![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Intel Core i7 920 Overclocking and Heat @ [H]
- Thread starter FrgMstr
- Start date
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
If I'm going to install something that massive in there I'd probably just move on to a watercooling setup.
Anyhow I look forward to seeing your numbers once you get it on there.
I'll update my post when I get on my desktop ( currently on laptop ). What's most peoples ambient temps at? My 4.0 OC seems to be running a lot cooler than most peoples. off the top of my head I idle at 33-38 and under load I believe highest I've seen is 60.
Edit: Heres my idle temps. @Farkle: Yes that was under stress testing. During stress testing I saw it hit 60 once and then settled down to stay at 58-59.
Edit2: Coretemp and Realtemp are pretty close together, but the gigabyte program is quite a bit lower. Is it because thats the socket? anyone know how it works?
Edit: Heres my idle temps. @Farkle: Yes that was under stress testing. During stress testing I saw it hit 60 once and then settled down to stay at 58-59.
Edit2: Coretemp and Realtemp are pretty close together, but the gigabyte program is quite a bit lower. Is it because thats the socket? anyone know how it works?
Last edited:
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
Was that stress testing? I definitely don't get above 60 unless I'm doing stress testing. My ambient temp is about 21C (70F). What voltage are you running at?
My other post got big enough. Not sure exactly my ambient temp, I'll have to grab a thermometer. To answer your other question Farkle I'm not quite sure, it's been awhile since I've gone into my bios and looked at the voltages. I'll update this when I check and post some load temps for people too.
I'll update my post when I get on my desktop ( currently on laptop ). What's most peoples ambient temps at? My 4.0 OC seems to be running a lot cooler than most peoples. off the top of my head I idle at 33-38 and under load I believe highest I've seen is 60.
Edit: Heres my idle temps. @Farkle: Yes that was under stress testing. During stress testing I saw it hit 60 once and then settled down to stay at 58-59.
Edit2: Coretemp and Realtemp are pretty close together, but the gigabyte program is quite a bit lower. Is it because thats the socket? anyone know how it works?
You haven't got HT enabled.
Ok, but that'll be why your 920 is so cool, HT adds quite a lot to heat.
The V10 smokes the H50--smokes it. The H10 doesn't do much better than the Cogage.
I got it in but it was a whore trying to get the fan clips inserted with the Cooler Master 820 Stacker side fan door. I finally ny-tied the fucker to the fins. But it was really time consuming. I could have taken the board out, installed the V10, and reassembled everything faster than that. Pisses me off.
The push pin method was easy, and it sits perfectly in the heat sink top hole. I guess they redesigned that problem for the new Cogage's. But I can twist it left and right rather easily. I guess that's how all push pin coolers are?
Unless I got a ban fan, the Cogage's fan is way too noisy. And even slowed down it hums. With all the problems I had mounting the fan, I'm not looking forward to putting another one on it. I have a PWM Gelid Silent 120mm coming tomorrow too. Did you hear yuor Cogage fan? Was it loud?
Not really no, It's barely better than the TRUE and you have to make sure that the TEC can be cooled well and also its huge and expensive. Totally not worth it.
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
I never used the COGAGE fan. Went straight to the Scythe Gentle Typhoon 1850rpm. All my temps are with the Gentle Typhoon.
As for case fans I've got 3 Gentle Typhoon 1450rpm fans. 2 intake, one exhaust. The only other fan is the HX850 exhaust fan. I'm achieving positive pressure quite nicely (can feel the air coming out of the cracks in the case).
Since I have the Cooler master 830, it's like Swiss cheese. So I pointed all my fans inward except for the Stacker's blow hole, which used the Cooler Master fans that came with the case.
I ramped my rig to 3Ghz last night and I wasn't pleased with the temps. I think it hit 74 tops, and averaged around 69-70 running Prime for 30 minutes. After 30 minutes it leveled off. I just don't think there is anyway I'm going to hit 4Ghz temp wise. Although you said you were hovering at 89C, so that's another 15C I can push it.
Also my CPU fan isn't throttling itself. Have no idea why that is happening. I'm gonna go back in and see if everything is okay tonight. What are your idle temps?
I ramped my rig to 3Ghz last night and I wasn't pleased with the temps. I think it hit 74 tops, and averaged around 69-70 running Prime for 30 minutes. After 30 minutes it leveled off. I just don't think there is anyway I'm going to hit 4Ghz temp wise. Although you said you were hovering at 89C, so that's another 15C I can push it.
Also my CPU fan isn't throttling itself. Have no idea why that is happening. I'm gonna go back in and see if everything is okay tonight. What are your idle temps?
http://www.legitreviews.com/article/926/4/
http://www.legitreviews.com/images/reviews/926/i7.gif
Direct comparison V10 vs TRUE 120
10C difference between the V10 and TRUE 120. The thing is, you have to run that sucker hot to get the Peltier system to work, and once it's working, it kicks ass. If you look closely at the OC where the V10 doesn't do as well, you will see that it is a much lower heat test than using the 920 at 4Ghz. The test above is at a heat pumping 4Ghz.
Actually, here's a review that compares the H50 and V10 and the H50 actually does outperform the V10, by .2C and yes, that's "point2C" as in dot2 or .2!! And that's at, like I said, below 4Ghz, at 3.66Ghz. The hotter it gets, the more the TEC system kicks in.
http://www.rwlabs.com/article.php?cat=&id=203&pagenumber=5
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
That was at 4.2ghz though. Not 4.
dvsman
2[H]4U
- Joined
- Dec 2, 2009
- Messages
- 3,628
Good thread guys. I'm running a Cocage True Spirit (with stock Cocage fan) and am thinking of swapping out the cooler to the Corsair H50. I just don't like the looseness of the stock mounting brackets (not sure how this passed their QC) and I didn't want to have to resort to the penny mod or replacing the brackets with "push through" mounts which I see some guys mention now and again.
The RW Labs link was a good read. Looks like the H50 will provide a decent improvement over the stock True Spirit so that's good to know. Will report back once I have the H50 installed and get a chance to see how temps are (all other factors being equal).
The RW Labs link was a good read. Looks like the H50 will provide a decent improvement over the stock True Spirit so that's good to know. Will report back once I have the H50 installed and get a chance to see how temps are (all other factors being equal).
Last edited:
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
Why not just get the bolt through kit? Cheaper than going with the H50. Also why are you afraid of the washer mod?
I took out the cooler, cleaned it and the heat spreader, and reseated it. Then I put all my fans in, two GTs, two Gelid Silent PC/TCs, and twwo CM fans that came with the Stacker 830.
I also did a face test to see which fans blew the most, and the winner goes to the GTs. I held the fans, one on each hand, and tilted them back and forth (very scientific). If they were really close, I would not have felt any difference. The Cogage fan put out the next best amount, but it's loud. It's also really light. So I used it for the sink fan.
I also mounted each fan to the heat sink and tested how much air was flowing through just by feeling with my hand. The GTs were the best, and that was no surprise because they have the best static pressure. They are also NOT silent at 1850 RPM, but very quite running around 1000-1200.
The Gelid fans are quite, very well built, have really long lead lines, and a 5 year warranty. Even at 1500RPMs,which is their max, they were quieter than the GTS at 1850. They didn't move as much air though, even though they are rated about the same. The Gelid TC with its heat probe is REALLY nice. You can hold the probe in your fingers and it ramps up nicely to max. You could use this as the CPU fan by plugging it into a sys fan 3 pin header and by a process of trial and error, get the probe right where you need it, in relation to the CPU, and have it throttle up and down nicely. That would open the CPU header for another fan with PWM and that would then throttle that fan as the CPU got hotter.
Throttling doesn't seem to matter with my MB (Gigabyte X58 v5) since I'm watching the fans right now and they are all blowing max at idle. Really stupid. The only 4 pin connector that throttles is the sys2 fan, which is the only 4 pin fan riser other than the CPU--which doesn't seem to throttle at all.
Oh yes, one more thing. I made a sink shroud out of heating and sir conditioning extra heavy aluminum tape. I first sealed the top of the Cogage to fan junction, since a lot of air was simply gushing out of the top space. Then I ran the tape down the sides to block the air escaping from the side of the fins. I couldn't do the bottom though because there is no way to access it. The difference at the back of the sink fins was probably an increased 2x, which means twice as much air is now flowing through the fins instead of around them, or passing up and over them.
I ran again at 2800Ghz and the temp never got past 59C high. Idle temps dropped to 27C- 30C, whereas they were 55C at idle at 3GHz.
I'm going to go for it when I get a chance. I might just do a dumb BIOS OC now to see what temps are while the BIOS auto clocks it.
I also did a face test to see which fans blew the most, and the winner goes to the GTs. I held the fans, one on each hand, and tilted them back and forth (very scientific). If they were really close, I would not have felt any difference. The Cogage fan put out the next best amount, but it's loud. It's also really light. So I used it for the sink fan.
I also mounted each fan to the heat sink and tested how much air was flowing through just by feeling with my hand. The GTs were the best, and that was no surprise because they have the best static pressure. They are also NOT silent at 1850 RPM, but very quite running around 1000-1200.
The Gelid fans are quite, very well built, have really long lead lines, and a 5 year warranty. Even at 1500RPMs,which is their max, they were quieter than the GTS at 1850. They didn't move as much air though, even though they are rated about the same. The Gelid TC with its heat probe is REALLY nice. You can hold the probe in your fingers and it ramps up nicely to max. You could use this as the CPU fan by plugging it into a sys fan 3 pin header and by a process of trial and error, get the probe right where you need it, in relation to the CPU, and have it throttle up and down nicely. That would open the CPU header for another fan with PWM and that would then throttle that fan as the CPU got hotter.
Throttling doesn't seem to matter with my MB (Gigabyte X58 v5) since I'm watching the fans right now and they are all blowing max at idle. Really stupid. The only 4 pin connector that throttles is the sys2 fan, which is the only 4 pin fan riser other than the CPU--which doesn't seem to throttle at all.
Oh yes, one more thing. I made a sink shroud out of heating and sir conditioning extra heavy aluminum tape. I first sealed the top of the Cogage to fan junction, since a lot of air was simply gushing out of the top space. Then I ran the tape down the sides to block the air escaping from the side of the fins. I couldn't do the bottom though because there is no way to access it. The difference at the back of the sink fins was probably an increased 2x, which means twice as much air is now flowing through the fins instead of around them, or passing up and over them.
I ran again at 2800Ghz and the temp never got past 59C high. Idle temps dropped to 27C- 30C, whereas they were 55C at idle at 3GHz.
I'm going to go for it when I get a chance. I might just do a dumb BIOS OC now to see what temps are while the BIOS auto clocks it.
Last edited:
Fair enough, thoguh thats the BIGGEST difference i've seen. The main point is the size and price...
£91 for the V10
£60 for the H50...
£44 + fan
Now if you consider the fact that all of these will be able to handle an overclocked i7 at the same speed stable and good temperatures, regardless of a maximum of 10c cooler for the V10, i think you'd be a bit silly to go for it, unless of course you hav the money and anal enough to want say an improvement from 80c to a hopeful 70c.
I agree with that. But the V10 still smokes the H50.
I disagree, strongly. But each to their own.
Well I'm open to changing my mind, but I have yet to see any evidence to the contrary. I've given you my evidence, based, yes, on authority, but evidence nevertheless.
Where's yours?
Actually you are disagreeing with facts using an opinion. It's not my "opinion." It's the facts that I am agreement with. Opinions don't need facts, and as such, they have no persuasive power.
So what you are saying is that you disagree with the facts for no reason except that you disagree. That's an irrational position.
OK that's fair. When I use the word "smokes" I mean it was 10C lower in the test I provided, if and only if the CPU generates enough heat. It that test, the 920 was clocked to 4Ghz, and it is common knowledge that the 920 runs hot at 4Ghgz, which, in turn, provides a good control to test fans.
So yes, my arbitrary word "smokes" is an opinion. Given the evidence, don't you think 10C could be considered "smoking" the competition when even 1C could make or break and OC?
So yes, my arbitrary word "smokes" is an opinion. Given the evidence, don't you think 10C could be considered "smoking" the competition when even 1C could make or break and OC?
OK I have it clocked to 3.8. It posted and booted into Windows at 4Ghz, but BSODede when I fired up P95. It been running P95 now for 15 minutes and high heat is 88, 87, 84, 80.
I have the BIOS set to a Vcore of 1.5, but it shows only 1.36 tops in CPUz. Can someone tell me what is up with that? If CPU z is reading it correctly, then it was simply undervolted at 4Ghz.
I lowered my DRAM to 570x2 to make sure RAM isn't a problem. Plus, I have 2x6 sticks of RAM. Last time I OCed peple were saying you really have to slow the RAM down and lossen timings to get 12GB of RAM stable. Is this still correct?
I have the BIOS set to a Vcore of 1.5, but it shows only 1.36 tops in CPUz. Can someone tell me what is up with that? If CPU z is reading it correctly, then it was simply undervolted at 4Ghz.
I lowered my DRAM to 570x2 to make sure RAM isn't a problem. Plus, I have 2x6 sticks of RAM. Last time I OCed peple were saying you really have to slow the RAM down and lossen timings to get 12GB of RAM stable. Is this still correct?
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
I found CPU-Z to consistently under-report my CPU voltages compared to OCCT. If you are undervolting and really just at 1.36 then whatever cooler you have is performing worse than my TRUE Spirit, if you really are at 1.5v then that's not bad at all.
I found out what's up. Actually, your setting in the performance section does not correspond directly to true vCore because of vDrop (not droop). Set yuor voltage and then go look at it in the BIOS under PC Health, and you will see it is much lower than you set it.
OK so I went back in and set the vCore so it read 1.408 ion CPUz and BIOS. Clicked to 4Ghz, ran prime for about 10 minutes and was getting a high heat score of 90, just ticking it and then falling back to 87 or so. And that was only the 1st core.
I'm gonna run Vantage Pro right now and see what the temps do.
Then I'm gonna go back and set the vCore a little lower, to something like 1.39 and see if I can get it to run Prime for a few minutes. If I like my heat, then I burn it in over night. Then tomorrow I'll see if I can clock my RAM up a bit.
Taking vDrop into consideration, reading from CPUz, vCore of 1.392. P95 running for 15 minutes and High is 95. Average core 0 is 93C. Running Future Mark Vantage it never gets over about 60C. !!!!!!!!!!!!!!!!!! BSOD about 20 minutes into P95. Have no idea if it was heat or an error. I think I got one of those "binned" CPUs
Running again at vCore1.376 vCore (CPUz at load) at 3.9Ghz. I think this is going to be as high as I can go w/my POS. 15 minutes into P95 High temp is 89-90C.
FaRKle0079--any ideas on what to do now? I'm not too knowledgeable about OCing.
Last edited:
Your underestimating the h50, it works very well. Most run in the 70's or below at 4ghz. Corsari claims 10c better than the "best" air coolers. Thats a bold claim and I don't say its definitively true. But it is a excellent designed cooler. Where as I don't read many good things about the v10 at all.
Last edited:
Yeah the "incredible v10 hulk", Which cause of Its size probably doesn't fit all that many cases, I have an H50(still boxed), As I'm slowly building an i5 computer with It as the case needs filters and the house needs a new window and new weatherstripping[the old window & weatherstripping leaks like a sieve], So I'll sell off some parts to get the cpu and ram(older parts: Intel Q9300[3.51GHz capable], Asus P5K Deluxe, 2GB Patriot Viper pc2-8500 ram[runs at 440MHz @ 2.12v] and a heatsink w/mounting) The motherboard(a DFI LANPARTY DK P55-T3eH9) I'll have enough for on the 1st at least.
I have a fan on each side of the h50 rad both blowing out the back of my antec 300.Yeah the "incredible v10 hulk", Which cause of Its size probably doesn't fit all that many cases, I have an H50(still boxed), As I'm slowly building an i5 computer with It as the case needs filters and the house needs a new window and new weatherstripping[the old window & weatherstripping leaks like a sieve], So I'll sell off some parts to get the cpu and ram(older parts: Intel Q9300[3.51GHz capable], Asus P5K Deluxe, 2GB Patriot Viper pc2-8500 ram[runs at 440MHz @ 2.12v] and a heatsink w/mounting) The motherboard(a DFI LANPARTY DK P55-T3eH9) I'll have enough for on the 1st at least.
Your going to like it, I have the i5750 at 4ghz 1.40v it runs in the 60's prime95+Ibt full load.
Compared with the stock hsf, lol - It would go to the 80's at 3.4ghz
Dual fans, push pull, yum, I've got two all picked out too, amazon does a good job on that, I just have to keep from adding to the lists there. I've seen at least one i5 750 @ 4.216GHz.
The rad has holes on both side to mount fans, it comes with 4unique screw that fit the holes. I used 2 diagonally on each side to mount the fans. Which left two holes open on the antec rear fan to mount to inside of case (its original place) from the outside with the fatter case screws.Dual fans, push pull, yum, I've got two all picked out too, amazon does a good job on that, I just have to keep from adding to the lists there. I've seen at least one i5 750 @ 4.216GHz.
Does anyone have any experience using the H50 on a 920 at 4Ghz? Better yet, a 920 C0? Some 920s run a lot hotter than others, like mine.
I really wish somebody would do a head to head between the H50 and V10. They could use a Pentium D heater and a 920 both insanely OCed. As stated earlier, the Peltier system in the V10 WILL NOT WORK until the heat gets to a specific level. This is because Peltier systems usually work by adding current to a "cool plate" that creates a convection inside the heat block. the V10 does not have any current to do this; however, it does create a current on it own, a micro current, after the two plates--hot plate and cold plate--get to a certain temperature. Thus, before the hot plate heats up enough in relation to the cold plate, the Peltier system isn't working, and you just have a huge "V10 Hulk."
It's a very technical air cooler incorporating something no other air cooler ever has, and it works. A way to prove this is too look at the V10 scores on cooler CPU tests. It always gets outperformed by the good air coolers because it isn't warm enough to kick in the Peltier system. Once that system starts to work, the difference in cooling is dramatic. So yes, your idle temps will probably be a few C higher with the V10, but the high heat systems will always go to the V10. After a while, the water is going to heat up in the H20, or whatever they put in it, in a high heat system. In other words, the heat will exceed the H20's capabilities to cool.
If the H20 came with a larger radiator or simply a small tank with more fluid reserve, then it would outperform any air cooler.
I wish you guys would Google "Peltier Cooling" and read up on it.
Now, on the other hand, you do have my interest with the H20. I just haven't seen ANY information that says it outperforms air systems enough to try it on HOT systems.
I really wish somebody would do a head to head between the H50 and V10. They could use a Pentium D heater and a 920 both insanely OCed. As stated earlier, the Peltier system in the V10 WILL NOT WORK until the heat gets to a specific level. This is because Peltier systems usually work by adding current to a "cool plate" that creates a convection inside the heat block. the V10 does not have any current to do this; however, it does create a current on it own, a micro current, after the two plates--hot plate and cold plate--get to a certain temperature. Thus, before the hot plate heats up enough in relation to the cold plate, the Peltier system isn't working, and you just have a huge "V10 Hulk."
It's a very technical air cooler incorporating something no other air cooler ever has, and it works. A way to prove this is too look at the V10 scores on cooler CPU tests. It always gets outperformed by the good air coolers because it isn't warm enough to kick in the Peltier system. Once that system starts to work, the difference in cooling is dramatic. So yes, your idle temps will probably be a few C higher with the V10, but the high heat systems will always go to the V10. After a while, the water is going to heat up in the H20, or whatever they put in it, in a high heat system. In other words, the heat will exceed the H20's capabilities to cool.
If the H20 came with a larger radiator or simply a small tank with more fluid reserve, then it would outperform any air cooler.
I wish you guys would Google "Peltier Cooling" and read up on it.
Now, on the other hand, you do have my interest with the H20. I just haven't seen ANY information that says it outperforms air systems enough to try it on HOT systems.
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
You said you have a C0 chip? A lot of C0's had a hard time getting above 3.8ghz no matter what the multiplier/bclk combo was.
I do have that. I'm going to test today and see if i can hit 3.9. I guess the Gigabyte tech PDF guy got lucky. He cocked his C0 to 4Ghz and then used Turbo boost to hit 4.2Ghz and was stable. Makes me want to buy a D0 now. lol. But I don't think 200Mhz is worth it. Even if I have to go to 3.8, I think that will be nice. I can probably lower the voltage a little to and the combo will produce less heat all around. Even 3600 with turbo boost to hit 3800 under stress would be fine with me.
Thanks for that info. Do you have any exotic combos I could use to see if I can get it stable at 3.9? I have no idea if it is voltage or something else causing it to fail at 4Ghz. anyway, it's too hot at 4Ghz.
Don't you have a C0?
Thanks for that info. Do you have any exotic combos I could use to see if I can get it stable at 3.9? I have no idea if it is voltage or something else causing it to fail at 4Ghz. anyway, it's too hot at 4Ghz.
Don't you have a C0?
FaRKle0079
2[H]4U
- Joined
- Jan 11, 2005
- Messages
- 2,686
I have a D0 chip. Specifics are in my sig.
Also that "4.6ghz on air" chip won't last. He mentioned highest it would boot into the OS stable, not highest stable OC. I can get a large OC to boot into windows, but if it's not proven stable by stress testing then what's the use?
I'm extremely skeptical of his "under load not higher than 74C" figure as the Dark Knight isn't THAT good, especially at 1.55v.
I'm not sure any "exotic" combos will help you clock any higher. Chips are different and it's just luck of the draw.
Also that "4.6ghz on air" chip won't last. He mentioned highest it would boot into the OS stable, not highest stable OC. I can get a large OC to boot into windows, but if it's not proven stable by stress testing then what's the use?
I'm extremely skeptical of his "under load not higher than 74C" figure as the Dark Knight isn't THAT good, especially at 1.55v.
I'm not sure any "exotic" combos will help you clock any higher. Chips are different and it's just luck of the draw.