thewhiteguy
Supreme [H]ardness
- Joined
- Jun 17, 2002
- Messages
- 4,551
They're pretty quiet at 12V, but I run them at 7V and my Zalman GPU cooler (CNPS supposedly) at 80% is still louder than anything else.el rolio said:
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
SystemCooling + AC
- Thread starter Erasmus354
- Start date
zer0signal667
2[H]4U
- Joined
- Jan 11, 2002
- Messages
- 3,847
thewhiteguy said:
The target audience for reviews and data like this IS the extreme overclockers. Non-watercoolers don't care, non-overclocking watercoolers might kind of care, and mild overclockers probably aren't that concerned with the data at hand. If you don't think that 8C is a big deal, I don't see why you're even here arguing about it...
thewhiteguy said:
Two to three degrees is something that we watercoolers typically lose sleep over. Four degrees, and many of us are pulling our hair out. At eight degrees, we're talking about mass murdering in postal offices
This obviously isn't the case for everyone, but temperature is important to any extreme cooler, extreme overclocker or not.Yes, I still consider our water cooling setups to be extreme cooling solutions. The products have changed dramatically over the last few years, but the facets of what we do has remained the same. We still play a very dangerous game, and the majority of us do it for the drastic reduction in core temperature.
I agree that 8 degrees is excessive, but I have trouble believing it was 8 degrees. When I was using the Zalman 9500's on my Opty's my idle temps were 50-52C, with two XT's my idle temps are 29-31C. I believe it has a lot to do with my Rads and the Aquastream being OC'ed. I still consider those temps to be good with four blocks and two Rads. The temps in both cases were reported by the Tyan Monitor, according to my inline temp probes, the temps are even lower.
EQDuffy said:
This doesn't make the Cuplex XT a bad block. I take issue with this sort of blind, meaningless
intention to trash AquaComputer products. The intention is fruitless and childish.
We can go back to the Masserati/Ferrari analogy if you really think we need to do this.
thewhiteguy
Supreme [H]ardness
- Joined
- Jun 17, 2002
- Messages
- 4,551
Remember I was talking about the biggest difference (actually rounded up slightly) in blocks on that chart, from the Maze3 to the Storm, with the Maze3 at the lowest flow and the Storm at the highest flow. If you talk about everything but the Maze3 in the same way as I did before and again on 100W, it's like 4.5 degrees. I don't care how hardcore you are about temps, but that's within the margin of error of many sensors.zer0signal667 said:
zer0signal667
2[H]4U
- Joined
- Jan 11, 2002
- Messages
- 3,847
thewhiteguy said:
I don't agree to that, especially not with quality test equipment (like I would hope is used in such data collection). Even cheap thermistors may might only be consistently "off", not prone to 5 degree fluctuations without some other change in conditions.
Poor end-user measurement does not change performance.
thewhiteguy
Supreme [H]ardness
- Joined
- Jun 17, 2002
- Messages
- 4,551
No, but an overclock is very unlikely to be affected by 4.5 degrees (and we're talking about low temps already, not 4.5 degrees on top of 60), so if you post temps 4.5 degrees higher than other people they're not going to say OMG YOUR BLOCK SUCKS because we all know how bad temperature sensors typically are.zer0signal667 said:
You're centered on absolute performance regardless of whether or not it matters, I'm talking about the real world where some people might care about 4.5 degrees, but it doesn't really have any bearing on anything.
- Joined
- Jan 2, 2001
- Messages
- 38,507
R1ckCa1n said:
shhhhhhhhhh.....don't bring common sense and logic into play, it fucks up the aquatrolls
That is what the first graphs on the review tell us, how it performs in an apples to apples. Now you put it in a 3 year old motherboard and try to compare, it is rediculous. As stated, my temps are better with a Cuplex Pro, Cuplex X1900, Cuplex X1900 in one loop than shown in the last graph so who's temps are right?thewhiteguy said:
Erasmus354
[H]F Junkie
- Joined
- Mar 12, 2004
- Messages
- 9,450
R1ckCa1n said:
I find it hilarious that now that this review is out real world testing is all of a sudden worthless and only the die sim results matter. Where is that link to TN's large smily that is rolling around laughing when you need it? The back peddling is so furious I am surprised some people here haven't hurt their neck from the whiplash and the stiff about face they made with their positions.
Erasmus354 said:
QFT. And no I'm not flaming anyone. Even when Cathar was on here trying to help folks be reasonable concerning real world performance, he was ignored. Many other folks who were telling people to use their brain a little, and not just listen to propaganda spread by certain individuals, were ignored.
Lee's review should enlighten people to the fact the the "whole watercooling system" should be taken into consideration when purchasing equipment. Stick a Storm on a Bigwater kit and it will perform poorly. This is what happened with the Cuplex XT, very good block stuck with poor components, as I said before.
This also shows that retailers, generally speaking, are dumb as dirt and only want to sell product. If Sharka had done any kind of research what so ever, they would know what components to choose to get optimal performance. The information is out here and freely available. Unfortunately most people find it easier to let someone else do the legwork for them.
Again not to step on anyones toes, but as of the last few months or so this forum has become a huge pile of missinformation. A laughing stock so to speak. Why? Because those who are ignorant or just plain wrong are making the most noise, drowning out the few voices of logic and reasonableness. Sad situation indeed...
ikellensbro
[H]ard|Gawd
- Joined
- Feb 14, 2004
- Messages
- 1,410
Maybe it's good that no one noticed the thread I started a few days ago linking an Overclockers.com article by Bill Adams about what we've been discussing the last 20 posts or so. My views:
8*C is a lot, and even 1*C in a block makes a difference. When you have a waiting list of people for a $300+ water block (Storm G5/G7), you can tell that a little bit more performance is a highly valuable commodity. Water blocks have become so advanced and efficient, there isn't much more room left to go. If some new block came out that could beat a G5 by a single degree, it would become sold out almost instantly (assuming it was priced as most regular blocks right now, of course).
When you take that 1*, add it to the benefit of properly arranging your fans and radiator placement, use larger fittings with the tubing stratched over the barbs, reduce your tubing lengths and bends as much as possible, etc., you wind up with temperatures a couple degrees lower (possibly). Now, with an average OC, you're mostly concerned with keeping your temperatures below whatever level you deem 'cool'. When you're truly pushing your cpu/gpu to its limit, the cooler you keep the chip, the better chance you have of keeping that high OC stable, longer. Why are so many high OC's done on phase, dry ice, or LnO2, when a proper loop can keep the temperatures around 40*C?
Another idea, that I think could work pretty well: If reviewers have access to most of the popular pumps used today (Eheim 1048, Mag, DDC, D4/D5/50z), why not design a loop (radiator and toobing lengths kept constant) with either an actual cpu or a die simulator, and test each block with each pump? Numbers and graphs are nice, but actual sytems that readers will use would give them a much better prediction of which block would fit them best (and if they already had a block picked out, which pump to go with it). This is a lot we're asking from SystemCooling, but some nice tests (with accurate temp sensors) would settle a lot of arguments, and take out the innaccuracies of trying to predict how much flow x block will receive compared to y, which performs better, but is more restrictive.
One more thing: I always liked ProCooling's interactive graph, and testing down to 0.25gpm flowrate, because it seemed like the XT may have been receiving around that much flow with 1/4"id tubing in SystemCooling's review.
8*C is a lot, and even 1*C in a block makes a difference. When you have a waiting list of people for a $300+ water block (Storm G5/G7), you can tell that a little bit more performance is a highly valuable commodity. Water blocks have become so advanced and efficient, there isn't much more room left to go. If some new block came out that could beat a G5 by a single degree, it would become sold out almost instantly (assuming it was priced as most regular blocks right now, of course).
When you take that 1*, add it to the benefit of properly arranging your fans and radiator placement, use larger fittings with the tubing stratched over the barbs, reduce your tubing lengths and bends as much as possible, etc., you wind up with temperatures a couple degrees lower (possibly). Now, with an average OC, you're mostly concerned with keeping your temperatures below whatever level you deem 'cool'. When you're truly pushing your cpu/gpu to its limit, the cooler you keep the chip, the better chance you have of keeping that high OC stable, longer. Why are so many high OC's done on phase, dry ice, or LnO2, when a proper loop can keep the temperatures around 40*C?
Another idea, that I think could work pretty well: If reviewers have access to most of the popular pumps used today (Eheim 1048, Mag, DDC, D4/D5/50z), why not design a loop (radiator and toobing lengths kept constant) with either an actual cpu or a die simulator, and test each block with each pump? Numbers and graphs are nice, but actual sytems that readers will use would give them a much better prediction of which block would fit them best (and if they already had a block picked out, which pump to go with it). This is a lot we're asking from SystemCooling, but some nice tests (with accurate temp sensors) would settle a lot of arguments, and take out the innaccuracies of trying to predict how much flow x block will receive compared to y, which performs better, but is more restrictive.
One more thing: I always liked ProCooling's interactive graph, and testing down to 0.25gpm flowrate, because it seemed like the XT may have been receiving around that much flow with 1/4"id tubing in SystemCooling's review.
thewhiteguy said:
im sorry man, i gotta tell you 4.5*C can really matter alot. i have 3 opterons, and am stuck in the middle of a entire loop/system upgrade to acheive that kind of difference. like im talking newer more powerful and expensive radiator (thermochill triple), movin where the radiator is to pull as cool air as possible and even changing my additive!
why? cuz the hotter opterons begin to run into a scaling wall as soon as you pass 40-42*C....
i mean this is why we uncap them in the first place!
like i said. the differnce between say 40* load and 44.5* load can mean a lot to a lot of us, you just cant dismiss it like that. 2*C difference in a calibrated test is enough to separate the the leaders in waterblocks.
nikhsub1 said:
And nikhsub is right.
A maximum overclock, if not limited by other components, is based solely on the effectiveness of a cooling system to reduce the temperature of the (or each) core. If there is an 1 degree variance between two systems (assuming that a thermal diode's output represents the hottest portion of a die), the system with the lower temperature will yield a higher maximum overclock, and this can be far more substantial than you may think.
It's not entirely a temperature diode game, as the block still has to be effective as transferring the majority of heat from the IHS or the die into the coolant, but overall die temperature is everything.
There's no chasing tails about this - it's pure science. It's the same science that states that high flow is more effective than low flow, and there are simply no buts about it.
thewhiteguy
Supreme [H]ardness
- Joined
- Jun 17, 2002
- Messages
- 4,551
So you're going to get yourself to a borderline temperature that will cause instability if the ambient temperature goes up a couple of degrees? That doesn't sound wise to me.el rolio said:
DFI Daishi
2[H]4U
- Joined
- Feb 14, 2005
- Messages
- 2,839
and then you back it off as spring/summer approaches. what the big deal? if you don't want to stomache any risk then just get your self a high end dell and have done with it.thewhiteguy said:
that's how i opperate, at least......i've actaully had to back off from 300 even to 295 as the weather has gotten warmer. if i'm doing something that requires the utmost stability, i turn off the PC and fire up the trusty old powermac G4.
i'd say that watercooling and serious overclocking is for those who don't feel like waiting 6 months for the next advancement in processors to come around until they can run their games at full quality with playable framerates. wisdom has little to do with it, the desire to "go fast" has everything to do with it.
so far as the review goes: the A-C gear is pretty. we all knew in advance that A-C gear is pretty. there are some problems to be addressed with the present review, and i hope that robotech will address those issues in the coming months.
untill they have run the review with the pump at full speed, i don't have much to say other than that calls for the use of a rad other than then airplex kind of runs contrary to user claims that A-C provides "a complete and optimized watercooling SYSTEM."
fish or cut bait A-C fans: either you admit that no one manufacturer does everything the best and that entusiasts get their blocks from one company, their rads from another, their pumps from a third and fatter tubing yields better performance (up to a point), or you accept poor performance in reviews of your favourite equipment because they use ALL A-C equipment, "in it's intended usage."
According to a test done on a three year old motherboard, yes. Did it get beat at the same sound level, no.RotorHead said:Been following this thread for a while......
Correct me if I'm wrong......But did the AC just get beat out by an XP-120 with a 12v L1A ???
And $300 cheeper
Not really, if the realworld was done with same loop different block and a current generation motherboard to see which gives optimal overclock it would hold water with me. The throw it in a 3 year old motherboard means nothing to me.Erasmus354 said:
Not counting how much louder each solution is compared to the configuration tested Sharka supplied.
R1ckCa1n said:
To summarize:
- Lee's data is useless due the fact that he used a motherboard that is three years old
- Thermocouples have a bias against small bore, German made cooling products
- The science of processor cooling using a liquid heat transfer medium does not apply equally
- Cooling systems that decrease the die temperature (as much as eight degrees) do not yield higher maximum overclocks
Put it this way, my temps are much lower with three times the heat load in the loop (X2 2.6ghz @ 1.5v, X1900XT x 2), so I call the realworld, wallyworld on this review. It appears Lee must have rushed and shortcutted the mounting?phide said:
But hey, the graphs show all you need is a gargen hose and marine grade pump to make it work like a Storm AND look good
R1ckCa1n said:But hey, the graphs show all you need is a gargen hose and marine grade pump to make it work like a Storm AND look good
Hey, there's no denying that
I originally didn't like the look of the Cuplex (a little too macho, maybe), but it has somewhat grown on me. I suppose it still bothers me that they have to stamp "Aqua Computer. Made in Germany" on every feasible portion of everything they make, but it's not so bad.
Erasmus354
[H]F Junkie
- Joined
- Mar 12, 2004
- Messages
- 9,450
R1ck, I would quit while you are ahead. You keep changing what to you is defined as acceptable testing, and frankly I am not even sure what you think is acceptable anymore.
Whether the test was performed on a brand new, 1 year old, 2 year old, or 3 year old motherboard doesn't make a lick of difference. The temperature readings will be just as accurate as they are on motherboards today. Every different motherboard seems to have its own bias on temperatures, as long as all these kits were tested on the same configuration (yes) then it doesn't matter. Besides the fact that the temperature readings are taken from a thermocouple attached to the IHS if I am understanding Lee correctly.
Simply because your results are different you immediately are throwing out the entire test calling it useless. Something many of us have been trying to bang into your stubborn little head is that you cannot compare one real setup directly to another. There are too many variables to take into consideration. I think Lee has mounted waterblocks hundreds if not thousands of times by now. He could probably do it in his sleep. In case your selective memory blocked it out, Lee tested the repeatability of the mounting of the Cuplex XT. So there goes your one theory as to the invalidity of the test data.
Also, the testing was done on the kit provided, as a kit. The real world testing is to test the kit as a whole against other commercially available kits. Notice how there wasn't a "Storm + PA120.2 + AQX-50Z" in the test, everything was a kit. Therefore it is perfectly accurate for him to compare each setup as its own individual entity. The die simulator testing was done to provide an idea of how the Cuplex XT compares on its own merits. And the answer to that is and has agreed upon that it is a well designed, albeit very expensive, waterblock. It is also worse than the Storm so please stop trying to call it superior and/or equal, it is close but still worse.
FWIW there have been a lot of very valid concerns brought up by myself, phide, TN and other more level headed people. I pointed out that the acoustic testing doesn't go into a range low enough to really show off how quiet the AC setup probably was. TN pointed out that the pump was probably not running at full potential.
Finally, testing overclocking really doesn't have much merit here. The reason being that every piece of hardware is different. Some hardware will benefit from that extra bit of cooling while others may hit a wall that takes drastically lower temperatures to move forward. What isn't really debatable is that more cooling is always better and could possibly improve your overclock. Therefore if you simply test the cooling capabilities of a kit and/or waterblock that is enough for consumers to make an informed decision. Just because on your computer a different solution didn't yield a higher overclock does not mean that it wont for someone else. All that means is that the worse of the two solutions was good enough for your setup, not good enough in general.
Whether the test was performed on a brand new, 1 year old, 2 year old, or 3 year old motherboard doesn't make a lick of difference. The temperature readings will be just as accurate as they are on motherboards today. Every different motherboard seems to have its own bias on temperatures, as long as all these kits were tested on the same configuration (yes) then it doesn't matter. Besides the fact that the temperature readings are taken from a thermocouple attached to the IHS if I am understanding Lee correctly.
Simply because your results are different you immediately are throwing out the entire test calling it useless. Something many of us have been trying to bang into your stubborn little head is that you cannot compare one real setup directly to another. There are too many variables to take into consideration. I think Lee has mounted waterblocks hundreds if not thousands of times by now. He could probably do it in his sleep. In case your selective memory blocked it out, Lee tested the repeatability of the mounting of the Cuplex XT. So there goes your one theory as to the invalidity of the test data.
Also, the testing was done on the kit provided, as a kit. The real world testing is to test the kit as a whole against other commercially available kits. Notice how there wasn't a "Storm + PA120.2 + AQX-50Z" in the test, everything was a kit. Therefore it is perfectly accurate for him to compare each setup as its own individual entity. The die simulator testing was done to provide an idea of how the Cuplex XT compares on its own merits. And the answer to that is and has agreed upon that it is a well designed, albeit very expensive, waterblock. It is also worse than the Storm so please stop trying to call it superior and/or equal, it is close but still worse.
FWIW there have been a lot of very valid concerns brought up by myself, phide, TN and other more level headed people. I pointed out that the acoustic testing doesn't go into a range low enough to really show off how quiet the AC setup probably was. TN pointed out that the pump was probably not running at full potential.
Finally, testing overclocking really doesn't have much merit here. The reason being that every piece of hardware is different. Some hardware will benefit from that extra bit of cooling while others may hit a wall that takes drastically lower temperatures to move forward. What isn't really debatable is that more cooling is always better and could possibly improve your overclock. Therefore if you simply test the cooling capabilities of a kit and/or waterblock that is enough for consumers to make an informed decision. Just because on your computer a different solution didn't yield a higher overclock does not mean that it wont for someone else. All that means is that the worse of the two solutions was good enough for your setup, not good enough in general.
I am not going to waste my time with you on this one. I am humored that I am not happy with the motherboard tests and all the aquatrolls, like yourself, are not concentrating on the "die" test which shows the Cuplex right there with the best of the best.Erasmus354 said:
Anyways, the test was a serious disservice to the AquaComputer product once it was put on a motherboard.
thewhiteguy
Supreme [H]ardness
- Joined
- Jun 17, 2002
- Messages
- 4,551
WTF does the motherboard have to do with the performance of a water cooling system? The CPU is what makes the heat, and it will make the same amount of heat on any motherboard that supports it.
CCUABIDExORxDIE
2[H]4U
- Joined
- Jan 30, 2005
- Messages
- 2,957
it has everything to do with it! especially since the same motherboard was used on each test setup.
The "problem" with this review is that the Aqua Computer genie has been let out of his lamp and guess what, he didn't grant any wishes. The miracle of small bore tubing, weak pumps, and tube style rads just didn't happen. Instead it performed just like any other small bore setup, average.
Welcome to TT Bigwater country fellas!!!
Welcome to TT Bigwater country fellas!!!
Punx_Clever
Gawd
- Joined
- Jan 10, 2005
- Messages
- 579
R1ckCa1n said:
Actually, what the whole review shows is what E354 and the others have said all along: The block is a good block, one of the best out there. But when you put it with 1/4" tubing, a weak pump, and an outdated tube-style radiator, well, it dont matter how good the block is, the loop is gonna suck. The watercooling "system" that TN has preached about AC providing has been blown out of the water, not the block.
zer0signal667
2[H]4U
- Joined
- Jan 11, 2002
- Messages
- 3,847
thewhiteguy said:
That's right, as long as the same equipment is used throughout the test, I don't care if it's 0 or 3 years old.
- Joined
- Jan 2, 2001
- Messages
- 38,507
Punx_Clever said:
ummm, from what i've seen TN has "preached" that the AC rads are not the best, so she uses BI ones, and the pumps can be OC'ed while the one in the review was not, and she is running several different loops in order to limit flow loss due to too many componets in one loop.
Care to rephrase?