So whats the top performing cpu block out for i7?
Printable View
So whats the top performing cpu block out for i7?
you guys have a test result for this? because i cant think of many i7 tests out there..
ive done a ek supreme and a koolance 350... so far i like the koolance a little bit better...
performance is a little off, but i like the mounting system (super easy) and results i get from it are reproducable...
I'm running a gtz and have had my 920 up to 4.2ghz (turbo) with 1.55v and never broke 65C. Although my apt is pretty chilly right now.
I've been testing the GTZ extensively for the past week and with a MCR220, my 920 at clocks in sig using 1.36V crests at around 73C. With the Cooler Master V8 these clocks gave me 97C load both using prime blend (blend hammers the i7 much more that small or large fft's ;)
http://i125.photobucket.com/albums/p...FinalTemps.jpg
well until someone can get reproduce dead even temps like this... (its easy on the 350 i just helped someone yesterday and found a borked sensor on his cpu)
im not ready to pass any block greater then koolances 350... 1-2C means nothing without balance..
How can anyone reproduce results with so many unknown variables? What voltages? What's the ambient? What TIM is being used? I can go on and on lol. Also, how about an updated version of Realtemp.
Balance my arse...more like luck. Looking at the minimum temps pretty much confirms that. I've used stock, a V8 and this GTZ and the delta between Core 0 and 3 have been identical across all three coolers.
oh man... gotta be nice...
luck eh?
Running a custom dual pump setup with a cpu only loop.
http://i125.photobucket.com/albums/p...a/IMG_0161.jpg
Also i helped a guy out yesterday and we reproduced even temps.
Luck? no its called hours of implimentation and testing...
Im not on the same level as the rest of this forum when it comes to watercooling parts..
When i do tests.. i dont end it with just mounts..
I also retest with different blocks also... i have had 3 supremes at one time..
http://i125.photobucket.com/albums/p...a/IMG_1147.jpg
So you have anything else contructive to say to me?
So luck eh?
reverand... because naekuh is in nice mode..
i did this for ya in the version you requested...
http://i125.photobucket.com/albums/p...rla/Temp-1.jpg
i cant show you guys anymore then this..
http://i125.photobucket.com/albums/p...rla/Update.jpg
Read Gainestown and not bloomfield.. i applogize, but id rather keep my contacts then violate a full blown NDA..
If you havent tried a 350 block, dont talk smack about it... its a dayam fine block...
EDIT: uhh wtf... im getting lower load temps?? :rofl:
*sigh* great... now my load tests will be invalid if i use this version.. :\
wow, you're in nice mode...should I feel special because you're not in another mode? lol. Use an i7 Bloomfield which are notorious for having a 4C delta between core 0 and 3 then get back too me. I never talked smack about the Koolance...but you seem suspiciously defensive of a product so I'll find someone who will give me proper information before I make ANY judgment. I think you have me pegged as a Swiftech fanboi...couldn't be further from the truth.
well.. if anyone responded like this who wouldnt take it as an attack?
I dont deal with luck.
Luck is bull:banana::banana::banana::banana: and not reproducable..
What i care about is whats reproducable. :up:
And yes i am being nice right now... trust me, if i wasnt, id be a lot harsher...
and if i was mean, i would of raped ya on that statement.. :P
how is lowest temps of any concern when looking at block temps?
Which valid tester do you see and trust that takes idle temps when doing his calculations?
Why is idle temps even needed besides to prove a point that you got your block mounted?
You seem to know a lot of testing methodology... So i'll explain mine.
I take ambient.. then the delta of watercoolant from the ambient.. that gives me a variable... (Rad efficiency)
Then you take the loaded coolant temp, and take that delta with the loaded cpu block (gives you another set of data)
Then you compare all the sets of data with the ambients taken at inlet of your rad, and get a third set of data...
And i do this with multipul blocks and mounts..
I was using the old realtemp because well, my other testing was done on near same hardware minus the board cpu and using old real temp.
When cross testing the biggest mistake a tester will do is CHANGE his methodology and screw up his comparision. (reason why anandtech took so long to adpot to the new methodology).
Do you see a flaw somewhere that i should be aware of?
LOL :owned: Poor rev
Gainestown is looking very nice. Nice testing methodology
Guys, it's one day before Christmas Eve. If you're not at work I suggest a good drink...like a double. We don't need a hardware war. The question was, which was best. The only way to answer that is through testing. Testing is good. This hardware is good. The OP is looking for the best out there right now. That's gonna take testing of multiple blocks. I would venture a guess that we don't currently know what the best is quite yet, but it won't be much longer before we do. :yepp:
Right now, I'm too buzzed to think about messing with computers. If you are of legal age, now is a good time to get that way too. :)
^
we need more testing to show whats the best...
i cant not tell you wants the best..
i can tell you what my favorate is right now...
im in the middle of testing the supreme.. i have 4 mounts down and 2 90 directions..
i need 2 more rotations on the block b4 i can conclude this...
So no i cant tell you whats the BEST... The best can change with the addition of 1 other component.. :up:
I do suspect like many that the blocks with the larger pin array or whatever surafce area enhancements they use will be a bit better, but that is only a guess.
Thanks for the testing Naekuh. :)
well it kinda sucks how i dont have mounts for the cuplex, or my d-tek.
ima trying to get bei to work on another project...
i was thinkn D-tek promount for the EK supreme LGA1366 style.. You would have to use the thermalright spring screws and backplate tho..
Skinnee and i are planning it out right now...
Don't shoot me for bringing up that word "air" but these things are monsters on air. The stock cooling is just not efficient for overclocking use. I spoke with Alex about this, and the water blocks in use here at XS are MUCH more efficient. It's a completely different kind of running. Heck for that matter water itself is much more efficient...well, quality water cooling setup correctly anyhow.
I'm debating between the DTek and GTZ. I'm definitely gonna get the rest of my stuff in January. In fact, that's why I called Alex. I promised him that I have not forgot about it, and that I would purchase it then.
These i7's are no joke. I really really want to try and keep them below 60 load, but I don't know whether that is even possible. It's gonna be hard to do.
Valuable post there elmore. :clap:
T Flight, you're right; we need more testing, but in all of Naeku's expertise, he's failed in actually giving anyone viable information. I can talk in circles all day and present loaded info like it's actually valuable to the public but I'll refrain.
Happy Holidays T :)
so the load temps of a "cherry chip" that isnt released yet? Im sure the IHS is dead balls even on that thing so people like you that get them post what amazing overclocks they get so people buy them.
Why does everyone in the i7 thread have their core 3 (4th temp) the coolest? Interesting
wow...
and this is exactly why i quit my first wave testing..
If you guys all feel this way, i'll keep my results to anandtech only then.
thanks for messing up another fun thing for me to give you guys...
have fun with i7 testing YOURSELVES then...
Good luck with the mounts... i have them... but you guys wont see results anymore from me on in this section.
Martin doesnt have i7... skinnee doesnt have one either..
Andy's i7 is not fully setup, and also he doesnt have all the mounts...
Evil only has mounts for the supreme..
So have fun waiting for one of my friends to do it instead.. Oh martin quit testing btw..
so this is how you deal with adversity? Nice
Listen man, if you have a cpu that can produce identical loads across all 4 cores, don't try to come off like it's the coolers doing. It's the cpu.
When you get an i7 and actually know what you're arguing against bring on the data; no one should ever discourage that.
1.36v after droop for 3680mhz is a bit much though, might want to back that down a bit.
I fail to see how you can eliminate the 350 if you don't have both the block and i7 to test against. Just because in your current setup you have uneven temperatures across cores doesn't mean you will have the same results with the koolance i7. Perhaps you'll be kind enough to scientifically explain why it's the i7 cpu over the 350 waterblock??
Lol maybe because he is the only 1 out of the entire i7 thread that has equal temps
have a little read :)
http://www.xtremesystems.org/forums/...postcount=1136
and how does that effect me when i am going to use my cpu as a standard which does have even temps?
am i using YOUR CPU to do the load test?? Wouldnt you think a unbugged cpu which does have even load temps be a better basis of testing?
So how does OTHER PEOPLE'S STUFF come into ANY PLAY when doing a side by side comparison?
Do you know how noobish you guys are sounding?
im done being nice in this thread...
You noobs really want to try to outwitt me? go for it..
Naekuh, dude, let's chill out and calm down. This place is about sharing our experiences and thoughts; please don't overreact. We would all love your input, but others have a right to share their views too.
Go easy Naekuh...Easy man. I really wanna see your results.
For those that don't, please keep that to yourselves. If anybody has personal issues with a user, please take that to PM's. We don't need to loose people that have access to this stuff and the ability to test it.
Let's not have any division in this forum. We've already seen the results of that. Let's all get along...right now. If you need to chill (and I'm not pointing any fingers) go get you a drink or something.
This is a thread that EVERYBODY that has an i7 has been asking about. Don't get it "cleansed", and cause a breakdown here. I stay off the report button, but I'm for real man, if this gets out of hand I'm gonna be forced to use it, and I really don't wanna do that.
I'm not getting in the middle of this or taking sides. Let's cool it.
Naekuh, will you please reconsider for the benefit of people like me? I value this kind of input, and it doesn't have to be a Gov't funded military experiment. Your testing is just fine. We can all make our own decisions based on that, but your info is still valuable here.
Remember, this is XS folks. Let's keep this topic going and make it worthy of where it's at. :yepp:
Someone needs to start a NEW THREAD on Intel i7 water cooling. Temps,equipment,fans used etc.. The reason for these threads to help each other. I wish Martins Liquid Lab was not closing (12-26-08) he could help us greatly in this matter. I am planning on using a Swiftech GTZ with a BlackIce 240GTX with a push/pull at 78CFM. I am wondering if this will put out allot of heat? Below are the loops I will be using to water cool. If I had to I can add another 120GTX to the mix if I need it. I was planning on doing the last one but decided to much.
Loop #1 BI 240GTX - Swiftech GTZ - 355/XSPC Acrylic Reservoir
Loop #2 BI 360GTS - eK4870 X2 - 355/XSPC Acrylic Reservoir
BI 240GTX - Swiftech GTZ - BI 120GTX - eK4870 X2 -355/XSPC Acrylic Top - BI 360GTS - 355/XSPC Acrylic Reservoir - back to BI 240GTX
I have the koolance 350 and GTZ, I prefer the koolance block,
NaeKuh thanks for the tip on getting the TIM spread easy thats works like a charm.. 4 mounts now and TIM been perfect each time
As would I, but without the assumption that everyone ELSE needs to shut up afterwards, T. Anand's cooling forum is dead due to behavior like this. I'd hate to see that happen here too.
What's best? :D I'm very pleased with the GTZ for mutiple reasons and won't be switching until the i7 specific blocks see the light of day.
Thanks for your help man!!!! NaeKuh Don't give up!
I changed from the GTZ to the Koolnace 350 block and dropped my temps by 8c
judge for your self, im selling the GTZ and staying with the 350, but with out the QD fittings.
Ambient is 22C for both blocks.
Aslo a friend has the 350 block with temps in the low 50's at full load with the same installation from NaeKuh.
Koolance 350 after NaeKuh help with the mount!
This is with 7/16 x 5/8 tubing with 5 Koolance Quick Disconnect fitting that restrict water flow alot, and i will be taking them out after X-mas and will most likely see a 2-3 temp drop
http://i28.photobucket.com/albums/c2...om/64temps.jpg
GTZ temps: These temps are with 1/2x3/4 tubing with bitspower compression fittings.
http://i28.photobucket.com/albums/c239/vkvenom/GTZ.jpg
Hi All:
I saw this post, and noticed that there are inconsistencies between the two tests that need some explanation. An 8C Delta is simply much too large, which raises questions on the test setup and methodology.
I did a little cut and paste in photoshop with your test results to highlight what I am talking about:
http://www.swiftech.com/assets/images/350vsGTZ.jpg
1/ core voltage is higher in the GTZ test
2/ north an south bridge, as well as GPU & memory temps are much lower in the koolance 350 test.
The core voltage will account for higher temps against the gtz, how much is unknown at this point
The lower temps for NB, SB, GPU, and GPU memory reflect a significant environmental difference in favor of the 350.
Merry Christmas to all.
great eyes Gabe :up:
And a happy New Year, Gabe. Also, good catch! :yepp:
Fair enough, some differences.
But did you see the mobo temps on the GTZ test are lower by 6C?
Also i added an 80mm and ram fans to cool the memory and NB , that is why the temps on the NB and vid card are lower with the 350.
This was just taken with same voltage 1.35 as gtz( but the mobo overvolts) thats why we see a fluctuation in voltages in Everest
Look at the CPU-Z the voltages are almost the same
http://i28.photobucket.com/albums/c2.../350retest.jpg
I for one hope Naekuh doesn't post his results here.....waiting for real tester results....
I personally cannot wait for Naekuh (a real tester) to post his results. Not only is he doing repeatable things openly (full methodology writeup forthcoming with final results, I presume), he's doing it with NDA silicon.
Thanks for the contribution, Nh. Don't let the 13 year olds push you out of doing something truly valuable for the community.
so 13 year olds cant contribute here?
And keep waiting all you want...
Not a lot of people have this hardware, and you probably dont, so i guess it plays out real well with you..
PM me.. i already got the results.. i just refuse to post them..
But only you.. and people on my friend list... :up:
I only had 2 blocks to compare with, and thats the supreme, and 350...
No, but 13 yr olds killed it for the rest of the people... Because i start acting like one also...
And yeah you may think my data is of very little use... but a lot of people do think its a lot of good, and you just lost it for them. :up:
Nice job!
:mad::mad::mad::mad::mad: Nice goin'. Maybe the ones who don;t want to see testing we can all help with that. Next time they ask a question, let's see about about making it a zero reply thread. I think that's fair. Anybody onboard with this idea?
Wow....you guys are stooping to begging for test results on a block?
If he wants to be childish and not post them, then let him.
I thought the community was supposed to be about sharing knowledge, not egos. :shakes:
come on guys relax for a bit its the holiday season this thread is easily becoming a 1 post info and a 20 post flame thread
I myself like to see everyone's result with different configurations and different water blocks, res,rad combination
how about we make it like a news thread and take everything here with a little grain of salt.
belated Merry Christmas and advance Happy New Year
just finished reading all the comments in this thread.
We are all here to share, learn, and contribute. Some of us are all too familiar, experienced and very well versed in what we do. Some of us are not and are on a journey to learn and expand on our knowledge.
"Share your knowledge. It’s a way to immortality".(Dalai Lama)
We all have the right to comment and add our opinions. Common sense, respect and courtesy should always be remembered though.
happy new year to all.
Is Swiftech the only one with a back plate for the 1366 motherboards? I didn't see any other ones, even on the Koolance.
Swiftech is probably the only one with a decent backplate.
Aside from that, there are a few blocks with support for socket 1366: EK Supreme, Aqua Computer blocks, Koolance 350.
I know a few have the hold down plate, but other tha the Swiftech, I didn't see any that offered the back plate.
koolance 350 comes with a backplate
http://i24.photobucket.com/albums/c5...5/bracket2.jpg
http://i125.photobucket.com/albums/p...a/IMG_0136.jpg
Here is the entire package..
the rubber with the hole is used for i7... the non hole, everything else..
i like gabe's backplate tho.. :X
That's the one. ^
Guys i need to appologize.
My test result was botched.
I just found out i did an improper mount for the 350. I THOUGHT the block was flat. So i didnt put a lot of mounting pressure on it. Its kinda pointless i feel on a flat block.
But i just found out that its BOWED.
So i need to retest the block all over again.
So at this moment in time no one really knows what is the best 1366 block out there...Also will d-tek me bringing out a backplate for its v2?
:rofl:
No...I didn't know the thing was bowed, but your message to me was more than enough of a reason to laugh at you.
remounted with pressure.. (KL-350)
:rofl:
thats all i need to say.. i cant get even cores (only first mount), but look at my LINPACK max temps.. :rofl:
Ambients taken at 25-26C. Via temp probe at the front intake of my rad.
http://i125.photobucket.com/albums/p...orla/Final.jpg
Coolant was recorded at 26.7-27C it was dancing in between those numbers...
im not gonna edit anything anymore since you guys want to see unedited stuff..
i'll post my supreme retest results when i get them, most likely next year,
but dude, i was getting those temps on prime95 on the supreme..
OCCT temps, @ same ambients 25-26C, coolant temp was lower.. like 26.3-26.8C
http://i125.photobucket.com/albums/p.../OCCTFinal.jpg
Load was held for 10min to achieve max delta... add 1C to those numbers... when i took the pic, the temps fell 1c below max.
Thanks, Naekuh, for the results; do you have any other blocks results (please)?
Sweet mother of poopysauce! Those temps are bananas.
Dang, with temps like that IntelBurnTest is pointless.
Hang on a second...did I just say that? I never thought I'd say something like that.
The CPU isn't even warm. That's unbelieveable! I mean I believe Naekuh, but that's incredible. :clap:
Man if I can get temps like that I'm gonna do Jumping Jacks and back Flips.
EK results will come next year.. promise..
today i get drunk.. tomorrow i stay in bed with hangover...
the next day i throw the supreme back on. :up:
Is that with a single DDC or the dual DDC setup? Iwaki? Please, do tell speciffics. You have about the same ambient I do in my room right now.
Nice temps There NaeKuh :up:
No offense, but my temps are a bit better IMO (using the GTZ), but I am with the Feser 480.
I am getting the same temps (relating to OCCT temps) at about 4100Mhz with about the same voltage using wPrime (HT and Turbo are on, full 8 Cores load).
But maybe my ambient is lower so we are not talking apples-apples comparison, I need to check again using the probe.
What rad/fans/RPM are you using ?
p.s.
I realize its going to take a while until you guys wake up and sober from hangover :rofl:
rofl..
yeah i have a bad hangover... :\
mojito's with beer... have no fear? yeah right...
atm I'm thinking bout the orientation of my ek block on the i7 but I haven't found a useful hint so far.
how are the 4 cores orientated? lengthwise the ihs like this?
http://media.arstechnica.com/news.me.../nehalem-1.jpg
if yes do u think it would perform better if I turn the block 90° clockwise?
http://alexiel.stor.li/x58/ek%20block.jpg
ouch... making me think while i have a hangover..
http://i125.photobucket.com/albums/p...a/IMG_0842.jpg
There we see a supreme... you see the direction of the injector plates?
You want that to lay on top of the longest side of your ihs..
Here you see one delided:
http://www.ixtremtek.com/ixtek/Hardw...A447/test1.jpg
You can see that its a rectangle.. so you want that plate slit right down the middle if possible..
This would give you the best implact on all 4 cores..
If you have it the other way, you can see you will intersect the die, and not spread over the die..
So flip your block 90 degree's and then a 180 to see which side it likes..
perfect, thanks =)
gonna test it
I'm going to build the first WC systems for my i7. At first,i chose Apogee GTZ; however, after reading this thread, i reconsider for the Koonlance 350 block. But, what does Koolance 350 block look like? What is it?
u mean the internals?
http://www.xtremesystems.org/forums/...37#post3484537
I tested several mounts with the ek:
- turned it 90° clockwise: temps were even worser than before.
- turned it another 180°: best temps so far.
But well...I'm still thinking bout getting a Koolance 350 too to compare the temps.
he flipped it 90 degree's... and he got bad temps.. meaning his injectors were carved in the other direction.. So he fliped it and got best temps. 180 degree's again.
I am not gonna say this block is better then a GTZ.
At the same time im not gonna say a GTZ is better than this block.
(i really dont have a GTZ to throw on my bed)
The only thing i can say is, most likely the 350 is better then the supreme on a lga1366..
Ok, I do not think I am reading this correctly. I thought he was flipping the actual block around!? Is he changing something in the block its self? If so I just want to know what I need to adjust to ensure that I get the best temps possible. :shrug:
Edit: I just read your post over here: http://www.xtremesystems.org/Forums/...7&postcount=13 I think I understand now
Hey Andyc,
Long time no see for sure. I'm just getting back into the swing of things to build myself two new WC'ed rigs before going back into my own self exile from anything remotely fun.
I'm glad to hear that Koolance finally wised up and cut the Alu crap. Let's hope the trend continues for all other former alu utilizers.
Anyhow, I just bought myself a new GTZ to mess around with for my new i7 builds. However, I'm wondering if I'm just wasting money since Niksuhb1's testing shows that my Fuzion V1 can still rock the socks off of any block on the market.
I love the temps the v1 fuzion gives me. My only problem is that it takes me at least 5-10 mounts before I'm finally happy with my results. The bowing with the o-ring creates a much larger variance in "good" vs "bad" mounts from my own experience.
Also, Enzotech has their new adapter for the Saphhire and Luna out. NCIXUS.com is the only one I know of with it though, only $5 though. I like my Saphhire block, easy too mount and is flat.
I'm sorry, here's a lil photo. I tested all 4 possible orientations this one works the best for me:
http://alexiel.stor.li/x58/ek%20test%20mount.jpg
And yes I'm using the EK backplate.
Here are some idle and load temps at stock settings and 8x100% load.
This mount gives me the lowest temp difference between all cores and the lowest core temps.
http://i180.photobucket.com/albums/x...wagain/3-2.jpghttp://i180.photobucket.com/albums/x...wagain/4-1.jpg
I'm glad you posted that last shot. I couldn't see how having the block vertical (inlet to outlet) could be better than horizontal.
Horizontal vs Vertical (in to out)
http://i180.photobucket.com/albums/x...wagain/1-3.jpghttp://i180.photobucket.com/albums/x...wagain/2-1.jpg