10850k OC help

[CasoF13]

AFter many struggels i got a working system with my 10850k on a rog strix z490g gaming wifi and a 760w psu fractal ion

because when i bouight the 850k i though i could get a 900k for ceap and in stock. now i am trying to oc currently i am all core 4.8 at a minus voltage offset of 0.03 and llc4. with r23 stresstest my temp creah a 85c max at peak at the cores with my fuma 2. i am new to oc and therfore  have read and read forumarticels what works and what not. any advice would be great. (for stresstesting i use r23 for 10m ,occt for 1h and realbench for 2h)

when trying to bump to 49x i reduced the voltage minusoffset to 0.015 - 0.01 but ether my voltage seams for a noob like my to close to max or my temps skyrocket to 93-94 at peak from what i read in the www is bad. in windows my coltage is reported by hwinfo be at 1.172-1.234 sometimes 1.252 with 85c peak

my goal would be a x49 or x50 stable if possible

 

i also trying asus ai auto oc at it clocks cores differntly some to 5.2 i beleive with a adapt vcore voltage of 1.4... what doesnt seams to be safe so i turned it off right away

 

any help would be great.

[Guest]

You won't get more speed by reducing v-core. You will be closer to 1.3v at 5ghz maybe a touch more. 

So you should look to reduce the temps as much as possible first. But you're doing good so far!

 

What's the case and configuration of the cooling??

[Somerandomtechyboi]

13 minutes ago, CasoF13 said:

AFter many struggels i got a working system with my 10850k on a rog strix z490g gaming wifi and a 760w psu fractal ion

because when i bouight the 850k i though i could get a 900k for ceap and in stock. now i am trying to oc currently i am all core 4.8 at a minus voltage offset of 0.03 and llc4. with r23 stresstest my temp creah a 85c max at peak at the cores with my fuma 2. i am new to oc and therfore  have read and read forumarticels what works and what not. any advice would be great. (for stresstesting i use r23 for 10m ,occt for 1h and realbench for 2h)

when trying to bump to 49x i reduced the voltage minusoffset to 0.015 - 0.01 but ether my voltage seams for a noob like my to close to max or my temps skyrocket to 93-94 at peak from what i read in the www is bad. in windows my coltage is reported by hwinfo be at 1.172-1.234 sometimes 1.252 with 85c peak

my goal would be a x49 or x50 stable if possible

 

i also trying asus ai auto oc at it clocks cores differntly some to 5.2 i beleive with a adapt vcore voltage of 1.4... what doesnt seams to be safe so i turned it off right away

 

any help would be great.

Yea theres no headroom anymore with these new platforms so id suggest just setting a high static freq like you are doing right now and just running a relatively low volt, diminishing returns zone is at around 5ghz for these cpus so you are good for now

 

Just dont bother ocing the cpu anymore since 5ghz 1.25 is already pretty good, now just move on to overclocking the rams

[CasoF13]

my case is a meshify 2 compact ui revert most settings so now i am sitting at 4.8ghz at a minusvoltageoffset of 0.03 with stresstest max temp at peak 85.

--------

so for stresstesting is a temp of over 90 ok ? with the 850k because with a setting of

4.9ghz at minusvoltage of 0.01 my stresstest temp went in the cinebench alone already at past 90.

 

sorry for the questions but is my first oc and my first pc build ever

[Somerandomtechyboi]

7 minutes ago, CasoF13 said:

my case is a meshify 2 compact ui revert most settings so now i am sitting at 4.8ghz at a minusvoltageoffset of 0.03 with stresstest max temp at peak 85.

--------

so for stresstesting is a temp of over 90 ok ? with the 850k because with a setting of

4.9ghz at minusvoltage of 0.01 my stresstest temp went in the cinebench alone already at past 90.

 

sorry for the questions but is my first oc and my first pc build ever

Surprising that a newbie would try ocing, but i guess new bioses are much simpler now so that helps

 

Tbh its only 1.25v, i wouldnt really worry up until ~100-105c, 85c and below is prefferable for longevity and stability, but in this circumstance id say its fine to run it ~95c, 105c is also fine but its abit sketch so i would not reccomend that

 

at a higher volt like 1.4 or higher you def want the temps  to be 85c or lower for longevity and stability, go even higher to 1.52v (intel spec max volt) or the absolute max "safe" of 1.6v (not reccomended) then you will want 75c or lower, pref 70c cause at this point if your temp is too high your system will destabilize

[jones177]

In the tests I did with my i9 10900k 4.9ghz all cores is equal to a 5gz all cores on a i9 9900k or i7 8086k in gaming performance. With TVB on(below 70c) it beats them.

 

The get a better result with a traditional overclock, 5.1ghz is needed. I had no cooling issues at 5.1ghz with a 360mm AIO so very doable. 

 

I am not a fan of Cinebench for testing since at 5.3ghz all cores my i9 10900k stays relatively cool but it throttles in Time Spy. It is good for comparing scores but that is about it.

 

[CasoF13]

7 minutes ago, Somerandomtechyboi said:

Surprising that a newbie would try ocing, but i guess new bioses are much simpler now so that helps

 

Tbh its only 1.25v, i wouldnt really worry up until ~100-105c, 85c and below is prefferable for longevity and stability, but in this circumstance id say its fine to run it ~95c, 105c is also fine but its abit sketch so i would not reccomend that

 

at a higher volt like 1.4 or higher you def want the temps  to be 85c or lower for longevity and stability, go even higher to 1.52v (intel spec max volt) or the absolute max "safe" of 1.6v (not reccomended) then you will want 75c or lower, pref 70c cause at this point if your temp is too high your system will destabilize

so if i got over90c in stresstesting at a voltage belowe 1.4 all good ? also i read that the vcore in hwinfo is not the actual vcore should i add 0.1 to my calc ? so for example if hwin says 1.12 the cpu is more lilkly at 1.22 ?

[Mister Woof]

I would use something like Real Bench to stress test + your actual apps. Cinebench will do all-core CPU, but zero GPU related heat. If you have a bigger modern GPU, when under load, it will heat up your CPU a few degrees more. Realbench stresses both components with real world tasks, so it's a fairer more realistic representation of your uses.

 

CB R23 isn't really a stress test beyond being fast to find obvious instabilities before attempting something more intense.

 

As for voltages: I was able to get 5.1ghz all-core on my 10900KF with as little as 1.32v with 0 AVX; however, to get up to 5.3ghz stable in lighter loads, I had to bump it up to almost 1.4v. Of course, under all-core stress, it vdroops to under 1.3v and 5.1ghz.

 

With a 10850k, i'd suspect slightly worse results. Maybe 100mhz lower at the same voltages. Each sample varies, of course.

 

The Fuma 2 is a great cooler but probably something I'd use for stock operation on a big 10 core chip, tbh.

[Montana One-Six]

12 minutes ago, CasoF13 said:

my case is a meshify 2 compact ui revert most settings so now i am sitting at 4.8ghz at a minusvoltageoffset of 0.03 with stresstest max temp at peak 85.

So basically at stock you get 85C which seems kinda high. My 10850k reaches ~75C max during stress-tests. The Fuma 2 is marginally worse than the NH-D15 but not that much. So I would investigate that before anything else.

[CasoF13]

1 minute ago, Montana16 said:

So basically at stock you get 85C which seems kinda high. My 10850k reaches ~75C max during stress-tests. The Fuma 2 is marginally worse than the NH-D15 but not that much. So I would investigate that before anything else.

yeah my stresstest and cinebench and realbench 2h maxs out at 85c with all core fixed at 4.8ghz with the minusvoltae alrady mentioned

[CasoF13]

oh and by the way i have a rtx 3060 gigybyte so i belkeive that is now shoudnt cause much heat (i know a 3060 is "trash" but my old 3070ti was faulty and the stock/prize situation leads my to buy a 3060 which i got from a local retailer at 510 euro)

[Mister Woof]

4 minutes ago, CasoF13 said:

yeah my stresstest and cinebench and realbench 2h maxs out at 85c with all core fixed at 4.8ghz with the minusvoltae alrady mentioned

Other than checking mounting pressure, I'm not sure those temperatures are really unreasonable. It's still a pretty hot chip.

[CasoF13]

So here a clean overview of my system.

10850k

ROG Strix z490 g Gaming Wifi

2tb Nvme SN550 WD Blue

500TB Nvme XPG Spectrix S40G Systemdrive

1TB SSD SanDisk DH3

4TB Toshiba X300 HDD

6 Case fans CF120 Deepcool

RTX 3060 Gigybyte Eagle OC

64GB HyperX DDR4 Ram set to XMP 1 at 3200mhz 4x16GB Sticks

FUMA 2 with fans replaced with a 120mmx15 silverstone argb fan front and a coolermaster sickleflow 120 in the middle.

Meshify 2 Compact Case

PSU 760W Fractal Ion Platinum

[Mister Woof]

Have you tried the Fuma2 with its original fans?

[CasoF13]

1 minute ago, Mister Woof said:

Other than checking mounting pressure, I'm not sure those temperatures are really unreasonable. It's still a pretty hot chip.

So which temps are fine in daily use with a 10850k max 85 or more or less ? like i said my goal was a 10900 for cheap thats the main reason why i want to oc

[CasoF13]

Just now, Mister Woof said:

Have you tried the Fuma2 with its original fans?

Not jet should i replace the middle fan with the original or the front ? (because of the argb)

[Mister Woof]

2 minutes ago, CasoF13 said:

So which temps are fine in daily use with a 10850k max 85 or more or less ? like i said my goal was a 10900 for cheap thats the main reason why i want to oc

Well stress test temps are not typically representative of real life usage temps unless your uses are rendering all day long.

 

If you're just general use and gaming, you won't be nearly as hot as what those temps are. Stress test temps typically represent the worst case scenario. 

 

Safe and max temps are all relative to your cooling and your use cases.

 

For example I would personally feel that a cooling setup that throttles under P95 but stays within normal operating temperature during your actual applications to be "within spec" because with the stress test, you are asking more of it than you need it to do. 

 

Like transmissions on a car - you can thrash them by towing a buncha shit in low speed and high gear on a passenger car for as long as you can to see if it survives....but that's never gonna happen and doesn't make sense to judge it based on that criteria. 

[CasoF13]

4 minutes ago, Mister Woof said:

Well stress test temps are not typically representative of real life usage temps unless your uses are rendering all day long.

 

If you're just general use and gaming, you won't be nearly as hot as what those temps are. Stress test temps typically represent the worst case scenario. 

 

Safe and max temps are all relative to your cooling and your use cases.

so if a oc and stresstest for stability like with realbench for 2h and got over 90c at 49x its fine ? or should i stay at 48x with 85c (also realbench test)

at gaming temps at 48x seams to max out at 68c and idle at 35c

[Mister Woof]

1 minute ago, CasoF13 said:

so if a oc and stresstest for stability like with realbench for 2h and got over 90c at 49x its fine ? or should i stay at 48x with 85c (also realbench test)

What is your actual use case

[CasoF13]

Gaming for now. Maybe additional Creative work in the future (photoshop blender videoediting ...)

[Mister Woof]

2 minutes ago, CasoF13 said:

Gaming for now. Maybe additional Creative work in the future (photoshop blender ...)

I'd leave it as is (90c+ in stress is fine) while you're just gaming. Try your other tasks. If it gets too hot doing those tasks....then you can readdress your situation.

 

But I don't think anything is "wrong" with your current configuration. The closest you'll get in gaming to your stress test temps is during unpacking/shader optimization and that doesn't happen to often. Most other games aren't going to saturate a 10 core chip.

[CasoF13]

3 minutes ago, Mister Woof said:

I'd leave it as is (90c+ in stress is fine) while you're just gaming. Try your other tasks. If it gets too hot doing those tasks....then you can readdress your situation.

 

But I don't think anything is "wrong" with your current configuration.

ok thanks so i can try 49x and if it is hot not only during stresstest i should dial it back to 48x (like right now) and call it a day. also you write 90+ during stresstest is fine so if a stresstest has 90+ during 2 h the cpu should survivce it ? or is the heat for this long bad for the cpu

 

also what is considers hot in normal daily use ?

[Mister Woof]

3 minutes ago, CasoF13 said:

ok thanks so i can try 49x and if it is hot not only during stresstest i should dial it back to 48x (like right now) and call it a day. also you write 90+ during stresstest is fine so if a stresstest has 90+ during 2 h the cpu should survivce it ? or is the heat for this long bad for the cpu

TJ Max is 100c (per Intel, but others have said 115), so anything under (I would go with 100) that is within spec. Most laptops are bouncing off of this metric on the reg. It will throttle itself if it gets too hot.

[CasoF13]

so aslong cpu is under tjmax and vcore isnt to high and the system is stable i am good to go

also what is considers hot in normal daily use ? just want to know for a reference point

[Mister Woof]

1 minute ago, CasoF13 said:

so aslong cpu is under tjmax and vcore isnt to high and the system is stable i am good to go

also what is considers hot in normal daily use ? just want to know for a reference point

personally i start feeling uncomfortable if my desktop system hits 85+

 

but my laptop as I stated is hot AF and I don't care