I9 9900k

moorhen2 · 21 Nov 2018 at 10:52

Sorry about above post, trying another image hosting site as Photobucket down as bl###y usual.

Distracted · 21 Nov 2018 at 11:13

I've had a very brief opportunity to test a few 9900k .

All on a single 40mm 360 rad in a custom loop with VRMs and GPU in the same loop.
The board was an Asus Maximus XI Formula, RAM 3200 CL14 4x8GB, GPU 1080Ti, plus nvme boot drive.

LLC 7
Cache 4300
All limits maxed as per Der8auer youtube guide.

No Prime testing unless stated, just quick and dirty Cinebench runs due to limited time, no temp records either. Squandered opportunity I know.
With CPU #1 I got carried away and lost track of time :rolleyes:

Bear in mind that for every complete cinebench run I list below there were at least 3-5 failures prior.

5.0 GHz Cinebench @ 1.25v in Bios (may have gone lower, started @ 1.25v)
5.1 GHz Cinebench @ 1.28v in Bios (Did a couple of minutes with Prime 95 with -1 AVX offset)
5.2 GHz Cinebench @ 1.32v in Bios
5.2 GHz Cinebench @ 1.35v in Bios (Did a couple of minutes with Prime 95 with -1 AVX offset)

5.0 GHz Cinebench @ 1.35v in Bios (Did a couple of minutes with Prime 95 with -1 AVX offset)

5.0 GHz Cinebench @ 1.28v in Bios
5.1 GHz Cinebench @ 1.35v in Bios

None of these are my CPU. I might be able to add one more plus my own to the list later. Still waiting on a delivery of watercooling parts before I shift the PC again.

moorhen2 · 21 Nov 2018 at 12:22

Cache at 49 now, an hours run of Aida64 stress test. As usual no FPU/AVX, don't see the point.

Sem · 21 Nov 2018 at 14:10

Been doing more testing and it seems i have a pretty good chip...
so for 5.2 AVX i only need 1.28v to be Realbench stable

for 5.3 AVX @ 1.3v it passes cinebench and 3dmark but not realbench will probably need more volts or maybe reduce AVX by 1-2 but temps are already a problem

https://www.3dmark.com/spy/5098074
https://www.3dmark.com/fs/17110802

i tried 5.4 @ 1.32v and it passed 3d mark but not CB wont even attempt RB

https://www.3dmark.com/fs/17110927

i reckon i can get it CB stable at 5.4 i guess that kind of clock speed needs more tweaking than voltage and LLC only
maybe in afew days i will try it

now for my 24/7 setting things are pretty damn good
5Ghz prime and realbench stable (overnight runs)
when its not using AVX it needs 1.244v when i run AVX load volts jump up abit to 1.269 all rock stable so far
not tweaked any other settings just set adaptive voltage to 1.24v in the bios with LL6

0121danwilliams84 · 21 Nov 2018 at 14:20

...And you all instacrash in games ;-)

Distracted · 21 Nov 2018 at 14:38

0121danwilliams84 said:
I am finding that games are usually in the high 60's and 70's but I get random 1 second spikes into the high 80's. Average temps over an extended heavy game session are about 70c. Seem about right? or a bit on the warm side?

I don't know. Theres so much variance between cooling and silicon that i still have no idea what the norm for a 9900k is.

Distracted · 21 Nov 2018 at 14:39

0121danwilliams84 said:
...And you all instacrash in games

I read that as "and i hope you all instacrash in games ;-)

"

0121danwilliams84 · 21 Nov 2018 at 14:40

And you are all only stable at 4.9GHz in games ;-)

moorhen2 · 21 Nov 2018 at 15:13

Sem said:
Been doing more testing and it seems i have a pretty good chip...
so for 5.2 AVX i only need 1.28v to be Realbench stable

for 5.3 AVX @ 1.3v it passes cinebench and 3dmark but not realbench will probably need more volts or maybe reduce AVX by 1-2 but temps are already a problem

https://www.3dmark.com/spy/5098074
https://www.3dmark.com/fs/17110802

i tried 5.4 @ 1.32v and it passed 3d mark but not CB wont even attempt RB

https://www.3dmark.com/fs/17110927

i reckon i can get it CB stable at 5.4 i guess that kind of clock speed needs more tweaking than voltage and LLC only
maybe in afew days i will try it

now for my 24/7 setting things are pretty damn good
5Ghz prime and realbench stable (overnight runs)
when its not using AVX it needs 1.244v when i run AVX load volts jump up abit to 1.269 all rock stable so far
not tweaked any other settings just set adaptive voltage to 1.24v in the bios with LL6

With AVX used its not a true 5.2 overclock, but nice anyway.

Sem · 21 Nov 2018 at 15:27

moorhen2 said:
With AVX used its not a true 5.2 overclock, but nice anyway.

not sure what you mean
why does using avx not make it a true overclock when no offset is used

orbitalwalsh · 21 Nov 2018 at 15:30

^^^^^^

Distracted · 21 Nov 2018 at 15:37

Sem said:
Been doing more testing and it seems i have a pretty good chip...
so for 5.2 AVX i only need 1.28v to be Realbench stable

for 5.3 AVX @ 1.3v it passes cinebench and 3dmark but not realbench will probably need more volts or maybe reduce AVX by 1-2 but temps are already a problem

https://www.3dmark.com/spy/5098074
https://www.3dmark.com/fs/17110802

i tried 5.4 @ 1.32v and it passed 3d mark but not CB wont even attempt RB

https://www.3dmark.com/fs/17110927

i reckon i can get it CB stable at 5.4 i guess that kind of clock speed needs more tweaking than voltage and LLC only
maybe in afew days i will try it

now for my 24/7 setting things are pretty damn good
5Ghz prime and realbench stable (overnight runs)
when its not using AVX it needs 1.244v when i run AVX load volts jump up abit to 1.269 all rock stable so far
not tweaked any other settings just set adaptive voltage to 1.24v in the bios with LL6

Was this all with ll6?

Sem · 21 Nov 2018 at 15:49

Distracted said:
Was this all with ll6?

only changed to LLC6 during my 24/7 stability LLC5 was almost stable but prime crashed after 3 hours
before that for the other runs was either on auto or manually set between 2-4
LLC7 for the 5.4 run

Distracted · 21 Nov 2018 at 15:51

moorhen2 · 21 Nov 2018 at 16:54

Sem said:
not sure what you mean
why does using avx not make it a true overclock when no offset is used

Not knocking your overclock, its very good, but when AVX instructions kick in, the core multiplier will drop to 47, as per your screenshot. So it wont be running at 5.2 constantly.

What voltage are you setting in the bios by the way.

moorhen2 · 21 Nov 2018 at 17:03

Can I just say I would never dismiss or knock any overclock, we are all in the same game are we not. So apologies if I have offended in any way or user on here.

Sem · 21 Nov 2018 at 18:10

moorhen2 said:
Not knocking your overclock, its very good, but when AVX instructions kick in, the core multiplier will drop to 47, as per your screenshot. So it wont be running at 5.2 constantly.

What voltage are you setting in the bios by the way.

Please look again

Especially at the "current" tab on hwinfo
Realbench is running so the "current" load is 100% with AVX instructions and the "current" clock speed is 5.2ghz the 4.7ghz is only on the average tab which takes into account the lowest idle speed of 800mhz

Under sustained load it stays at 5.2

cpu-z and coretemp confirm this also

The voltage i set was either 1.275 or 1.3 i forgot but LLC would have been on 4 or 5 tho

robin66 · 21 Nov 2018 at 18:15

I've been running my chip at 5GHz on games for 2 days. Also ADIA 64 for a couple of hours. And not one crash.

This is at 1.36v however with no AVX offset. I had noticed the odd crash here and there at 1.35v.

Yeah, not the best of overclockers for sure.

Now that I know what is stable at 5GHz. I might knock it back down to stock for now until I feel I need a boost later down the road.

moorhen2 · 21 Nov 2018 at 19:12

Sem said:
Please look again

Especially at the "current" tab on hwinfo
Realbench is running so the "current" load is 100% with AVX instructions and the "current" clock speed is 5.2ghz the 4.7ghz is only on the average tab which takes into account the lowest idle speed of 800mhz

Under sustained load it stays at 5.2 cpu-z and coretemp confirm this also

The voltage i set was either 1.275 or 1.3 i forgot but LLC would have been on 4 or 5 tho

Ok, can we clarify that you are or you are not using AVX offset,?

The reason im asking is because if you were using an AVX offset, for instance -5, like I have just set in my bios for testing purposes, as soon as Realbench starts and the AVX intructions kick in the core clock will drop by a multiplier of 5, and hence a full load core clock of 4.7ghz from 5.2ghz.

So whatever AVX offset is used, this will be subtracted from the core clock under AVX load. And you did state 5.2 AVX.

Or do you mean 5.2 running a stress test that uses AVX.

Majek · 21 Nov 2018 at 19:53

robin66 said:
I've been running my chip at 5GHz on games for 2 days. Also ADIA 64 for a couple of hours. And not one crash.

This is at 1.36v however with no AVX offset. I had noticed the odd crash here and there at 1.35v.

Yeah, not the best of overclockers for sure.

Now that I know what is stable at 5GHz. I might knock it back down to stock for now until I feel I need a boost later down the road.

Hi,

I am so glad to see this post! I am sorry but it gives me some solace that there are other people too with 9900k who don't get stable 5.0 below 1.3v.

It just seems that any forum I visit everyone is doing 5.0 to 5.2 at 1.2+v.

What board are you on?

I'm on z390 Aorus Master and I suspect I may need more than 1.32v at 5.0 (my current, seemingly stable setting).