Temps for a 4790k with current gen watercooling ? Need numbers to compare plz

Hey guys.. I just got my system below setup finally after lots of issues.. And im using watercooling like I have been for the last like 15 years now.. I never use fan coolers.


And you guys are probably going to smack me in the face when I tell you the water block and such im using.. I first got into watercooling in the Athlon Xp days back in like 2001.. And I have been using the same watercooling setup for every cpu jump I make.. All I have been doing is modding the brackets for each gen cpu I get.. I have a Maze 3 block and got I cant even remember the heatercore I have but the pump is an Eheim 1048 I believe the heatercore is a single 120mm fan size core.. Im using 2x 120mm fans on it.

Anyway, I have used this block on Athlon xp 1800-2500, Athon 64 3000, Opteron 64 165 duel core, and recently for 7 years my Core 2 Duo e-8400 wolfdale.. Mind you I was able to overclock that beast to 4400MHZ with that block which was very high for an E-8400, and I kept that clock speed for 7 years, to the point where I don't even remember my bios settings so I had to take pics in case I had to reset it..

Anyway, long story, but the mounting system on the C2D and the 1150 chipset is the same, only slightly bigger as far as the holes go on the board so I didn't have to change any of the mounting.. Im currently running Prime95 now for the last hour or two, and the cpu is clocking itself to 4400MHZ at 1.24v, and all 4 cores are around 65-73c.. They did get to 85C but all I had to do was tighten the block down to the max and the temps dropped to what they are at now.. I didn't have it tightend down all the way because I didn't want to bend the cpu socket which can happen if you mount something too hard. Im still not sure if its enough pressure though.. The block to the touch is warm but not hot. The water is warmish but not hot either, yet the CPU is 67-73c.. Probably normal though. As for my C2D got about those temps too.

What do you guys think ? Yes no ? What are your temps with watercooling using current blocks and so forth.

Oh and according to core temp.. Core #3 always seems to be like 5c cooler then the other cores.. If one is at 73c core 3 is 68. Is this normal ?

it is very common for cores to have 5c differences in temps.
I would say that you arent doing bad on temps if 73 is a high after a couple of hours of prime

Seems fine to me.

It's hard to compare systems though because you have a heater core and are not sure of its size versus something more traditional. We also need to know your ambient temps. What did you use for temps?

That said, temps are fine for the processor regardless. Pretty good for what was me tioned to be a single 120mm rad, actually.

Whatever heatercore's they sold back then with the maze 3 blocks is what im using lol.. I think I'll invest in a newer block down the road though.. 15 years out of mine was a good haul.. lol The reason they call them MAZE blockes is because they dont have a zillion fins fins or pins like newer blocks these days have.. instead it was a massive block of copper with a loopy maze going through it.


Here is a picture of the block.. Its an old picture from 2008 this is when I modded the block to fit on my C2D motherboard.. I had to cut away at the bottom because the caps were in the way and add new brackets and so forth.. But I didn't have to do anything to the 1150 socket just plop it right on.. But yea this is the block im using.

And this link is the radiator im using.. sorry not a heatercore its a black ice radiator

http://www.modthebox.com/review221_2.shtml

Wow classic. I had several of those blocks back during the Socket A (AMD) days. Very creative with modding your own mounting

It works right lol ? Like I said I used this block on my AMD Athlon Barton 2500+, A64 3000+ Newcastle, Opteron 165 duel core, C2D E-8400, and now my I7 4790K LOL.

I got my moneys worth.. I was trying to remember who made this block, it was dangerden.. I didn't realize apon looking them up they went out of business.. Oh well.. Their blocks are still rocking my system lol

Here are the updated pics.. As you can see its mounted on my I7 lol, and in the back is my old black ice radiator and the eheim 1048 pump.. You know that pump has been running nonstop for nearly 15 years now lol. I love Eheim products I have one of their canister filters on my 92g fish tank.

Not too shabby! About your first post, temps are right where they're supposed to be considering you're using a single 120mm rad. Also worth noting is the temps on each core may differ from 2-5c. On mine core #2 is up by 3c. I'll bet your temps will improve if you had a 240 or a 360 rad for sure. That DD Maze3, tho quite dated, is still doing a pretty good job.

I had an Ehiem 1048 too, but it died long ago.

Yea plus my bedrooms temp is a little high.. I have 3 fish tanks in the room and its a smaller bedroom upstairs and my fish tanks I keep at like 84-86f one of them is a 92g corner tank without a top so it gets a tad humid in the room too if I have the door closed, so my bedroom is always like 75f or higher

I think Im going to probably get a new block when I get settled with money the 1200$ I spent on this system is killing me.. Those temps were on Blend in Prime 95 when I went to the FPU test my cpu temps went to like 85c from 67-73c on blend.. Playing video games though my temps have only gotten to 53-54c so far. Or is my block actually cooling it well..

I know the small FFT test gives the cpu a unrealistic load im just not sure how hot other people's 4790k get using small FFT's... How hot does yours get ? Or is small FFT something we shouldn't be using anymore.. Im fine with the temps I get using blend.. But small FFT'S nearing 90c I just turned the test off I didn't like those temps

Yeh...running Small FFT test with P95 v28.5 will invoke the AVX instruction set within the cpu. When doing this the cpu will get hot instantly especially if you are overclocked and running higher Vcore than normal.

With my 4790K @ 4.8GHz (1.34 Vcore) my cpu would jump to 90C+ as soon as starting Small FFT test. This is with the cpu watercooled with a high capacity water cooling setup.

I ended up delidding my 4790K and this resulted in a dramatic drop of over 20C, by far the best thing I did for cooling the cpu.

I can now run Small FFT test in Prime 95 all day long and not break out in a sweat....

if the block is still doing it's job well, why change it?
It looks like you only need a little more rad.

Well the water temp was only barely luke warm... So I dunno if a new radiator and faster pump would help but probly.. I could always try that though.. The Eheim 1048 is a really low GPH pump.. at the time was the bottom of the list that dangerden was selling..

From what im reading these temps are pretty good/normal for the most part.. I guess im just used to my cpu's not jumping up so high so fast, and they do seem lower then aircooling at least.. I turned the vcore down a tad.. It was 1.24-1.25v at 4400mhz, I put it down to 1.17v now and the temps using Blend are now in the 60-65c range, and small fft's test went down about 10c as well.

Man this auto overclocking almost takes the fun out of overclocking your cpu from what I gather these 4790's don't get much higher then 4400mhz lol maybe 4600 or so and that's about it.. Ah well Fkat 6 what are your overclocks and voltage now to where you can run small FFT's 24/7 im curious because I want to compare current gen watercooling to what I have now.. It actually seems like 13-15 year old setup is about on par with current stuff for the most part anyway.

I know with my C2D everyone laughed at me for using a maze 3 block said it couldn't be done, but I was able to overclock that thing from 3ghz to 4400mhz and that was a really high overclock

Cnaydmancangt said:

Flat 6 what are your overclocks and voltage now to where you can run small FFT's 24/7 im curious because I want to compare current gen watercooling to what I have now..

Click to expand...

I run my 4790K at 4.8GHz with the cache set to 4.6GHz. I utilise 'adaptive' voltage control after I carefully worked out my final max. core and cache voltages carefully using manual voltage settings.

Under load my core voltage hovers around 1.34v and at idle it sits at 1.28v. Note that I had enabled C-States. The idle pic below lists the core voltages at 0v - This is normal behaviour when C-states are enabled.

At IDLE [with C-States Enabled]:




Running P95 v28.5 Small FFT test.

It would be helpful to know your ambient temperature, and what cooler you are using....

(if its not the same, then it isn't applicable information for the OP I hate to say)

It gives me a general idea, I can garutnee my ambient temps are much higher then his, living upstairs on the 3rd level in a house in a tiny bedroom in Virginia sucks.. That an I have as I said 3 fish tanks all at 82-86f my 92g tank doesn't have a top so it gets humid in my room with the door closed. My ambient temps are always around 80f in the room. I just used my infarred probe gun on the wall of my bedroom and its reading 79.5f I know that doesn't read actual air temps but it gives you an idea lol.

And this is with my bedroom door open, when its closed the room gets even hotter. But even if my room were 15f hotter then his that doesn't really mean much in the celcius temps

Right, you said that earlier .

There are sooooooooooooooo many variables that what someone else's system does is rarely applicable to your own. Even though you are looking for a 'general idea' without knowing those details (and more) it may not even do that.

EDIT: As far as the new block, it would likely improve things a couple of degrees C, but I wouldn't expect much more honestly.

I have an idea where my cooling stands, its not the best but it sure does seem better then the heatsink temps I have seen.. I don't blame the radiator and pump that much.. the water is luke warm at best that doesn't equate to much in cpu temps when your so high in the C range. If the cpu temps sky rocket to the 90's within a split second with a higher voltage.. Then its probably just how the cpu acts.. I did some research and found a long thread on the intel forums about people complaining the 4790s sky rocketing like this.. So intel ran tests of their own and they came to the conclusion that it was because the voltage was too high sometimes reaching 1.34v on some board using default settings.. I know mine over volts my memory to 1.54v, and my cpu also jumps to 1.25v or so. Once I lowered it to 1.17v the wattage usage went way down and the temps using small FFT's instead of 90c were like 75c. Also intel doesn't use the prime 95 benchmark they use some other benchmark they created and it acts more like prime 95's blend test

Now his system runs 4800mhz 1.34v and gets 75c or so. So my only conclusion to this is that my water block isn't as able to fully transfer the heat from the cpu to the block itself.. Probably because the block is so thick and doesn't have the smaller little chambers that all these new blocks have.. instead its mostly a flat surface with a snake tube ya know ? And so it doesn't transfer heat as well. I can only imagine my temps if I had a vcore at 1.34v and overclock at 4800mhz lol id probably reach over 100c

But in all honesty small FFT'S is unrealistic, I've never had a cpu in my 15 years building computer be put to so much extreme in gaming or whatever.. I mean I just played Planetside 2 which is heavily CPU dependant and my cpu temps were 54c the cores were even split at like 30% usage max..

You've mentioned (all) that before too, yup.

As far as small FFT's being unrealistic, it is for MOST things. But that is the point! We stress test to QUICKLY find any instability in our CPUs and to see absolute worst case temps. If it passes the stress tests, it will likely pass anything else you do.

EarthDog said:

There are sooooooooooooooo many variables that what someone else's system does is rarely applicable to your own. Even though you are looking for a 'general idea' without knowing those details (and more) it may not even do that.

Click to expand...

Agree, too many variables.

Cnaydmancangt said:

So my only conclusion to this is that my water block isn't as able to fully transfer the heat from the cpu to the block itself.. Probably because the block is so thick and doesn't have the smaller little chambers that all these new blocks have..

Click to expand...

If you're tubing feels a bit warm and you're getting those high temps while benching, my guess is that there's either isn't good contact, which I believe you already do without bending CPU pins and the fact that the block is an older design and as times have past, newer and better designs have since been released. Its up to you if you're willing to go into your pockets to fork over for a new block and what not. I'm sure you'll get better temps but can't say by how much as the rest of the loop will need to be updated as well to see a decent change or what delta-temps are being aimed for.

EarthDog said:

You've mentioned (all) that before too, yup.

As far as small FFT's being unrealistic, it is for MOST things. But that is the point! We stress test to QUICKLY find any instability in our CPUs and to see absolute worst case temps. If it passes the stress tests, it will likely pass anything else you do.

Click to expand...

Again, Ed is right. In my case since I fold, that is the closest real world environment that stresses the heck out of my OC'd CPU and GPU. When I'm not folding, I am doing what everyone else does, browse and game.

I will say though, that is a pretty old decent setup you got there OP for the temps you get from a 120.1 rad.

120.1 whats that ? Actually I think its a 140mm radiator I just use 2 120mm fans on it with tape to seal it up.

As for good contact I checked and double checked before I hooked the rest of my system up and the block was getting full contact over the heat spreader all of the AS5 was fully smushed.

I think as I said its just the block is too thick and doesn't have the fins that say your block has you know the tiny chambers that you current watercooling guys talk about getting clogged lol. Im content with the thing though.. At least it cools better then a heatsink still

I just played The witcher 3 for like 4 hours today and my GPU got to 73c so I made a fan profile for that game in the CCC and set it to run at 60% and it now runs at 65c, and the cpu never got above 54c, just like on Planetside 2 and World of tanks, which are the other games ive played so far. Speaking of which man... this PC I built is so badass compared to my old one... Geesh that 4gb of ram and the e8400 C2D and ATI 4890 were sooo old.. lol

Now that I have 16GB of ram I can see the real memory usage and im sitting at 6gb on some of these games. Im glad I opted for 16gb of ram and not 8GB like some ppl suggested.. I can see in a couple years video games using 8GB of memory and ill be going back to pagefile... Im sure you can imagine the pagefile and how slow my PC was with that 4GB of ram haha.. Witcher 3 runs maxed details, except that NVidia hair setting and I get 60fps.. The hair thing really crushes my PC

Cnaydmancangt said:

I believe the heatercore is a single 120mm fan size core.. Im using 2x 120mm fans on it.

Click to expand...

You had said that it was a 120mm rad so we assumed that's what you were using.

Also, 120.1 is terminology we use for radiator sizes. 120mm being the size of the radiator and the 1 being a single rad. If it was a dual radiator we would replace the 1 with a 2, so on and so forth.

Overall, your temps look good for what you do.

As for the pagefile, my IT buddy told me to turn mine off. Never understood why but I did on my main rig with Windows 8 64-bit Pro. I did some research around and read that its not good having a pagefile not only when you have so much RAM but having a SSD as well. I use a premium priced M.2 PCI-E OS drive and the last thing I need is that drive going bad, especially for the ridiculous price I payed for it. I've had the pagefile disabled now for over a week and haven't seen any negative impact. As a daily driver, I use about 6-8 GB of RAM out of 16 GBs myself in a triple monitor setup.