question :
Do i need to add some voltage or reduce some MHz to have some some headroom? i yes, how much?

thanks
 

·
Totally Tubular
Joined
·
1,508 Posts
Sup, I oc-ed mine R3 1200 to 3.7GHz at 1.25V (set in bios) on ASRock AB350M-HDV. I started stress testing with Aida64 and RealBench and everything was fine. After that I started a Prime95 test (small fft's) my vcore started to raise to 1.28V and go back to 1.264V *which i guess is normal* but the strange thing is that some of the cores went to 3794MHz *even I set the multiplier to 37 at bios*. Is that normal cuz I didnt saw things like that before when I was at 1.275V (I dropped the voltage to 1.25 my lowest stable)
The CPU is also going well below 3.7GHz (note the minimum column). The CPU shouldn't be varying that much in speed when you manually set the multiplier. From 125MHz below the target to 94MHz above the target is way too big a range -- it should always stay within a few MHz (3698-3702MHz). I've never seen a board where the BCLK fluctuated anywhere near 1MHz and that would be 37MHz below to 37MHz above the target in this system.

Make sure your Windows power settings are correct. The minimum and maximum processor speeds should be set to 100% if you want it to stay at the target speed.

If you haven't already, make sure the board's BIOS is updated to the newest one. Then reset the BIOS to defaults and go back in and set them back up the way you want them.

If the BIOS settings and Windows settings are confirmed correct and that large speed fluctuation keeps happening, I would replace the board.
 

·
Totally Tubular
Joined
·
1,508 Posts
Hi,

I just testing my overclock R5 1600 3.85 GHz @1.3V with OCCT large data set for 3 hours, running good
however it was crash on @1.2875V (only small different 0.0125V)

my question :
Do i need to add some voltage or reduce some MHz to have some some headroom? i yes, how much?

thanks
If it's stable at 1.3v but not stable at the next step down, then keep it at 1.3v. Make sure you do more testing than just OCCT as other programs stress the CPU in different ways.

I use Prime95 v29.4b8 Small FFT preset, RealBench v2.56 with your full RAM amount selected, and Y-Cruncher to stress test my systems. And then I make sure my games and other programs work, of course.
 
1 - 5 of 5 Posts
Top