Seeking advice regarding Z490 build

Ah man..  seems like every time I get comfortable with a good game

sig my arse!

You best be careful slick.. and this is no joke.

Han's just started getting back to me on some results in which I need to digest the data and make adjustments. We are seeing card manufacture differences appearing as well.

If all you got was the first list which I thought would be simplistic to start off with, be forewarned there were issues and a lot more communications after that so if you think you know what you're doing you go right ahead and do it.

There was a reason why you didn't get those from me and I am going to let find out on your own. Then we will see who's rubbing who's nose in it and who was protecting your smart arse by not giving you a loaded gun with the safety OFF and a hair trigger.

Ya know, if you would cut out the self destructing know-it-all crap I wont love you any more or less, but the wife sure as heck will.  I guarantee it!

To late gunnery sergeant! That browner Deviant already gave me the 3090 overclocking instructions that you where rubbing my nose in! Up yours sergeant!

Oh and I got the Nuke codes of Donovan too!

I'll post video card clocking when and IF I believe it is something safe for public display and use. Right now it's too soon and there is testing/evaluation going on around it. The drivers are suspected as being coded to downclock due to the hardware that was shipped out and questions about stable cards on the market and later they may remove that changing how the card is clocked...  then there are issues with properly reading the memory clock where some of these applications are reading x8 higher than the actual memory clock speed.

So until such time as I think no one will blow their 800-2000 dollar video card into the junkpile within a year or two thinking they clocked it right and are top dog on 3DM scores.. I'm only working with a select few on that as I can due to my current availability for it.

If anyone here thinks there are landmines in a clearly written 1-2-3 simplistic CPU clocking guide you ain't seen nothin yet if you try to clock one of these card on your own without proper engineering guidance or take advice from some idiot gamer or hardware site on the internet about how to do it right with the 3x card right now.

So, if there are going to be anymore hissy fits... 

I didn't realize anyone was in a serious relationship where they feel like they are a gaslighted emotional victim.

Now listen Karen,

You do know that we were never going to take warm showers together and hold hands. I never lied to you about anything, never said I love you and never led you on. So please, wipe the mascara from your eyes, I am sure that burns and hurts <here's a tissue>...  put the gun down on the floor and step back from it before anyone gets hurt. 

Those target VID's are to get you in the range and where you need to be as quickly as possible. From there your system and CPU may very well run a lower offset voltage..  in some cases may need a touch more offset voltage. 

Lets say you locate the VID for 5.0 (1.315) and the offset to achieve that on your system is  [-]  0.025. Your system passes this offset with flying colors and then you start reducing offset by .005 which then makes the voltage [-]  0.030 and it passes again...

Then drop another .005 to [-] 0.035 and passes ..  lets say you did that a total of 6 times till it failed and then raised it back to drop #5 where it passed. That would mean your offset voltage is now [-]  0.050 and the VID would most likely be much lower than the outline target.

That could be an indicator that you have lottery winner. It would depend on where the stable VID sits at 5.2GHz if it may be possible to roll into 5.3 safely (or at least reasonably safe). In a case where a CPU is stable @ a VID of say 1.40-1.42 for 5.2GHz there is a good chance 5.3 may be possible because the chip is very stable at lower voltage.

The example above is just scratch-off pipe dream  but could happen. It is the same mindset when applied to finding the lowest offset voltage for any target CPU speed. The target VID is to get you there as fast as possible, safe, and then go from there.

You may get a good drop in offset and run a lower VID which in turn will run a lower temp and watt load. Testers found successful offset drops as small as .003 from the target had a positive influence on temp and watt loads especially at higher clock speed.

The ranges I posted were based on clock reports for all the systems that were tested. Just because you may see a bit of difference doesn't indicate a problem, simply could mean the voltage regulator modules in use are a different manufacture run of components. 

My suggestion around that is one of two:

  1. When working each clock setup for the first time go ahead and rework the VID down in offset voltage steps during OCCT:LIN 64 AVX and repeat till fail, then raise back to last successful run and confirm. Then roll that into OCCT:CPU and see it if passes. If it doesn't slightly raise offset till it passes, no need to rerun LIN. From that point run HCI Memtest. If HCI Memtest fails try raising VCCSA .01 first and if that fails set it back and raise offset a bit.

   2. Complete all of your load testing up through memory testing for each clock level, then go back and try working the VID down using offset the same way I outlined in #1. At this point you have established a stable 5.0/5.1 and if possible 5.2 but are simply looking to improve the temp/watt load. Reductions in offset will typically have far more influence on temps as speed increases.

The VID target is an area where we know the stability point for each clock level to locate them fairly quick but because of component changes in manufacturing there can be differences and therefore a fine tune of the VID may be required.

To speed things up you could try dropping offset by .005 at first and if that passes all the tests again I would say that would have a good influence on temps and watt load. If it doesn't pass then try a smaller amount like .003.

The behavior you are seeing is correct. System will display a lower real time Vcore during OCCT:LIN and higher real time Vcore during OCCT:CPU so it is working as outlined. 

Merry Christmas  

There's no hurt feeling here, I just can't hold your hand through the process.

I pushed you hard throughout all this to get mistakes and the corrections for them to seriously sink-in and also make sure anyone else reading what was happening would learn and not make the same mistakes that appeared to be continued and repeated.

I was saving the last blank hurt feelings report copy I have here so you could fill it out... You can make copies and file many multiple reports...  just make sure to list all the mean things and the reason why too..  

Be that because they didn't have IHS distortions, didn't remove the TIM, run the H150i Pro, had large towers, have new (not fleabay old/used) new/or quiet/efficient high volume fans, had room ambient temps @ or below 78F.. all the recommendations I made along the way...     I don't know

After that work the VID and don't forget to run the 3D load tests..   it may still all be unstable for all you know at this point.

Now...please go to bed, I have work to do. 

Ah come on Nick! Its all in the George Costanza spirit. I'm sorry I didn't mean to hurt your feeling. I thought you didn't have any. Here's a big Christmas thank you from me and I'm sure everyone else you have immeasurably helped.

VID comes down, Vcore comes down, temps come down.  How much? 

I don't know, .. third BASE!

These companies have become ruthless and Newegg or Amazon would be the last place on the planet I would give any credence to tech reviews and responses after the last year or so of competitive debauchery when it comes to high desired tech product sales.

You can reset you watt load to 259/260, its fine

I would say if there is more room to work, get that penny size divot in the IHS as small as possible if not completely removed.

Note I said..  if there is room to work. If you are not sure, pull the board and check the clearance as any responsible engineer or computer technician/clocker, would. 

A flashlight and a good eyeball works for that unless its really tight.