AMD competition for Core i3 (Gamers thread)

[DrMrLordX]

My guess would be the IGP kills the overclocking potential. Even if it has separate clock and voltage control, the fact that it's on the same package probably screws around with stability in one form or another. Look at i5 overclocking with the onchip PCIe.

I am sort of thinking the same thing. s1156 i5/i7 is a bit hamstrung by the on-die PCI-e controller, though problems with it can be partially overcome by just raising vcore as indicated in Anandtech's i5/i7 review (basically, increasing blck = increasing PCI-e controller clock speed, so to stabilize it, add voltage to the chip where it resides, which on an i5 and s1156 i7 entails raising vcore).

On i3 it almost looks like you'll have to increase voltage to the die with the imc, PCI-e controller and IGP in order to offset PCI-e controller overclocks. The IGP will hopefully drop into a power-saving state and/or be shut off entirely whenever a vid card is used, something of which I'm sure Intel is capable.

I was thinking of using a Corsair H50 mini water cooler for these two reasons.

1. It would fit nicely in the Sugo Sg05 and the radiator could be put directly behind the only fan (120mm front intake fan bringing cool air in from the outside)

2. I think a water block might do a better job at cooling the hot spot than a symmetrical heat pipe Tower cooler.

But honestly Turbo mode on Core i5 750 sounds pretty good (after reading post #233).

That cooler looks interesting. I wish there were more benches to show where it performed versus the Megahalems, TRUE, and other top-tier HSFs. I've heard the H50 is nearly silent so if that's a problem, it's worth a look. Only problem I've heard about it is that it sometimes offers up high idle temps and it often takes awhile to deal with load swings/spikes.

 

[Markfw]

You make a good point about water being harder to move.

But I just found the information here-->http://www.driverheaven.net/reviews.php?reviewid=819&pageid=3 and it says the pump only uses 2.5 watts.

I looked that up. There is a fan for the radiator, and then a motor for the pump. This is more expensive than my megahalems cooler, and I venture to say from my watercooling knowledge that it won't cool as large a cpu. Why use it ? More expensive, doesn't work as well, and the same or noisier.

 

[cbn]

That cooler looks interesting. I wish there were more benches to show where it performed versus the Megahalems, TRUE, and other top-tier HSFs. I've heard the H50 is nearly silent so if that's a problem, it's worth a look. Only problem I've heard about it is that it sometimes offers up high idle temps and it often takes awhile to deal with load swings/spikes.

http://www.madshrimps.be/?action=getarticle&number=3&artpage=4176&articID=939

I haven't read that much on this cooler but here is a review comparing it to TRUE.

http://www.legitreviews.com/article/1025/6/

In this review Corsair H50 does have a higher idle temp than stock box cooler.

 

[cbn]

I looked that up. There is a fan for the radiator, and then a motor for the pump. This is more expensive than my megahalems cooler, and I venture to say from my watercooling knowledge that it won't cool as large a cpu. Why use it ? More expensive, doesn't work as well, and the same or noisier.

http://translate.google.com/transla...test_corsair_h50_fluessigkeits-cpu-kuehler/6/

This is a German test of Corsair H50 vs Megahalems using two different fans.

I have seen Corsair H50 for $55 on sale at Fry's before so it doesn't really have to be that expensive.

 

[Markfw]

http://translate.google.com/transla...test_corsair_h50_fluessigkeits-cpu-kuehler/6/

This is a German test of Corsair H50 vs Megahalems using two different fans.

I have seen Corsair H50 for $55 on sale at Fry's before so it doesn't really have to be that expensive.

Even if they are the same price, the H50 still does not perform as well as the megahalems, and uses at least a little more power, and is still restricted as to what it will cool. Example, a friend of mine, and elite member here(Duvie), has a Coolit watercooler. It will not keep his Q6700@3.4 as cool as a TRUE, since it maxes out its cooling capacity (175 watt), and the I7@3.9 is supposed to produce even more heat @ 100% load, but mine is cool at 60c. From the looks of this one, its not any better than the coolit, and it does not cool as well even by your link, so, again, I will stay with my megahalem.

 

[cbn]

Example, a friend of mine, and elite member here(Duvie), has a Coolit watercooler. It will not keep his Q6700@3.4 as cool as a TRUE, since it maxes out its cooling capacity (175 watt), and the I7@3.9 is supposed to produce even more heat @ 100% load, but mine is cool at 60c. From the looks of this one, its not any better than the coolit, and it does not cool as well even by your link, so, again, I will stay with my megahalem.

http://www.bit-tech.net/hardware/cooling/2009/07/29/corsair-hydro-h50-cpu-cooler-review/4

Are you talking about the Coolit Domino? Like I said I haven't read too much about Corsair H50 yet but according to this review it beat Coolit Domino even when it is on high.

"It also pretty much annihilates the Domino ALC when we consider noise and performance together - only at full speed can the Domino ALC keep up, but it's a few degrees off the much, much quieter Corsair H50 that uses a single slow spinning 1,500rpm fan, compared to the Domino's beastly 3,000rpm+ fan."

 

[cbn]

Even if they are the same price, the H50 still does not perform as well as the megahalems

The link I gave in post #254 for performance mode with Noctua 1380 rpm fan says different.

http://forums.anandtech.com/showthread.php?t=2029465

However, according to this thread I am linking Rubycon says Megahalems will beat Corsair H50. But in order to get the 6-8 degree advantage 200 cfm worth of fans have to be installed next to the Megahalem fins.

I wonder at that level of cooling and even below that (with Megahalem) if other parts on the mainboard might break first. Seems like that tower is over-engineered for 24/7 overclocking.

 

[edplayer]

Seems like that tower is over-engineered for practical 24/7 usage.

its not designed for practical 24/7 usage.

Its designed for serious overclockers.

 

[edplayer]

and that 6~8° C advantage is with the same fans on both (according to thread you quoted):

Yes the H50 is about 6-8°C warmer under higher load (W3520 @4.4GHz 1.3V) than Megahalems shadow. Same fans used for both - 1800RPM Gentle Typhoon in push pull. Tested open air on bench, ambient temps identical.

but why do you care about any of this since spending the extra money on a H50 cooler is like burning your hard earned money?

 

[cbn]

its not designed for practical 24/7 usage.

Its designed for serious overclockers.

I just edited that post.

It no longer says practical 24/7 usage. I changed that sentence to say 24/7 Overclocking instead.

 

[cbn]

and that 6~8° C advantage is with the same fans on both (according to thread you quoted):

Well Sure. You are comparing a tiny constant speed pump to a massive heatpipes.

The point is the radiator, pump and fan are matched to more realistic 24/7 OC conditions.

Furthermore according to my review with OC i7 the only way Megahalems can beat Corsair H50 is if the user increases fan speed and noise.

 

[edplayer]

and going up to 200cfm gives an additional 5° C advantage

I wouldn't say the H50 is more closely matched. You could say its limit is closer to more realistic 24/7 OC conditions. The Megahalem would cool even better under such conditions.

 

[cbn]

The Megahalem would cool even better under such conditions.

According to that German review Megahalems does worse than Corsair H50 @ 3.4 Ghz Overclock when fan speed is 1380 rpm.

If the fan is changed and speed is increased to 1860 rpm Megahalems will pull away.

How they would compare at 4+ Ghz Overclocks I have no idea. Hotter water being pumped by a constant speed pump to a radiator vs heatpipes moving expanding air to fins? Which one is more efficient heat transfer process with a 1380 rpm fan?

 

[edplayer]

Which one is more efficient heat transfer process with a 1380 rpm fan?

why do you want it to be specifically 1380rpm?

is it because that is the the only rpm range that it could beat the Megahalem (as in one out of five tests)? How about picking a decibel level that you would find acceptable?

 

[cbn]

why do you want it to be specifically 1380rpm?

is it because that is the the only rpm range that it could beat the Megahalem (as in one out of five tests)? How about picking a decibel level that you would find acceptable?

http://www.newegg.com/Product/Produc...-065-_-Product

I think this fan would be too much noise for 24/7.

 

[edplayer]

I think 35 dB is quieter than a stock Core 2 Duo...

the Gentle Typhoon you linked earlier look pretty good. 28dB for 98m³/h

 

[cbn]

http://www.xbitlabs.com/articles/coolers/display/corsair-h50.html

I finally did find a review testing Corsair H50 with 45nm quad core @ 4 Ghz.

The thermalright IXF-14 wins by a good margin. But look at how much noise the testing fans make in the chart? Those units are pretty loud even at very low rpms.

This is why I question how applicable these results are for 24/7 OC. It looks good on the bench but how valuable is something like IFX-14 (compared to other options) when it needs the fan(s) turned down to make it tolerable for 24/7?

 

[edplayer]

are you high?

Those acoustic tests have nothing to do with the heatsink and everything to do with the fan. At 2000rpm Thermalright claims 80.5cfm. 63.7cfm at 1600rpm, that is slightly more than the gentle typhoon.

What happens when you put those thermalrights on the H50?

it makes the same amount of noise!


and it still loses.

 

[cbn]

are you high?

Those acoustic tests have nothing to do with the heatsink and everything to do with the fan. At 2000rpm Thermalright claims 80.5cfm. 63.7cfm at 1600rpm, that is slightly more than the gentle typhoon.

What happens when you put those thermalrights on the H50?

it makes the same amount of noise!


and it still loses.

Why would anyone put those loud fans on the Corsair system?

H50 is limited by the radiator. If you read the xbit review lowering pump voltage from 12v to 7v does not affect cooling that much.

 

[edplayer]

Why would anyone put those loud fans on the Corsair system?

probably for the same reason as they would put it on the IFX-14

they don't know any better or they just had them laying around

H50 is limited by the radiator. If you read the xbit review lowering pump voltage from 12v to 7v does not affect cooling that much.

and the stock fan is still louder than the pump so who cares

 

[edplayer]

did you see how that stock fan scored 42.5 dBA at 12V?


and you thought 35 was unacceptable...

 

[cbn]

and the stock fan is still louder than the pump so who cares

The major point to understand is that the heatsink is the limiting factor of how well the fan(s) can be utilized. You said this yourself in post #268.

With IXF-14 there is much less efficiency drop off with loud fans. In Contrast, the H50 radiator cannot make use of higher airflow (see page 7 of Xbit review)

But off the bench and in the hands of someone using the system 24/7 this is of little consequence. The H50 is already loud enough in its efficiency range.

 

[edplayer]

So you get the best of both worlds

it doesn't cool as well as the best air coolers and it is louder.

At least it has bling

 

[cbn]

it doesn't cool as well as the best air coolers and it is louder.

Xbit never tested H50 against IXF-14 with a single 120mm fan on both set-ups. So we cannot say IFX-14 is better than H50 at lower decibels.

The lowest setting they used was two thermalright fans @ 1000 rpm. Two fans makes more noise than one fan.

You have to remember the radiator only has so much capacity for airflow.

 

[Markfw]

The radiator on the H50 is smaller than on the megahalem or the ifx-14. The radiator ultimately exchanges the heat. Without using chilled water (which the H50 does NOT) the other two can exchange more heat than the H50, how quickly, efficiently or quietly is another matter. The Coolit that Duvie has uses a peltier cooler on the water, and still can not beat the megahalem. Until I see a benchmark with both on at least a 4 ghz I7 @100%load with HT, I will not believe it is better. edplayer seems to agree with me.