Originally Posted by Takla
Dude, are you dense? Do you read only half of my sentences at a time, and than your brain just shuts of from overheating or what? I am using a fixed constant cpu core voltage of 1.30v and a fixed constant core clock of 4300mhz because that gives me the best performance ON TOP of also giving me a much lower max temp AND idle temp, compared to BOTH all stock and/or pbo+auto OC. I already watched all the youtube videos about ryzen 3k (seriously, all of them). stop linking these. you are not showing me anything new.
also, overclocking per ccx is only really viable for cpu's that use more than 2 ccxes like the 3900x for example. personally, I wasn't even be able achieve a mere 25mhz more for ccx0 (core 1, 2 and 3) over the 4300 all cores so to me this is yet another worthless feature.
edit: On the toxic part; well, to me, you do not appear to read my comments properly, and just skip over them, which is quite disrespectful to begin with. So this is what you get.
Dude, chill a bit please
That being said:
From my experience, there are a couple of possibilities at play here, and you may have already addressed some or all of these (apologies, if so).
Regardless of whether you are using PBO or not, setting a fixed voltage really wrecks the ability of the CPU to do low power states. Personally, I have worked around this by applying a negative voltage offset to the CPU (as opposed to a fixed voltage, and usually around -0.075 v). I have found that higher levels of LLC on the CPU also give a little bit better range for the CPU to maintain maintain a low voltage when idle, but also get what it needs when boosted. Experimentation here is key as your mainboard and CPU will certainly react a little differently than mine. Also, in Windows, make sure you are using the 'Balanced' power plan (not the 'Ryzen Balanced' power plan if it is there), and that the CPU low is set at 5%.
This has worked for me with the Ryzen 1000 and 2000 series CPUs. Nothing I have read has lead me to believe it will be any different when I finally upgrade to the 3000 series CPUs. Also, there are a few known issues that are being addressed via AGESA and firmware updates that can have a significant impact on CPU power. Launch issues with Ryzen stuff are old hat by now. History with prior launches has shown that given a couple of months, everything should be pretty well smoothed out, followed by a much slower, periodic progressive run of firmware refinement updates.
Does the situation suck? Yes. Do AMD and their mainboard partners ever seem to learn from their past mistakes in this regard? The answer really seems to be No. Should we expect and demand better? Absolutely - but until we actually GET better, this is what we have. It is what it is...