9900K delid overcloking at 5.2ghz or 5.3... What you think about those settings ? - Page 2 - Overclock.net - An Overclocking Community

Forum Jump: 

9900K delid overcloking at 5.2ghz or 5.3... What you think about those settings ?

Reply
 
Thread Tools
post #11 of 19 (permalink) Old 01-06-2019, 07:31 PM
Watercooled
 
lexer's Avatar
 
Join Date: Jul 2015
Location: Argentina
Posts: 865
Rep: 47 (Unique: 33)
Don't use AUTO setting for VCCIO / SA, With your RAM at DDR4 4266 it will set a very high voltage. Try with VCCIO 1.15 / VCCSA 1.20 if crashes 1.20/1.25. LCC is too high in my opinion try with LCC5, very high LCC increases temperatures.
Is up too you but I don't see a reason to leave the processor with manual voltage and locked to 5.3Ghz all the time. You can try this if you want. it will reduce slightly the temperature in partial loads and the power consumption at IDLE.

>SVID Beheivor: Best Case
>LCC5 o 6 max
>Voltage Adaptive > start with 1.26 but you will have to play with this value. With Adaptive the voltage behaves a little weird some times but is just a matter of trying and get a the right value. Control the Voltages and temps with HWINFO64
>AVX Offset - Personally i'm not a fan of AVX offset. But if processor can run 5.2Ghz AVX but crashes at 5.3Ghz, you can try negative -1

For stressing you can use HWBOT X265 / OCCT or since you have a 2080 Ti play Battlefield 5 that games uses AVX and is VERY cpu intensive and also sensible to unstable O.C. AIDA64 is too "light" in my opinion

I7 8700K@4.7Ghz / Z370-A / 2X8GB 3200Mhz ADATA Spectrix D40 / 1TB Intel 600P / EVO 850 250GB / 4TB WD Gold / RTX 2080 GAMING OC / NZXT Noctis 450 ROG / EVGA G2 850W@ECO
Heatkiller REV 3.0 / XSPC EX360 / DDC 3.25 + EK-XRES / 3x Venturi HF-12 PWM
lexer is offline  
Sponsored Links
Advertisement
 
post #12 of 19 (permalink) Old 01-06-2019, 07:40 PM
Overclocked
 
scracy's Avatar
 
Join Date: Aug 2014
Location: Cyberspace
Posts: 1,995
Rep: 116 (Unique: 83)
Quote: Originally Posted by lexer View Post
Don't use AUTO setting for VCCIO / SA, With your RAM at DDR4 4266 it will set a very high voltage. Try with VCCIO 1.15 / VCCSA 1.20 if crashes 1.20/1.25. LCC is too high in my opinion try with LCC5, very high LCC increases temperatures.
Is up too you but I don't see a reason to leave the processor with manual voltage and locked to 5.3Ghz all the time. You can try this if you want. it will reduce slightly the temperature in partial loads and the power consumption at IDLE.

>SVID Beheivor: Best Case
>LCC5 o 6 max
>Voltage Adaptive > start with 1.26 but you will have to play with this value. With Adaptive the voltage behaves a little weird some times but is just a matter of trying and get a the right value. Control the Voltages and temps with HWINFO64
>AVX Offset - Personally i'm not a fan of AVX offset. But if processor can run 5.2Ghz AVX but crashes at 5.3Ghz, you can try negative -1

For stressing you can use HWBOT X265 / OCCT or since you have a 2080 Ti play Battlefield 5 that games uses AVX and is VERY cpu intensive and also sensible to unstable O.C. AIDA64 is too "light" in my opinion
Agreed LLC 5 or 6 would be better (I use LLC 5), personally I would prefer a higher Vcore without load and let Vdroop do its thing when there is load present that way you reduce the chance of your CPU getting voltage spikes under load
Eg: My 9900K UEFI setting for Vcore is 1.344V for 5.2Ghz however using LLC 5 my Vcore drops to where it needs to be for stability to 1.328V under load, its better to let LLC drop the Vcore under load than let LLC push the Vcore up if that makes sense?

ROG Maximus X Formula, Intel Core [email protected]/4.8Ghz Cache 1.344V (OCCT Load) (OCCT Stable), Corsair Dominator Platinum ROG Edition [email protected], EVGA GTX1070 SLI, Samsung 960 Pro x2 (Raid 0) (O/S), Samsung 850 Pro x4 (Raid 0) (Data)
ROG THOR 850W PSU, EK Custom Loop 2x PE-360 Radiators, Corsair ML120 Pro Fans, Phanteks Evolv X, Creative Soundblaster Z, Woo Audio WA7D, Sennheiser HD660S, ROG PG279Q 2560x1440 165Hz IPS Monitor, Windows 10 Pro Build 1809.

Last edited by scracy; 01-06-2019 at 07:44 PM.
scracy is offline  
post #13 of 19 (permalink) Old 01-06-2019, 08:07 PM
Watercooled
 
lexer's Avatar
 
Join Date: Jul 2015
Location: Argentina
Posts: 865
Rep: 47 (Unique: 33)
I never had spikes while using Adaptive with the right LCC and AI DC/AC setting and my board is a fake doubled 8 phases. Just for curiosity I put my FLUKE 179 right on the backside of the socket and didn't register any considerable spike. With a oscilloscope probably I can see the spikes, since the multi-meter refresh rate is around 250/300ms but the ADC converter of the FLUKE still miles better than the motherboard sensor. On motherboards with very low-end VRM or not very good BIOS for sure can be a issue

I7 8700K@4.7Ghz / Z370-A / 2X8GB 3200Mhz ADATA Spectrix D40 / 1TB Intel 600P / EVO 850 250GB / 4TB WD Gold / RTX 2080 GAMING OC / NZXT Noctis 450 ROG / EVGA G2 850W@ECO
Heatkiller REV 3.0 / XSPC EX360 / DDC 3.25 + EK-XRES / 3x Venturi HF-12 PWM
lexer is offline  
Sponsored Links
Advertisement
 
post #14 of 19 (permalink) Old 01-06-2019, 08:36 PM
New to Overclock.net
 
white owl's Avatar
 
Join Date: Apr 2015
Location: The land of Nod
Posts: 5,347
Rep: 135 (Unique: 102)
Quote: Originally Posted by lexer View Post
I never had spikes while using Adaptive with the right LCC and AI DC/AC setting and my board is a fake doubled 8 phases. Just for curiosity I put my FLUKE 179 right on the backside of the socket and didn't register any considerable spike. With a oscilloscope probably I can see the spikes, since the multi-meter refresh rate is around 250/300ms but the ADC converter of the FLUKE still miles better than the motherboard sensor. On motherboards with very low-end VRM or not very good BIOS for sure can be a issue
How can it be fake if it's doubled?


As for the OP, unless you need to change something you can leave it.
There really isn't a wrong way to do all of this as long as it's stable and safe IMO. I would recommend a real stress test like P95 (no AVX) though.


Everything is pretty much summed up here:
https://www.overclock.net/forum/5-in...tatistics.html

Quote: Originally Posted by SpeedyVT
If you're not doing extreme things to parts for the sake of extreme things regardless of the part you're not a real overclocker.
Quote: Originally Posted by doyll View Post
The key is generally not which brands are good but which specific products are. Motherboards and GPUs are perfect examples of companies having everything from golden to garbage function/quality.
Hot n Bothered
(12 items)
CPU
4790k 4.7Ghz
Motherboard
Asus Sabertooth Z97 MkII 2
GPU
EVGA GTX 1080 SC
RAM
16gb G.Skill Sniper 2400Mhz
Hard Drive
2x Kingston v300 120gb RAID 0
Hard Drive
WD Blue
Power Supply
Seasonic 620w M12 II EVO
Cooling
Cooler Master 212 Evo
Case
Corsair 450D
Operating System
Windows 10
Monitor
Nixeus EDG27
Other
I have pretty lights.
▲ hide details ▲

Last edited by white owl; 01-06-2019 at 08:40 PM.
white owl is offline  
post #15 of 19 (permalink) Old 01-06-2019, 09:05 PM
New to Overclock.net
 
Falkentyne's Avatar
 
Join Date: Dec 2007
Location: Riverside
Posts: 5,667
Rep: 370 (Unique: 263)
Quote: Originally Posted by white owl View Post
How can it be fake if it's doubled?


As for the OP, unless you need to change something you can leave it.
There really isn't a wrong way to do all of this as long as it's stable and safe IMO. I would recommend a real stress test like P95 (no AVX) though.


Everything is pretty much summed up here:
https://www.overclock.net/forum/5-in...tatistics.html
Fake doubled= using two POWER stages per phase rather than a phase DOUBLER.

[email protected] ghz, RX Vega 64, 32GB DDR4, Gigabyte Aorus Master, Seasonic Platinum 1000W, Corsair 760T
Alt: MSI GT73VR Throttlebook with 7820HK @ 4.7 ghz, GTX 1070 MXM TDP mod to 230W, 32 GB RAM
Falkentyne is offline  
post #16 of 19 (permalink) Old 01-06-2019, 09:28 PM
New to Overclock.net
 
white owl's Avatar
 
Join Date: Apr 2015
Location: The land of Nod
Posts: 5,347
Rep: 135 (Unique: 102)
Quote: Originally Posted by Falkentyne View Post
Fake doubled= using two POWER stages per phase rather than a phase DOUBLER.
Interesting. I think a reviewer (like GN) should review boards with the same CPU each time so we can see how the VRM OC's rather than judging them by phase count. I find it odd that we have PSU reviewers that test PSUs but no one reviews motherboards in this way.
Knowing how precise the voltage readings are would also be helpful.

Quote: Originally Posted by SpeedyVT
If you're not doing extreme things to parts for the sake of extreme things regardless of the part you're not a real overclocker.
Quote: Originally Posted by doyll View Post
The key is generally not which brands are good but which specific products are. Motherboards and GPUs are perfect examples of companies having everything from golden to garbage function/quality.
Hot n Bothered
(12 items)
CPU
4790k 4.7Ghz
Motherboard
Asus Sabertooth Z97 MkII 2
GPU
EVGA GTX 1080 SC
RAM
16gb G.Skill Sniper 2400Mhz
Hard Drive
2x Kingston v300 120gb RAID 0
Hard Drive
WD Blue
Power Supply
Seasonic 620w M12 II EVO
Cooling
Cooler Master 212 Evo
Case
Corsair 450D
Operating System
Windows 10
Monitor
Nixeus EDG27
Other
I have pretty lights.
▲ hide details ▲
white owl is offline  
post #17 of 19 (permalink) Old 01-08-2019, 03:10 PM
Memory Overclocker
 
Join Date: Nov 2018
Posts: 467
Rep: 23 (Unique: 9)
Quote: Originally Posted by ThrashZone View Post
Hi,
Not sure if 99..k is the same as 8700k but cache is supposed to be within 3 of core multiplier for stability
If you want 5.2 core, cache max should be no lower than 4.9 supposedly.
On my 9700k, running at 5.0, I cannot get stability with the cache above 4.5. The moment I raise the cache to 4.6, Prime95/Blend w/no AVX will BSOD quickly, no matter how high I raise CPU voltage. I have CPU voltage at Fixed/1.37V now (which give 1.28-1.34 on VRVOUT under Prime95 load). Drop the cache back to 4.5 and I'm stable for as long as I run Prime95.

Motherboard is ASRock Z390 Taichi, 2x8 memory running at 4000/17-17-17-37. VCCIO and VCCSA at 1.18.

Is there anything else I can do besides CPU voltage to help make it stable with a higher cache speed?
jfriend00 is offline  
post #18 of 19 (permalink) Old 01-13-2019, 04:59 AM
New to Overclock.net
 
fat4l's Avatar
 
Join Date: Nov 2014
Posts: 1,439
Rep: 73 (Unique: 38)
All I can say is .....
Do NOT use AIDA 64 for stressing as it is not stressful enough.

Use either prime 95 26.6 1344k to 1344k, in place FFTs checked. Also u can use Realchench stress test, 2.44 or 2.56. Version 2.56 is more stressful as it uses (afaik) AVX so it will probably downclock your cpu if you have it set up in the bios(avx offset) so watch out on that.
First test if you can pass 1 hour, then if you are happy test 8 hours.

Anyway....if you pass prime or realbench for 8Hours, you are good to go.

PS:
For me for example, i7 4790K, 5100MHz..
.....for realbench 2.44, I need 1.335V for 1 hour.
But for realbench 2.56, I need 1.347v for 1 hour.
You can see the difference(Bear in mind the cpu is always at 5100MHz no matter of AVX).

Prime 95 26.6, 1344k, is similar to realbench 2.44, in terms of "how hard the test is". Temps are also similar.

Good luck!

CPU: Intel i7 4790K 5100MHz @1.335V MB: Asus ROG Maximus VII Formula RAM: 4x4GB Corsair Dominator Platinum 2666MHz CL10 VGA: Asus ROG GTX 1080Ti Strix 2152MHz @1.131V SDD: Samsung 870 PRO 1TB NVME + 860 EVO 1TB PSU: Asus ROG Thor Platinum 850W MON: Asus ROG Swift PG278Q 27"
fat4l is offline  
post #19 of 19 (permalink) Old 03-22-2019, 11:40 AM - Thread Starter
Trying to learn ^^
 
Fraizer's Avatar
 
Join Date: Oct 2012
Posts: 338
Rep: 0
finaly i fixed some problem on my waterccooling

you advise me here to not use aida for stress test then i follow your advise ^^ like occt 1hou and realbench 2.44 non avx for 1hour but some wrote i must run it for 8 hours and full memory (i have 32gb) some half memory

originaly i have Asus Maximus XI Extreme, my cpu delid direct die under custom watercooling : 9900k @5.3ghz @1.320v with cpu cache at x49. VCCIO 1.2500 VSA 1.2500. i am on LLC 7 (the asus need to be at 7 not 6 or 5). i disable the c-states etc...

- 8h stress test under OCCT = ok stable

but i dont understand wat happen, when i do stress test under realbench 2.44 non avx.... i put the vcore at finaly 1.325v and i make an stress test for 2h = is ok stable another time i think it was almost 4h before i get an blue screen.... another test (always after an full reboot) 2min and i have an blue screen...i increase at 1.330v is the same... under realbnench i go to 89°c on few core

+ something to do with the "CPU PLL" who is by default at auto ? must to fix it at the default valu 1.050v ?..

+ i am not using the right stress test sofware or setting ? or time of stres test who 1 hour is enough ?


note:

- i put an low speed and timing of the memory to not get problem on the test

- my graphic card is not on the custom watercooling

Last edited by Fraizer; 03-22-2019 at 11:43 AM.
Fraizer is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off