my new 7700k dosnt run that cool even on a lower vcore

[trog100]

maybe not a fair comparison but at stock which seems to be 4500 mhz on 8 threads with the vcore showing at 1.184 volts its at 68 C whilst running the cpu burner from furmark..

my old de-lidded 4790K at 4.6 mhz with the vcore at 1.28 volts doing the same thing would be at around 58 C..

this chip at stock shows a tad faster in benchmarks than my 4790K at 4.6 gig.. but in real life use the difference isnt even vaguely noticeable and that is with 3200 memory..

the only thing i have noticed is my cpu runs at least 10 C hotter.. he he

i still havnt found out for sure what part of my old system died on me yet.. the motherboard or cpu.. but i have a cheap celeron chip coming to find out.. i will rig it up on a table in the old motherboard and find out..

trog

 

[Liviu Cojocaru]

Hi, did you apply the same thermal paste if yes, same qty?

 

[FYFI13]

Hi, did you apply the same thermal paste if yes, same qty?

It doesn't really matter, i7 7700K is known for running much hotter than i7 4790K. De-liding does the trick

 

[trog100]

It doesn't really matter, i7 7700K is known for running much hotter than i7 4790K. De-liding does the trick

i know it dosnt matter.. i am just noting that the newer smaller chip on less voltage still seem to run as hot as the old devils canyon chip which was known for running hot..

not something i would have expected to be honest.. i doubt i will push this chip.. just let it run well within it limits and not bother to de-lid it.. apart from playing the benching game there isnt much to gain from an extra couple of hundred mhz..

trog

ps.. interestingly i can still browse and do my normal stuff with the cpu burner running on 8 threads.. last night i forgot the burner was running and watched a netfix movie and never noticed the burner was still running.. its running now as a general stability test.. i cant really tell.. the system still seems sprightly enough.. he he..

 

[Komshija]

This is perfectly normal during synthetic benchmarks. I wouldn't worry at all. As long as you are under 85°C in short term (eg. during 1 hour Prime95 test), you are OK. If you can keep it under 70°C in the long term, that's excellent.

 

[trog100]

This is perfectly normal during synthetic benchmarks. I wouldn't worry at all. As long as you are under 85°C in short term (eg. during 1 hour Prime95 test), you are OK. If you can keep it under 70°C in the long term, that's excellent.

i would lay odds that prime95 will take it up close to the 100 C mark during certain of its tests.. pretty much like it did with my 4790K.. but prime95 is not a good temp tester..

in fact i will give it a run and see..

trog

ps.. i just gave prime95 a fifteen minute run.. i was wrong about it going up to 100 C it only made 86 C.. that is at stock settings with very good air cooling and both case sides off.. i havnt bothered to put them back yet..

i am sure a small overclock would have prime95 hitting the 100 C mark.. as i said prime95 is a poor tool for temp testing.. my 4790K chip didnt like it and nether does the 7700K.. it cooks the later generation intels cpus.. it does things that nothing else does and generates temps that nothing else does..

why people recommend using it i havnt a clue but many still do..

 

[Folterknecht]

Delidding will solve that problem - done right, not the Linus way.

 

[Komshija]

ps.. i just gave prime95 a fifteen minute run.. i was wrong about it going up to 100 C it only made 86 C.. that is at stock settings with very good air cooling and both case sides off.. i havnt bothered to put them back yet..

What's your room temperature, did you correctly apply thermal paste and did you fasten all screws properly? What are temp differences between your cores? There's a small possibility that once core has flawed temperature sensor.

My temps with my i7 6700K OC'ed to 4,6 GHz @ 1,344V under BQ! Shadow Rock Slim reached 82°C with Prime95 28.9. If I remember correctly, that was 80°C on core#0, 82°C on Core#1, 80°C on core#2 and 72 on core#3 and that was in September last year with room temp approx. 23°C. These temps were the highest ever on my chip.
When I play Far Cry Primal on Ultra settings, my CPU hits 63-66°C on core#1 which is always the hottest and 54-55°C on coolest core#3; room temp is 24°C.

 

[FireFox]

maybe not a fair comparison but at stock which seems to be 4500 mhz on 8 threads with the vcore showing at 1.184 volts its at 68 C whilst running the cpu burner from furmark..

Those temps are too high at stock and 1.184V.

It doesn't really matter, i7 7700K is known for running much hotter than i7 4790K. De-liding does the trick

Exactly.

 

[trog100]

What's your room temperature, did you correctly apply thermal paste and did you fasten all screws properly? What are temp differences between your cores? There's a small possibility that once core has flawed temperature sensor.

My temps with my i7 6700K OC'ed to 4,6 GHz @ 1,344V under BQ! Shadow Rock Slim reached 82°C with Prime95 28.9. If I remember correctly, that was 80°C on core#0, 82°C on Core#1, 80°C on core#2 and 72 on core#3 and that was in September last year with room temp approx. 23°C. These temps were the highest ever on my chip.
When I play Far Cry Primal on Ultra settings, my CPU hits 63-66°C on core#1 which is always the hottest and 54-55°C on coolest core#3; room temp is 24°C.

my room temp is about 20 C.. i think i know how to fit a cooler properly and apply thermal paste.. you are just gonna have to take my word for that.. running anything other than prime95 realtemp shows a max of 68 C..

all four cores show roughly the same.. two at 66C and two at 68 C.. pretty good that.. mostly they are further apart.. prime95 taking the temps higher is as expected..

but on such a low vcore 1.184 with a good cooler i would have expected to see lower temps.. which is kind of why i started the thread.. i did expect the new chip to run cooler.. it isnt..

it is worth noting my cooler fans are not running flat out at 68 C.. at 86 C they.. 86 C is a place i dont want to go which why i have an expensive cooler with two radiators and two large slow running fans..

the cpu also idles higher.. my 4790K idled about 3 or 4 C above ambient.. this one idles a good 10 C over ambient.. something else i was not expecting..

none of this matters that much.. its all just something i have noticed and thought worth a mention..

just for interest my 4790K needed 1.128 to run a stable 4.6 gig.. which dosnt make your 6700k at 4.6 gig on 1.34 needing vcore very good does it.. it also gamed about 55 C.. i would expect this 7700K even at stock speeds to be 10 C hotter.. as i said earlier not exactly what i expected.. but it wont do any harm..

trog

 

[eidairaman1]

I hope your cooler isnt on too tight to cause board flex

 

[trog100]

i would like a few thoughts on this.. a while back i bought a windows 10 retail home.. i have just looked at it.. as soon as you open the packet you see a card with a prominent key symbol on it.. on the card it says product key.. clearly printed and of a size you can easily read.. no problems and as it should be.. the whole package is nicely presented.. even a dumbo would not have problems with it..

this latest win 10 pro thing came on a dvd.. nowhere on the package does it tell you how to find the product key.. nowhere does it tell you that the product key is hidden under a little gray sticker about the size of a little finger nail.. it says microsoft in the little sticker nothing else.. it dosnt look big enough to hide a product key..

nowhere does it tell you to (very carefully) scratch the paint off the little grey sticker and under it you will find a tiny printed product key..

i recon i am gonna send it back claim its not fit for purpose and demand a refund.. i may not get one but if i dont scan will never see another penny of my money ever again.. so far this week alone i have spent very close to thousand quid with them.. i will gain a tiny bit of pleasure knowing that their loss will be greater than mine..

as for MS.. i can hate them as much as i like but i am stuck with them.. their is no viable alternative.. they have us all by the short and curlies.. he he

their asian based so called support system with phone lines that sound like an echo chamber gone wrong and operatives that have difficulty understanding what you are trying to say to them are not only useless.. they are an insult to their customers..

rant over but i had to have one.. he he

trog

ps.. its worth noting if you buy something from the micro soft store they claim to have a money back policy.. if you buy from anybody else as i did.. they politely tell you to go away as they did not sell you the product.. everything points back to scan.. but it seems scan dont want to know..

 

[eidairaman1]

Take a picture of it including all the contents of the package check your email to for a key

 

[FireFox]

nowhere does it tell you to (very carefully) scratch the paint off the little grey sticker and under it you will find a tiny printed product key..

Yeap, i guess they do it purposely so if you damage it you wont get your money back

Note: I don't know why you didn't buy a Key on Ebay as many of us did

 

[Komshija]

@trog100 It's always a possibility that Intel poorly applied TIM. That would be Intel's trademark. It also sounds very strange that Prime95 runs 20°C hotter than other tests...

OT: Seems that you have a golden i7 4790K chip, while my i7 6700K is average.

 

[Mussels]

not something i would have expected to be honest.. i doubt i will push this chip.. just let it run well within it limits and not bother to de-lid it.. apart from playing the benching game there isnt much to gain from an extra couple of hundred mhz..

if you have 20% less heat production over a 20% smaller area, you get the same temps.

It's been the same every generation since ivy bridge once they stopped soldering the IHS - only delidding can fix it.

 

[Arctucas]

Probably the IHS sealant.

De-lid and replace the OEM TIM with a liquid metal TIM.

 

[trog100]

Probably the IHS sealant.

De-lid and replace the OEM TIM with a liquid metal TIM.

my 4970K system died on me at the beginning of the week.. my "enforced" upgrade including the having to buy another windows 10 copy (two to be honest) has cost me close to £1000.. in view of this i am bit reluctant to do any major meddling for a while.. he he

i have that feeling that luck aint running with me at the moment..

the devils canyon was close to two years old.. i just bought a cheap £37 celeron to pop in the motherboard just to find out for sure which part of the old system has given up the ghost.. motherboard or cpu..

i think i have just gotten used to the nice cool running of my de-lidded 4970 chip.. i did too good a job on it..

my asus rog board has four preset profiles.. one gives it a small overclock which takes it to 4.6 gig.. it ups the vcore to about 1.24.. doing this adds about 8 C on the temps.. quite a lot for the small extra speed gain..

as for prime95 running hotter than anything else.. all it depends what its doing.. the blend test runs four different sets of instructions.. the first pass it makes runs at about 68 C.. the second one running small blocks runs about 83 C then the temp drops down to 68 C for the pass after that and so on..

its a bad tool for judging real world running temp tests because nothing else make the chip do what prime95 makes it do.. it uses more power and makes the chip run hotter than anything else does..

one thing is for sure.. intel aint making much progress as regards heat generation with these chips.. i recon they have hit the same brick wall they hit back in the old single core P4 days.. heat stops them clocking the things any faster.. time for complete new chip design i recon..

trog

 

[Arctucas]

<SNIP>
as for prime95 running hotter than anything else.. <SNIP>

trog

On my rig, LinX 0.7.0 with the 2017.1.014 binaries at max memory runs hotter than P95, but maybe that is just my rig.

 

[trog100]

i think i have discovered something about the cpu temps as recorded by realtemp..

this may be an asus thing i dont know.. when you enable the xmp memory setting in the bios in my case 3200 it asks if you want the all core enhancement option and says good cooling is required for this setting.. it gives you the yes or no option..

anyways exactly what this xmp all core option does i havnt a clue but when selected it adds 5 or 6 C to the cpu temps.. quite a lot to be honest..

my cpu fan is temp (or load) controlled and at idle its barely spinning but at idle the temps go up from 25 C to 32 C.. my ambient room temps is around 20 C..

under an 8 thread cpu burner load at 4500 mhz they go up from 65 C to 70 C..

so memory settings significantly effect cpu core temps..

does anybody have clue what this memory connected all core enhancement does.. ??

anyways i have found one reason why this 7700K chip is running a little hotter than i expect it to..

trog

 

[Mussels]

probably raises IMC volts (or similar) to help with overclocking the ram past spec.

 

[trog100]

i will give it a try without the all core enhancement selected and see what happens..

my vcore is now flicking between 1.168 and 1.184 at 4500 mhz.. this at stocking settings.. quite low by my standards.. it also seems to boost to 4500 on all cores..

trog

ps.. done that little test.. i think i know what it does.. it affects the stock intel boost speed.. intel control only boosts to 4.400 mhz with memory at 3200.. the all core enhancement option keeps it at 4500 mhz.. it dosnt seem to affect the cpu temps much at all on my system..

running the memory at xmp 3200 does.. it adds around 5 C to the cpu temps.. both at idle and full load..

 

[FireFox]

xmp

I don't use XMP profile.

 

[Aenra]

Delidding will solve that problem - done right, not the Linus way.

Please expand?

And on a more general note, isn't 68 degrees for "stock" speeds rather excessive? If this is what passes for normal, kinda tells me they pushed that chip way too high.. no?

 

[Folterknecht]

Please expand?

https://www.techpowerup.com/forums/...-3570k-lowering-my-temps.221017/#post-3433400

That is how it's done right.

THIS is the "Linus way" of doing things, fucking up and presenting it as reasearch/conclusion or whatever he wants to call it.

And on a more general note, isn't 68 degrees for "stock" speeds rather excessive? If this is what passes for normal, kinda tells me they pushed that chip way too high.. no?

68°C at stock speed is a little on the high side with a good aftermarket cooling solution and highlights the problem with Intels thermal paste instead of using the good old fashioned way (soldering the IHS to the DIE).