Zen 2 Owners Thread!

muon · 25 Jul 2019 at 19:59

@Sin_Chase

I wouldn't rely on that hwinfo64 number. I'm pretty sure you just want to focus on the CPU Package Power.

Jokester · 25 Jul 2019 at 20:00

If it does this across the board and it’s not working as intended, then it invalidates pretty much all the reviews done with 3600MHz RAM.

Minstadave · 25 Jul 2019 at 20:04

Sin_Chase said:
The SoC Power usage [W] between 3533/3600/3733 is largely unchanged (<0.5w variance). There is a marginal uplift of Amperage for each speed bump (<0.5A) and SOC voltage is largely unchanged.

A Single Core Cinebench R20 is not power limited though. Max CPU Package Power of around 77w.....PLENTY left out of the stock 142W on a 3900X. So why the throttling of the cores?

This is what I want to know, I'm not hitting power or thermal limits but still cant boost to the advertised 4.4ghz.

Sin_Chase · 25 Jul 2019 at 20:04

muon said:
@Sin_Chase

I wouldn't rely on that hwinfo64 number. I'm pretty sure you just want to focus on the CPU Package Power.

Graphed that out above also.

Notable increase from 3533 > 3600 but not again from 3600 > 3733. Still only a single Core Cinebench R20 workload though with over 60-70w of PPT headroom to play with.

muon · 25 Jul 2019 at 20:05

Sin_Chase said:
Graphed that out above also.

https://i.imgur.com/lvC0w3p.png

Notable increase from 3533 > 3600 but not again from 3600 > 3733. Still only a single Core Cinebench R20 workload though with over 60-70w of PPT headroom to play with.

Depends on the PPT limit set for lightly threaded loads. I doubt they would want you to pump 140w through a single core, probably cause it to ~~melt~~ burn.

Graph does suggest it is coming up against a limit above 3533.

Sin_Chase · 25 Jul 2019 at 20:08

muon said:
Depends on the PPT limit set for lightly threaded loads. I doubt they would want you to pump 140w through a single core, probably cause it to melt.

True. But one would assume that the 142W of available package power can be distributed as needed, be it to a single core workload + SOC or multi core + SOC. Why should the SOC power requirements be limited by the wattage given to a single core or vica-versa.

Basically, why would a single core be PPT limited as a result of SoC when there is plenty of available headroom in the package as a whole for SOC to draw from.

muon · 25 Jul 2019 at 20:09

Sin_Chase said:
True. But one would assume that the 142W of available package power can be distributed as needed, be it to a single core workload + SOC or multi core + SOC. Why should the SOC power requirements be limited by the wattage given to a single core or vica-versa.

Basically, why would a single core be PPT limited as a result of SoC when there is plenty of available headroom in the package as a whole for SOC to draw from.

That is true. Obviously this is speculation but maybe it gets set as a whole. Therefore increasing the undie power consumption decreases that available for the core.

Could be a bug. But I think my explanation seems plausible.

muon · 25 Jul 2019 at 20:11

You could test it by removing PPT limits and running multicore comparisons.

muon · 25 Jul 2019 at 20:13

Sin_Chase said:
This simply is not universally true. Please see my post above https://forums.overclockers.co.uk/posts/32885009

Certain undervolting (too much, too little) can start reducing performance but there are sweet spots where one can reduce temps and maintain or improve performance.

This is also what I see.

muon · 26 Jul 2019 at 11:44

Got 1.0.0.3 bios for my board yesterday.

PBO + Auto OC now works. Seeing up to 4.4ghz for lightly threaded loads looks pretty good. Still 4.2ghz for all core where temps don't get too high.

opethdisciple · 26 Jul 2019 at 12:52

muon said:
Got 1.0.0.3 bios for my board yesterday.

PBO + Auto OC now works. Seeing up to 4.4ghz for lightly threaded loads looks pretty good. Still 4.2ghz for all core where temps don't get too high.

What board and cpu?

Minstadave · 26 Jul 2019 at 12:56

muon said:
Got 1.0.0.3 bios for my board yesterday.

PBO + Auto OC now works. Seeing up to 4.4ghz for lightly threaded loads looks pretty good. Still 4.2ghz for all core where temps don't get too high.

If that's a 3700X that's good!

muon · 26 Jul 2019 at 17:16

Minstadave said:
If that's a 3700X that's good!

Ryzen 3600.

opethdisciple said:
What board and cpu?

Asrock B450 Pro4.

Asrock just released loads of 1.0.0.3 bioses.

DragonQ · 26 Jul 2019 at 18:07

muon said:
Asrock just released loads of 1.0.0.3 bioses.

Supposedly 1.0.0.3AB but I have no idea how to verify that.

Danthus · 26 Jul 2019 at 20:03

Wrong thread.

havenrl · 26 Jul 2019 at 20:18

Been testing out the temperature of my 3700x.

The AMD Ryzen Master, says 95C is the max heat. Is that real?
During my 1.5 hour gaming session, I hit 75C.

Idle temps seem to be from 35C to 60C, runs hotter at idle (ok, no gaming), than my old i2500k.

What are the safe temps? 75C ok?
I'd appreciate thoughts, since I'm new to being back to AMD cpu.

Buffalo2102 · 26 Jul 2019 at 20:20

That's absolutely fine and normal from what I've seen.

havenrl · 26 Jul 2019 at 20:21

Buffalo2102 said:
That's absolutely fine and normal from what I've seen.

Thanks for your thoughts, cheers.

roysters · 26 Jul 2019 at 21:29

Yeah the results seem to be pretty much the same across-the-board.

I've stopped worrying about it until they release some sort of update or... well...It blows up.

Sin_Chase · 27 Jul 2019 at 02:56

AMD_Robert due to issue an update with regards to some issues Tuesday 30th.

https://www.reddit.com/r/Amd/comments/ciajef/placeholder_update_on_whea_warnings_destiny_2_and/