Got a new Z790 ordered up.
So on the temp board B660 with code x112 just let it run default VID. Yeah, it shows over 1.5v sometimes. Meh, no big deal.
When on a proper Z790 board, VID is left on auto and usually run about 1.4v all core looking for 5.8ghz P-core and 4.7ghz E-core. The area effecient cores really suck down the juice though.
In order to run 8P cores and no E-cores at 6.3 to 6.4ghz 16 threads will utilize between 1.475 and 1.50v. Requires better cooling than most people intend to work with.
Just to validate 7ghz on a couple cores, took about 1.70v.
[ukvv7b] Validated Dump by ShrimpBrime (2024-05-28 18:15:17) - MB: MSI Z790 GAMING PLUS WIFI (MS-7E06) - RAM: 32768 MB
valid.x86.fr
I suppose everyone still worried......
I see. When I first unshackled my cpu. ( I had it under a million restrictions when the panic was at its max, just cause there was so much smoke, and I had no idea how much fire. I had vid limit 1.3v ecores max 3.5 pcores max 5.0. pl1/pl2 max 200. AC/DC at 70.
When the update came out, and I realized, I should be getting what I paid for, I unshackled it, but I was getting ( as software reports - so probably not that accurate, vcore up to 1.44) which I didn't like. But now. With just the AC/DC loadline reduced just a little bit, from 110 to 90, and taking that vid limit down from 1.55 to 1.5, my vcore (as reported) doesn't go above 1.4. On occasion it might go up to like 1.41. Gotta always assume transients are higher, but I can't imagine them being any higher than say 1.46 (again, transients, just to be clear - sorry, been speaking with some people with real poor reading comprehension lately - not your fault).
Anyway... For example when I set a vid limit of 1.4, my software reports no vid above 1.37/138 (again, probably because of transients) and then, there's just my vcore generally being lower than my vid by... idk it depends. Just checking hwinfo it looks like my max vcore is .3 lower than my max vid (dont ask me why, but I'm glad). I know some people like to have them like exactly the same but I don't really care so much about that.
So with my cpu firing on all cylinders, so old ucode, high iccmax, big undervolt, memory overclocked to snot with sky high trefi... passmark cpu test would get me like 57,000 and some change. Take away the memory OC and put on intel defaults it goes down to ~55,000. Then set the vid cap 1.5 and reduce the AC/DC loadline from 110 > 90 gets me ~54,500. So yeah, about 5% slower than the cpu completely unrestricted, and about 1% slower than intel's normal performance profile. Doesn't seem like such a bad deal, eh? I just hope there's no invisible degredation, even though I feel there probably is.
I mean how do you tell it apart from just not the best bin? Then again this chip has never quite been able to maintain boosts under load like my first chip. But also this one works after 8 months and that first one died almost immediately so its gotta get points for that, right?
Btw... I still have a moderately inflated trefi on my memory.... I figured that wouldn't stress the memory controller, if anything it would do the opposite, right? So I put it at 40,000 with trefix9 or w/e its called at 40 so that my memory still has acceptable latency. Plus at this low frequency its not like its gonna get too hot or anything, which asfaik is one of the main risks with increasing it ( If somebody wants to correct my assumptions on this, feel welcome).
I mean a Z board would be nice but.... I don't feel like I should be putting any more money into this platform and rather start saving for the next. I'm pretty broke
I find it highly improbable for Intel/MB makers to actually show you in plain text what is max. VID of your CPU (at max. frequency CPU operates at). However, even if they won't provide number directly - there are ways to figure it out indirectly
I mean intel just did.... mind you, they are in a bit of a bind so they have to do something:
here's the link:
https://community.intel.com/t5/Proc...el-Core-13th-and-14th-Gen-Desktop/m-p/1622129
And the quote "The latest microcode update (0x129) will limit voltage requests above 1.55V as a preventative mitigation for processors not experiencing instability symptoms."