Core i7 2600K Golden Sample vs. Core i7 4790K - lamentation

Toggle sidebar Toggle sidebar
uplink-svk

uplink-svk

Distinguished
Mar 9, 2013
2,300
5
20,965
Greetings,

This thread is more or less a lament about how new cpus are faulty. In general I started to run Intel processors when Core technology arrived [remember C2D xxxxx, E6600, Q6600 etc? :)] and I must admit, that i7 920 together with 2600K were Intels state of art products.

By building tons of 3xxx and 4xxx desktops, it's kind of frustrating for me to say that:

Core i7 2600K@ stock clock in "same" case, with NH-D14 with U.L.N.A. and no pwm, is running at default settings at Prime by 50-52°C with ambient of 24°C and 55% relative humidity. This beast is GS [golden sample] able to withstand 5 GHz, tested it, worked, withstood prime, crashed at some other tests. 2600K is eating 95W at max, and is running at about 75-80W at max with Prime at it's factory settings and 1.201 vcore.

How is it possible, that every other "mid-high" end processor I tested ran hotter, with same, or better cooling system?

Core i7 4790K at 1,1 vcore hit 65-70°C running all 8 threads at 4 GHz at "almost" factory manual settings. It's in the same case, using NH-D15 both fans, they were running really loud.

Why is this happening? Why are all processors from 2600K worse in heat dissipation and heat loss then their precedessor, eventhough they're running in dramatically cooler environment and lower TDP?

Were all those tens and maybe hundreds of mid-high end cpus faulty? It's like I built over 20 2600K systems, all are running "superb". All of the 3xxx and 4xxx are running "hot", eventhough they have lower TDP and better cooling systems in general.

Please enlighten me, because it's getting really frustrating.

Thank You,

Bryan
 
Solution
G
yeah sandy bridge CPUs were soldered to the heat spreader. ivy bridge and haswell have an intel "in house" thermal paste applied between the IHS and CPU die. solder just works better at conducting heat away from anything vs thermal paste. for some reason intel thought that it would save them money in the long run using TIM instead of solder. greedy bastids.
Sort by date Sort by votes
G

Guest

Guest
yeah sandy bridge CPUs were soldered to the heat spreader. ivy bridge and haswell have an intel "in house" thermal paste applied between the IHS and CPU die. solder just works better at conducting heat away from anything vs thermal paste. for some reason intel thought that it would save them money in the long run using TIM instead of solder. greedy bastids.
 
Upvote 0 Downvote
Solution
spentshells

spentshells

Judicious
Apr 1, 2006
8,556
294
33,990
Performance wise, basicly nothing, oerclocking wise, the larger die of the 2600K wil take more voltage and give a better overclock depending on the chip. The new tmu for devils canyon doesn't make up for this. If you were gonna buy now, well clearly go with the newer tech as the 1155 boards are hard to find, everyone wants an arm and a leg for a used one thats half way decent.
 
Upvote 0 Downvote
rolli59

rolli59

Titan
Oct 27, 2007
47,790
283
130,890


^+1
 
Upvote 0 Downvote
G

Guest

Guest


die size has nothing to do with it. theyre basically identical in dimensions. its the material applied between the metal cap (integrated heat spreader) and actual CPU die that cause temperatures to differ.
 
Upvote 0 Downvote
uplink-svk

uplink-svk

Distinguished
Mar 9, 2013
2,300
5
20,965
Dirk_Kuyt - well that sux :/. Does Intel plan on returning to bigger die and solder it together with iHS in some next cpus on their roadmap? Heard that some people divided iHS and chipitself and put there their own TiM, but this one sounds a little hardcore for me. Thank You for Your reply and for excellent explanation.

spentshells - I upgraded to 3770K, when I saw the temperatures I returned the piece. Now it's 4790K and I need a new computer with it [mobo in general] and it's doing the same. Well, I guess I'll have to deal with it in my head. I'm really a temperature freak. I love keeping my HDDs under 35°C at burn, and my CPU under 60°C at burn, even when it's summer [I do have AC, but it's hotter in general]. Thank You for Your kind reply.

rolli59 - yup :)
 
Upvote 0 Downvote
G

Guest

Guest
the TIM with devils canyon is a "next generation polymer". still TIM all the same. nothing beats soldering though. except maybe getting your heatsink direct-to-die with no IHS in between. never tried it myself.
 
Upvote 0 Downvote
spentshells

spentshells

Judicious
Apr 1, 2006
8,556
294
33,990


Yes it does and you are wrong in relation to the overclock. As for the temps, your being silly, if there is more physical contact between the heat spreader and the chip due to a larger surface area of said chip it does indeed dissipate hear more easily. A wider boat floats better if that helps.
 
Upvote 0 Downvote
G

Guest

Guest


1 - no, not as far as anyone can tell. they *tried* to fix this problem with the "devils canyon" CPUs but never truly soldered the CPU die back onto the heat spreader. they claim to use a "next generation polymer" for the TIM but from what i've read, it isnt necessarily close to same efficiency as actual solder. so, no, for the foreseeable future, we may not see a true soldered IHS

2 - yes, the "division" you're referring to is what some people call "delidding". i myself have done this with a i7-3770K (ivy bridge) and i7-4770K (haswell). i definitely saw a reduction in temps around 10C. but that was with overclocking as well. its a dangerous procedure and i would only suggest crazy people do it.
 
Upvote 0 Downvote
uplink-svk

uplink-svk

Distinguished
Mar 9, 2013
2,300
5
20,965
Dirk_Kuyt - cpu chip can't take much pressure, You need special pads in order to use unintegrated ihs cpu. 've seen many dead cpus of people, who tried to do it :/.

//no need to explain as I see :) You already tried...

spentshells - in case the die size is relevant to transitor count at µm², it is indeed "irrelevant", in case they're using the main casing as a cover, that is bigger then it must be, it is "rellevant". Not sure how deep under the main case is "cpu". Whether it's just a 0,1mm, or more. It's just physics talking out of me. I might be wrong.
 
Upvote 0 Downvote
G

Guest

Guest


nope. look at actual pictures of an i7-2600k die vs a 4770k die. they are physically quite similar to the eye, if not identical. they are almost the same exact size. only the number of transistors per die is what matters. if thats what you're referring to with "relation to overclock" then idk what to tell you
 
Upvote 0 Downvote
G

Guest

Guest


surface area is not relevant because they have the same basic surface area. the difference is the number of transistors within the CPU. im trying to teach you something here...not argue...haswell has 22nm transistors within the CPU die. sandy bridge has 32nm transistors within the same size cpu die. more transistors = more heat. the more trasistors you pack into the same amount of space is going to generate more heat. thats just the way it works. plus you add in thermal interface material vs a fluxless solder.
 
Last edited by a moderator:
Upvote 0 Downvote
G

Guest

Guest


yeah i dont really know wtf that even means. sounds like gibberish.
 
Upvote 0 Downvote
G

Guest

Guest


oh, and, there are several "decent" z77 boards on newegg for totally reasonable prices...
 
Upvote 0 Downvote
uplink-svk

uplink-svk

Distinguished
Mar 9, 2013
2,300
5
20,965
Dirk_Kuyt - physics. When You put same number of transistors that are made with 32 nm on same space, as 22 nm [same count!] You get less heat, right?

So when the die is bigger, You have lesser heat, with same number of transistors. This is only a theory, they always stick transistors as close as possible together.

leeb2013 - heard rumours, never had it explained by anyone in a logical way. Now my friend told me that 4790K is "fixed" tim-wise, well, well it's obviously not. And honestly, 3770k running with "delidding" [just watched YT vid.] mod is still running 10-15°C hotter than my 2600K, eventhough it has lesser TDP. So there are still some "issues" I have to catch up with.

P.S.: Majority of my systems are fully passive low-voltage, low-tdp solutions, so there are only a few "tens" of real life experience with "mid-high" range of processors [high-end entry]. Most of the clients want pc for MSO and porn ["movies"], therefore if they want desktop, and not brix, I stick up with S processors most of the time. Just explaining my lack of experience and knowledge with mid-range "high-end" processors and their architecture/engineering.
 
Upvote 0 Downvote
G

Guest

Guest


no just forget it

just read this: http://www.tomshardware.com/news/ivy-bridge-overclocking-high-temp,15512.html

intel calls it "thermal density"

its quite simple when you think about it. you pack a larger amount of people in the same sized room, its gonna get hotter....
 
Upvote 0 Downvote
G

Guest

Guest
hey no problem my man. glad to help you. i just wanted to help you understand that youre not crazy and theres probably nothing wrong with the builds you have done. its just how it goes these days. cpu developers state that supposedly its going to be a bitch getting down past 14nm cpus, and intel doesnt necessarily know where its going to go once their "broadwell" cpus are released next year. im sure they have a plan though.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom