13700KF too hot? Dodgy CPU or Dodgy Cooler?

Associate
Joined
7 Jul 2016
Posts
124
Hi folks

Recently upgraded my PC and got a 13700KF

Now I know these chips run hot, but how hot how quick is considered normal?

Mine hits 100c almost instantly in Cinebench23 at stock settings - this is using a Arctic liquid freezer ii 280

The only way I can get it to not hit 100c and throttle for a run is to undervolt with a -0.125 and it still gets to 95-97ish

I have read that people are getting 85c runs using stock settings?? This is beyond me with what I am seeing - is it possible?

I idle about 32c in a 20c room. Games depend on the game but sit around 60 - 70c

I didnt know about the possible IHS bending issue before buying and installing, but since then I have bought one of the Thermalrite contact plates and using that, and the results are as above with it. I have reseated the pump multiple times now with various methods for applying paste and its never any better

So is this normal? Is my CPU broken? Is my Cooler broken?

Cooler was used previously on a 5900x and was fine there so cant see it being the cause but (?)

Any advice please before I drive myself crazy trying to "fix" this
 
PL1 and PL2, I had to put them on the default 253w manually. I hit 94c on cinebench.
I have set them to 253w and 300w

Also disabled multicore enhancement

currently able to pull cine23 run around 95c with a score of 30900

HWinfo showing a pull of around 210w during the run

Its truely crazy how hot how quick these things get
 
Hi folks

Recently upgraded my PC and got a 13700KF

Now I know these chips run hot, but how hot how quick is considered normal?

Mine hits 100c almost instantly in Cinebench23 at stock settings - this is using a Arctic liquid freezer ii 280

The only way I can get it to not hit 100c and throttle for a run is to undervolt with a -0.125 and it still gets to 95-97ish

I have read that people are getting 85c runs using stock settings?? This is beyond me with what I am seeing - is it possible?

I idle about 32c in a 20c room. Games depend on the game but sit around 60 - 70c

I didnt know about the possible IHS bending issue before buying and installing, but since then I have bought one of the Thermalrite contact plates and using that, and the results are as above with it. I have reseated the pump multiple times now with various methods for applying paste and its never any better

So is this normal? Is my CPU broken? Is my Cooler broken?

Cooler was used previously on a 5900x and was fine there so cant see it being the cause but (?)

Any advice please before I drive myself crazy trying to "fix" this
Hi there.

There are a few things you can do for 13th gen cpus as out of the box and with an out of the box motherbaord they will run super hot. This is nowhere ideal but I have tip and trick you can try to see if that helps and let me know how you get on.

First Off disable MCE aka Multi core Enhancement.
Second enforce all bios limits.

Now I read you use an offset of -0.125 is this correct? And you say you are using a contact frame. I use the Thermalright as well although they all do the same job. The drop in temps is minor by itself but combined with all the other adjustments in bios it can all add up.

I run a 13900k in one setup and a 12900k in another so both hot running cpu's by design.

On your motherboard do you have 2x EPS12v sockets? Now as strange as this may seem if you only have one in try plugging the other one in as well even though you are not overclocking you are undervolting and making bios adjustments so semantically speaking you are 'overclocking' I am not sure why but when I plugged both EPS 12v in the temps drop. Maybe a more experienced overclocker explain but give it a whirl.

Next we need loadline calibration setting I think on Asus motherboards it set to 3 but you want to set it to 4. We then need to find 2 settings to do with loadline calibration. One is called AC_LL and the other is called DC_LL, ocate these and change the values. I lowered mine but I can't remember what I lowered to. I will have to reboot and edit later.

***Edit***

Right after rebooting and searching for it, it is in the Internal CPU power management on Asus motherboards.
Change this setting IA AC Load Line to 0.30 and the second setting IA DC Load Line to 1.02. Now you can get these lower but that involves changing one setting at the time rebooting, testing for stability and so on. Some peeps have got theirs really low and it helps with the temps.

Also in windows check out what power plan you are on just to make sure. I run mine on Balanced.

next we need to locate system agent voltage. I have two motherboards by the same manufacturer and the sys agent voltage is different. You can either used fixed voltage or an offset your call. If it higher than 1.25v then lower to 1.25v. You can get it lower but you need to keep retesting for stability. I think mine is at 1.20v on both boards now.

Also i would suggest downloading XTU and HWinfo64. You can make basic changes in XTU in real time. But most of the other settings you will have to go into bios. Also could I suggest altering one thing at a time and then testing for stability that way you can determine stability and what setting you changed that made it unstable. If you do the whole lot in one go you won't know which setting tripped it up. So by doing the above steps you should be able to drop quite a bit. When I first got my 13900k it was like your run CB23 and straight to 100c. Now I am not saying you will get the same result as me it could be slightly worse or maybe even better (fingers crossed)

So in a cb23 runs i max out at 81-82c with a room ambient of 20-22c ish. funnily enough I just leave PL1 and Pl2 as stock as is. I am hoping my setting changes in bios will do enough as i am not really a fan keep changing the Pl1/pl2 depending on my use case even though it fast and painless to do in a program like Xtu.

Hope that helps.
 
Last edited:
Hi there.

There are a few things you can do for 13th gen cpus as out of the box and with an out of the box motherbaord they will run super hot. This is nowhere ideal but I have tip and trick you can try to see if that helps and let me know how you get on.

First Off disable MCE aka Multi core Enhancement.
Second enforce all bios limits.

Now I read you use an offset of -0.125 is this correct? And you say you are using a contact frame. I use the Thermalright as well although they all do the same job. The drop in temps is minor by itself but combined with all the other adjustments in bios it can all add up.

I run a 13900k in one setup and a 12900k in another so both hot running cpu's by design.

On your motherboard do you have 2x EPS12v sockets? Now as strange as this may seem if you only have one in try plugging the other one in as well even though you are not overclocking you are undervolting and making bios adjustments so semantically speaking you are 'overclocking' I am not sure why but when I plugged both EPS 12v in the temps drop. Maybe a more experienced overclocker explain but give it a whirl.

Next we need loadline calibration setting I think on Asus motherboards it set to 3 but you want to set it to 4. We then need to find 2 settings to do with loadline calibration. One is called AC_LL and the other is called DC_LL, ocate these and change the values. I lowered mine but I can't remember what I lowered to. I will have to reboot and edit later.

***Edit***

Right after rebooting and searching for it, it is in the Internal CPU power management on Asus motherboards.
Change this setting IA AC Load Line to 0.30 and the second setting IA DC Load Line to 1.02. Now you can get these lower but that involves changing one setting at the time rebooting, testing for stability and so on. Some peeps have got theirs really low and it helps with the temps.

Also in windows check out what power plan you are on just to make sure. I run mine on Balanced.

next we need to locate system agent voltage. I have two motherboards by the same manufacturer and the sys agent voltage is different. You can either used fixed voltage or an offset your call. If it higher than 1.25v then lower to 1.25v. You can get it lower but you need to keep retesting for stability. I think mine is at 1.20v on both boards now.

Also i would suggest downloading XTU and HWinfo64. You can make basic changes in XTU in real time. But most of the other settings you will have to go into bios. Also could I suggest altering one thing at a time and then testing for stability that way you can determine stability and what setting you changed that made it unstable. If you do the whole lot in one go you won't know which setting tripped it up. So by doing the above steps you should be able to drop quite a bit. When I first got my 13900k it was like your run CB23 and straight to 100c. Now I am not saying you will get the same result as me it could be slightly worse or maybe even better (fingers crossed)

So in a cb23 runs i max out at 81-82c with a room ambient of 20-22c ish. funnily enough I just leave PL1 and Pl2 as stock as is. I am hoping my setting changes in bios will do enough as i am not really a fan keep changing the Pl1/pl2 depending on my use case even though it fast and painless to do in a program like Xtu.

Hope that helps.
Thanks for the detailed write up!

I am using a -0.125 offset yes

All the power sockets on the board were already connected

When I tried LLC 4 and the IA/DC/AC settings it seemed to make the thermals worse for some reason, suspect vdroop not as high? I will have another play with this tomorrow to see properly though.

I had already set the SA to 1.20v

What cooler do you have?
 
@B1gdeano I am trying to look over what I posted and what you are posting now. So from what I see your idle temps are the same as mine and the gaming temps seems close to mine. So atm I am gonna rule out the cpu or cooler be faulty.

The -0.125v offset has piqued my curiosity though. Are you using an Asus MB? If you are can you tell me what your SP rating is? Now generally speaking most peeps I converse with about 13th gen temps don;t have that high an offset. I think I am on like -0.07v on my 12900k and about -0.08/-0.09 on my 13900k. So there are two possibilties. Either A) you have a very good cpu which can run -0.125v offset Which doesn;t explain your high temps or B) you have undervolt protection turned on in the bios which bypasses the -0.125v offset and that would explain your high temps.

Oh and to answer your last question I use a Corsair h150i.
 
Last edited:
Not faulty at all its exactly what I'd expect you've gone from a cpu with a max tdp of 105w to one with 253w thats twice the amount of power all that power has to go somewhere and it comes out as heat I had an arctic 280 with a 5900x and it was barely enough to contain it it could hit thermal limit under cinebench with the intel well its not hard to work out its one of the things that put me off them. Actually its the thing.
 
@B1gdeano I am trying to look over what I posted and what you are posting now. So from what I see your idle temps are the same as mine and the gaming temps seems close to mine. So atm I am gonna rule out the cpu or cooler be faulty.

The -0.125v offset has piqued my curiosity though. Are you using an Asus MB? If you are can you tell me what your SP rating is? Now generally speaking most peeps I converse with about 13th gen temps don;t have that high an offset. I think I am on like -0.07v on my 12900k and about -0.08/-0.09 on my 13900k. So there are two possibilties. Either A) you have a very good cpu which can run -0.125v offset Which doesn;t explain your high temps or B) you have undervolt protection turned on in the bios which bypasses the -0.125v offset and that would explain your high temps.

Oh and to answer your last question I use a Corsair h150i.
Yeah its an Asus board

I actually dropped the undervolt to -0.120 last night as I was getting intermittant crashing in a game that I suspected was possibly this, so dropped it a bit to see.

Im using adaptive negative offset

The SP score is 85 and undervolt protection is disabled
 
Not faulty at all its exactly what I'd expect you've gone from a cpu with a max tdp of 105w to one with 253w thats twice the amount of power all that power has to go somewhere and it comes out as heat I had an arctic 280 with a 5900x and it was barely enough to contain it it could hit thermal limit under cinebench with the intel well its not hard to work out its one of the things that put me off them. Actually its the thing.

Yeah I knew they were hot, but I have read people are able to do cine23 10 min runs and average between 85 and 95c - while mine will hit 100 almost right away and throttle. So something isnt right or people arent being honest about their results or what settings they are using

Its the speed of hitting the limit that gets me, there is no ramp up, just BANG almost immediatly at stock or within a few seconds with multicore enhancement turned off. Only the undervolt allows a run to be successful without throttling at 100c, but if I were to try a 10 min run it just ends up throttled at 100c again
 
Its the speed of hitting the limit that gets me, there is no ramp up, just BANG almost immediatly at stock or within a few seconds with multicore enhancement turned off. Only the undervolt allows a run to be successful without throttling at 100c, but if I were to try a 10 min run it just ends up throttled at 100c again

Its because they generate a lot of heat internally and all those transistors lighting up yes its immediate. The AIO coolers just can't shift the heat away from the chip fast enough you might get away with a big air cooler but in reality water cooling is the way to go with these as the flow is simply larger. I swapped my arctic 280 for a 360 and all that does is provide a buffer against the fans ramping up and down quickly it doesn't shift the heat any quicker because as said the flow simply isn't there it still heats up really fast and easily gets up to 90c if I don't lower the voltage on the chip.

Without context those results don't mean much most published reviews are done on open benches rather than closed cases or it could be a cold room there are plenty of possible variables. Benchmarking is very much a synthetic test there aren't many real world scenarios that would load up all cores intensively like that for any length of time the only thing I have is encoding several hours worth of audio that could load up all 12 cores multithreaded with all the tracks but even then it only lasts a few minutes and if it throttles it simply means it takes a little longer its not critical, during gaming which is a more normal scenario its typically 50-60c.
 
Last edited:
Yeah its an Asus board

I actually dropped the undervolt to -0.120 last night as I was getting intermittant crashing in a game that I suspected was possibly this, so dropped it a bit to see.

Im using adaptive negative offset

The SP score is 85 and undervolt protection is disabled
I use a fixed offset. Drop your pl2 to 253w and retest. Either than that I am out of ideas at this point.
 
Offset seems high to me but haven't tried a 13700k. i use LLC4 -0.050 Ac_ll 0.17 IA_LL 0.2 TVB +2 profile on the 13900k. SA set to Auto. AF2 360 and cb23 get to around 92
 
Last edited:
Have you got the LGA 1700 mount for the cooler and removed the plastic? Connected it to the cpu header?
Also the correct screws for the 1700 mount as they are different.
 
Last edited:
Ok well I can compare with a custom watercooling setup, I have an EK 360mm XE with 3 fans, an EK D5 pump / res combo and EK quantum velocity full nikkel. Plus the thermal grizzly contact frame.
The RAM is 6600mhz C32.

I bought it from random matt and i never reset any BIOS settings or changed anything except for fan profiles.

After running the full cinebench R23 multicore test the Max temp recorded is 70 c


20230130-165854.jpg

20230130_170024.jpg
 
Yeah the cooler has the correct 1700 backplate and standoffs etc - plastic wasnt on - the cooler was used before on my 5900x

After faffing and faffing im leaning toward the cooler and case arent helping

Basically Im not convinced the pump/coldplate is mating properly or the pump runs consitently, as I can reseat it and sometimes get worse and sometimes get better results. Also I have opened up the top and front door of the case and can get a good run on cin23 with no more than 95c and around 31k score. I can get about 3 mins into the 10 min run before it pegs at 100 and cant cool down again. Case is Define R6 btw

So I think its time to buy a bigger cooler. I think the only one bigger I can maybe fit in the case will be a 360, to get a 420 I would need to buy a new case, one of the XLs
 
Just looked at that case and it seem to have a solid front door and breaths through its sides. I can't imagine the airflow is great there in comparison to some others. The trick with 12th gen and 13th is to tune the volts and these run perfectly fine.

If I can get the 13900k running with a reasonable temp I'm sure the 13700k should be cooler. There isn't a huge difference between the 280 and 360 according to reviews. I think the 360 has about 7% more surface area or something.

Have you tried to reset the bios and start over?
 
Just looked at that case and it seem to have a solid front door and breaths through its sides. I can't imagine the airflow is great there in comparison to some others. The trick with 12th gen and 13th is to tune the volts and these run perfectly fine.

If I can get the 13900k running with a reasonable temp I'm sure the 13700k should be cooler. There isn't a huge difference between the 280 and 360 according to reviews. I think the 360 has about 7% more surface area or something.

Have you tried to reset the bios and start over?

Yeah the BIOS was reset last night as it happens - best I get is the -.120 offset and then open the case front up to let air into the radiator easier (front mount) - that gives 95c max runs generally with 31k scores

I think its the cooler in all honesty - the results are too all over the place with it after each reseat etc and even between runs it can fluctuate

If im buying a new one I may as well go bigger, hence the 360 over 280. i should be able to top mount the 360 as its thinner and can be offset to miss the RAM which should let it breath easier

But tbh I think I will just go all the way and get a bigger case and 420, will see.

What volts are you pulling during a run on the Vcore with those settings you use? @snarloas1982
 
Not sure off the top of my head. Check tomorrow. I'm still working my way down the ac_ll a step at a time to get the voltages lower. I know if I'm running stock with the settings above and mce disabled the max Watts is only about 225. Currentley I have mce enabled again with tvb +2 enabled. Cb23 will then pull to around 270w. I'll check the vcore under load tomorrow.

I'm happy with how I have mine. 6gzh light load. 5.8 medium and 5.5 heavy loads and scoring 40k.
 
Back
Top Bottom