|
When using this resource, please read FragKrag's opening post. The Tech Support forum regulars have helped create countless of desktop systems without any compensation. The least you can do is provide all of the information required for them to help you properly. |
I'm not sure if those figures make that much sense. I don't exactly follow water cooling benches (at all), but are people really testing these things, or just guessing? You may be asking the wrong forums about this kind of thing, as I don't know if any of us have done or regularly bother with custom water; that said, sometimes experience and personal anecdotes get in the way of broader thinking and common sense, so maybe not.
Asus ROG Ares II (dual 7970s) cools the two GPUs with an Asetek closed-loop system connected to a single 120mm radiator. VRMs and RAM are cooled by a separate fan, and the radiator is pretty thick (though doesn't look super-extraordinarily dense or anything like that), but again... single 120mm rad.
Noise not that crazy:
http://www.techpowerup.com/reviews/ASUS/ARES_II/26.html
Temps actually pretty low:
http://www.techpowerup.com/reviews/ASUS/ARES_II/32.html
Power draw mostly in line with what you expect from those two GPUs:
http://www.techpowerup.com/reviews/ASUS/ARES_II/25.html
I don't think it should be much different conceptually than most air cooling. Think about the temperature differentials involved. The water is there to transfer heat from the CPU block to the radiators, where fans force air of a lower temperature through the fins. Obviously it depends on the radiator thickness, fin density, fan speeds, and more. This is assuming the pump and coolant are doing their jobs and making some more simplifications. Things hardly scale linearly, but as far as I know, you can do some rough estimation just thinking in terms of fin surface area.
Push/pull is more important if the fans have lower static pressure (are worse, blades not optimal for the task, and/or lower fan speeds), meaning they have more trouble blowing through obstructions, and if the radiators are thicker and denser, meaning more restrictive. Same as with other kinds of heatsinks. If you have an overkill amount of radiator area, it probably shouldn't make much difference.
|
What is your budget?
600$ - 650$
What is your resolution?
Will be using TV for now. 1920x1080
What are you using it for?
Mainly gaming, Skyrim, Farcry 3 and Starcraft. Won't be streaming with it.
What is your upgrade cycle?
2 - 3 years.
When do you plan on building it?
Plan to build is sometime in the next month.
Do you plan on overclocking?
Possibly some slight overclocking. But nothing very high.
Do you need an operating system?
Nope. Already own one.
Do you plan to add a second GPU for SLI or Crossfire?
No.
Where are you buying your parts from?
All of it will be bought off of newegg.
|
OMG OMG OMG IVB CELERONS HAVE ARRIVED!@#!@!@#!@!@#!@
http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007671 600315409&IsNodeId=1&name=Ivy Bridge
On January 30 2013 09:51 Turtle104 wrote:+ Show Spoiler +
What is your budget?
600$ - 650$
What is your resolution?
Will be using TV for now. 1920x1080
What are you using it for?
Mainly gaming, Skyrim, Farcry 3 and Starcraft. Won't be streaming with it.
What is your upgrade cycle?
2 - 3 years.
When do you plan on building it?
Plan to build is sometime in the next month.
Do you plan on overclocking?
Possibly some slight overclocking. But nothing very high.
Do you need an operating system?
Nope. Already own one.
Do you plan to add a second GPU for SLI or Crossfire?
No.
Where are you buying your parts from?
All of it will be bought off of newegg.
You fine with GPU only overclocking? Overclocking with Intel is around ~$160 more than going with an i3 and there is no real need to truthfully. From the reviews I've read Farcry 3 is a graphically intense game, while an SSD makes Skryim much more enjoyable, so I'll base my recommendation around that.
i3-3220 $129.99
http://www.newegg.com/Product/Product.aspx?Item=N82E16819116775
MSI H77MA-G43 & Patriot IEM LE 8GB (2x4GB) DDR3-1600 1.5v c9 $112.98 ($5MIR)
http://www.newegg.com/Product/ComboDealDetails.aspx?ItemList=Combo.1183438
EVGA GTX 660 $219.99
http://www.newegg.com/Product/Product.aspx?Item=N82E16814130825
Samsung 840 120GB SSD $99.99
http://www.newegg.com/Product/Product.aspx?Item=N82E16820147188
NXZT Source 210 $31.99 w/promo code NZXTJAN20, ends 1/30
http://www.newegg.com/Product/Product.aspx?Item=N82E16811146075
Antec VP-450 $38.99
http://www.newegg.com/Product/Product.aspx?Item=N82E16817371045
DVD Burner $16.99
http://www.newegg.com/Product/Product.aspx?Item=N82E16827106289
Total: $650.92 ($645.92AMIR)
There is only 120GB of storage, if you deem that not enough, drop the SSD and just get a regular HDD instead. I believe 120GB is enough, I only use 65GB and I have 3 games installed (sc2,skyrim,GTA4), win7 and 4GB of music.
|
On January 30 2013 08:58 Craton wrote:
Some watercooling-related questions.
From what I read, a typical radiator can dissipate about 150W worth of heat per 120mm, such that a 3x120 rad can dissipate ~450W of heat.
Some research suggests that an overclocked 3770K + pair of 7970 Ghz Ed. would consume somewhere around 600-700 watts of power (on high load) and thus one should get 480+ worth of radiator.
What I can't seem to find is what kind of difference I would get from different permutations, meaning what kind of temperature difference is there between a single 360, a single 480, or a 360 + 240 on a single loop. Additionally, what, if any, difference would a push/pull setup make on the single 360?
Its just not that easy. Your expected delta T(and thus cooling performance) (water load temp - ambient temp) is going to dictate how much radiator you put in the loop.
A typical 360 radiator can cool 550w with a 10c Delta T with high speed fans.
Have you read this? It explains pretty much everything crystal clear.
http://www.tomshardware.com/forum/277130-29-read-first-watercooling-sticky
On January 30 2013 09:49 Myrmidon wrote:+ Show Spoiler +I'm not sure if those figures make that much sense. I don't exactly follow water cooling benches (at all), but are people really testing these things, or just guessing? You may be asking the wrong forums about this kind of thing, as I don't know if any of us have done or regularly bother with custom water; that said, sometimes experience and personal anecdotes get in the way of broader thinking and common sense, so maybe not. Asus ROG Ares II (dual 7970s) cools the two GPUs with an Asetek closed-loop system connected to a single 120mm radiator. VRMs and RAM are cooled by a separate fan, and the radiator is pretty thick (though doesn't look super-extraordinarily dense or anything like that), but again... single 120mm rad. Noise not that crazy: http://www.techpowerup.com/reviews/ASUS/ARES_II/26.htmlTemps actually pretty low: http://www.techpowerup.com/reviews/ASUS/ARES_II/32.htmlPower draw mostly in line with what you expect from those two GPUs: http://www.techpowerup.com/reviews/ASUS/ARES_II/25.htmlI don't think it should be much different conceptually than most air cooling. Think about the temperature differentials involved. The water is there to transfer heat from the CPU block to the radiators, where fans force air of a lower temperature through the fins. Obviously it depends on the radiator thickness, fin density, fan speeds, and more. This is assuming the pump and coolant are doing their jobs and making some more simplifications. Things hardly scale linearly, but as far as I know, you can do some rough estimation just thinking in terms of fin surface area. Push/pull is more important if the fans have lower static pressure (are worse, blades not optimal for the task, and/or lower fan speeds), meaning they have more trouble blowing through obstructions, and if the radiators are thicker and denser, meaning more restrictive. Same as with other kinds of heatsinks. If you have an overkill amount of radiator area, it probably shouldn't make much difference.
All Asetek Closed loop systems use Ethylene glycol as the coolant which is terrible compared to water. Comparing a closed loop system and custom loops is pretty bad. A single 120 Radiator can cool that ASUS card because ASUS is fine with the Delta T being redicioulously high (>20c) because it will still out perform air.
|
On January 30 2013 13:32 iTzSnypah wrote:
Have you read this? It explains pretty much everything crystal clear.
Yes, I've read that. I was able to find what I needed from a combination of other sources.
For those that are curious: yes you can cool it with a 120.3 and have good temps, but you will want a second radiator (something around a 120.2+) to get ideal performance.
Right now I'm leaning toward the Xigmatech Elysium case and am figuring out how I might fit a second radiator in it, but I might just go with a single 120.3 and see how everything pans out before worrying about a second one.
|
On January 30 2013 13:32 iTzSnypah wrote:Show nested quote +On January 30 2013 09:49 Myrmidon wrote:+ Show Spoiler +I'm not sure if those figures make that much sense. I don't exactly follow water cooling benches (at all), but are people really testing these things, or just guessing? You may be asking the wrong forums about this kind of thing, as I don't know if any of us have done or regularly bother with custom water; that said, sometimes experience and personal anecdotes get in the way of broader thinking and common sense, so maybe not. Asus ROG Ares II (dual 7970s) cools the two GPUs with an Asetek closed-loop system connected to a single 120mm radiator. VRMs and RAM are cooled by a separate fan, and the radiator is pretty thick (though doesn't look super-extraordinarily dense or anything like that), but again... single 120mm rad. Noise not that crazy: http://www.techpowerup.com/reviews/ASUS/ARES_II/26.htmlTemps actually pretty low: http://www.techpowerup.com/reviews/ASUS/ARES_II/32.htmlPower draw mostly in line with what you expect from those two GPUs: http://www.techpowerup.com/reviews/ASUS/ARES_II/25.htmlI don't think it should be much different conceptually than most air cooling. Think about the temperature differentials involved. The water is there to transfer heat from the CPU block to the radiators, where fans force air of a lower temperature through the fins. Obviously it depends on the radiator thickness, fin density, fan speeds, and more. This is assuming the pump and coolant are doing their jobs and making some more simplifications. Things hardly scale linearly, but as far as I know, you can do some rough estimation just thinking in terms of fin surface area. Push/pull is more important if the fans have lower static pressure (are worse, blades not optimal for the task, and/or lower fan speeds), meaning they have more trouble blowing through obstructions, and if the radiators are thicker and denser, meaning more restrictive. Same as with other kinds of heatsinks. If you have an overkill amount of radiator area, it probably shouldn't make much difference. All Asetek Closed loop systems use Ethylene glycol as the coolant which is terrible compared to water. Comparing a closed loop system and custom loops is pretty bad. A single 120 Radiator can cool that ASUS card because ASUS is fine with the Delta T being redicioulously high (>20c) because it will still out perform air.
So for a number of reasons, CLCs have compromised performance compared to custom cooling: lower specific heat of the coolant, less of it because no reservoir and less tubing, weaker pump, probably a whole bunch of things I forgot or never knew about, etc. That makes the point further clear, that people speccing out custom loops are not running anywhere near the heat exchange capacity of the systems.
So my questioning of one square 120mm unit (uh, with standard z-height? fpi?) per 150W as being arbitrary still stands? Or is there any justification given for that kind of figure?
Depends on a lot of things, including desired temperatures. Which may not relate much to desired (computational) performance. Again, priorities.
|
Calling it "arbitrary" is probably overdoing it; the figure seems to be fairly well accepted.
There are definitely a number of factors, but from the tests I've seen and other material I've been reading it seems that because each of the radiators and pump options have performance that's basically dead-even, and that tubing seems to almost always one of two sizes (which have a negligible difference in effectiveness), that it becomes possible to make this kind of "rule-of-thumb" statement.
You also have to factor in noise levels with regard to how loud you want to run your pump/fans, but I don't think that detracts any from this.
|
By arbitrary, I mean that nobody actually calculated this value or extensively tested a wide variety of loads and radiators setups to come up with 150W as opposed to 175W or 200W or 125W or whatever else, right? Or is there a rigorous justification that I don't know about? Maybe that wasn't the right choice of words.
|
There are definitely some very rigorous studies out there regarding water cooling and the many related variables. I don't know which if any comes up with this figure, but I would be surprised if there isn't one saying something along those lines.
|
On January 30 2013 15:32 Myrmidon wrote:
By arbitrary, I mean that nobody actually calculated this value or extensively tested a wide variety of loads and radiators setups to come up with 150W as opposed to 175W or 200W or 125W or whatever else, right? Or is there a rigorous justification that I don't know about? Maybe that wasn't the right choice of words.
http://skinneelabs.com/water-cooling-radiators/
Extensively. It all comes down to what you want your delta T to be.
With a 1000rpm yate loon fan the RX360v2 can dissipate 362w at 10c delta T (10c is just meh). While at 2c delta T it can only dissipate 72w.
![[image loading]](http://skinneelabs.com/assets/images/Radiators/XSPC/RX360v2/RX360-v2_10cDelta_Chart.png)
![[image loading]](http://skinneelabs.com/assets/images/Radiators/XSPC/RX360v2/RX360-v2_2cDelta_Chart.png)
|
On January 30 2013 13:11 iTzSnypah wrote:OMG OMG OMG IVB CELERONS HAVE ARRIVED!@#!@!@#!@!@#!@ http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007671 600315409&IsNodeId=1&name=Ivy BridgeShow nested quote +On January 30 2013 09:51 Turtle104 wrote:+ Show Spoiler +
What is your budget?
600$ - 650$
What is your resolution?
Will be using TV for now. 1920x1080
What are you using it for?
Mainly gaming, Skyrim, Farcry 3 and Starcraft. Won't be streaming with it.
What is your upgrade cycle?
2 - 3 years.
When do you plan on building it?
Plan to build is sometime in the next month.
Do you plan on overclocking?
Possibly some slight overclocking. But nothing very high.
Do you need an operating system?
Nope. Already own one.
Do you plan to add a second GPU for SLI or Crossfire?
No.
Where are you buying your parts from?
All of it will be bought off of newegg.
You fine with GPU only overclocking? Overclocking with Intel is around ~$160 more than going with an i3 and there is no real need to truthfully. From the reviews I've read Farcry 3 is a graphically intense game, while an SSD makes Skryim much more enjoyable, so I'll base my recommendation around that. i3-3220 $129.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16819116775MSI H77MA-G43 & Patriot IEM LE 8GB (2x4GB) DDR3-1600 1.5v c9 $112.98 ($5MIR) http://www.newegg.com/Product/ComboDealDetails.aspx?ItemList=Combo.1183438EVGA GTX 660 $219.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16814130825Samsung 840 120GB SSD $99.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16820147188NXZT Source 210 $31.99 w/promo code NZXTJAN20, ends 1/30 http://www.newegg.com/Product/Product.aspx?Item=N82E16811146075Antec VP-450 $38.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16817371045DVD Burner $16.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16827106289Total: $650.92 ($645.92AMIR) There is only 120GB of storage, if you deem that not enough, drop the SSD and just get a regular HDD instead. I believe 120GB is enough, I only use 65GB and I have 3 games installed (sc2,skyrim,GTA4), win7 and 4GB of music.
Would it be better to go with AMD because the prices are about the same? And would the i3 be bottlenecking the GPU?
|
On January 31 2013 10:26 Turtle104 wrote:Show nested quote +On January 30 2013 13:11 iTzSnypah wrote:OMG OMG OMG IVB CELERONS HAVE ARRIVED!@#!@!@#!@!@#!@ http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007671 600315409&IsNodeId=1&name=Ivy BridgeOn January 30 2013 09:51 Turtle104 wrote:+ Show Spoiler +
What is your budget?
600$ - 650$
What is your resolution?
Will be using TV for now. 1920x1080
What are you using it for?
Mainly gaming, Skyrim, Farcry 3 and Starcraft. Won't be streaming with it.
What is your upgrade cycle?
2 - 3 years.
When do you plan on building it?
Plan to build is sometime in the next month.
Do you plan on overclocking?
Possibly some slight overclocking. But nothing very high.
Do you need an operating system?
Nope. Already own one.
Do you plan to add a second GPU for SLI or Crossfire?
No.
Where are you buying your parts from?
All of it will be bought off of newegg.
You fine with GPU only overclocking? Overclocking with Intel is around ~$160 more than going with an i3 and there is no real need to truthfully. From the reviews I've read Farcry 3 is a graphically intense game, while an SSD makes Skryim much more enjoyable, so I'll base my recommendation around that. i3-3220 $129.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16819116775MSI H77MA-G43 & Patriot IEM LE 8GB (2x4GB) DDR3-1600 1.5v c9 $112.98 ($5MIR) http://www.newegg.com/Product/ComboDealDetails.aspx?ItemList=Combo.1183438EVGA GTX 660 $219.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16814130825Samsung 840 120GB SSD $99.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16820147188NXZT Source 210 $31.99 w/promo code NZXTJAN20, ends 1/30 http://www.newegg.com/Product/Product.aspx?Item=N82E16811146075Antec VP-450 $38.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16817371045DVD Burner $16.99 http://www.newegg.com/Product/Product.aspx?Item=N82E16827106289Total: $650.92 ($645.92AMIR) There is only 120GB of storage, if you deem that not enough, drop the SSD and just get a regular HDD instead. I believe 120GB is enough, I only use 65GB and I have 3 games installed (sc2,skyrim,GTA4), win7 and 4GB of music. Would it be better to go with AMD because the prices are about the same? And would the i3 be bottlenecking the GPU?
AMD is almost never the better choice. So no, keep the i3.
|
On January 30 2013 16:48 iTzSnypah wrote:Show nested quote +On January 30 2013 15:32 Myrmidon wrote:
By arbitrary, I mean that nobody actually calculated this value or extensively tested a wide variety of loads and radiators setups to come up with 150W as opposed to 175W or 200W or 125W or whatever else, right? Or is there a rigorous justification that I don't know about? Maybe that wasn't the right choice of words. Extensively. It all comes down to what you want your delta T to be.
Of course, but the idea of a "rule of thumb" is for your typical case, not when you've already got a significant amount of dissipation and are already well in the realm of diminishing returns.
|
For anyone thinking of building, ASRock z77 Extreme4 is $74.99 at Micro Center this month when you buy any i5 processor. You can actually get a 2500k and this mobo for under $200 for you overclocking enthusiasts. Also, $20 off any Samsung SSD as well when you buy all 3 (at least at my MC).
|
On January 31 2013 11:31 upperbound wrote:
For anyone thinking of building, ASRock z77 Extreme4 is $74.99 at Micro Center this month when you buy any i5 processor. You can actually get a 2500k and this mobo for under $200 for you overclocking enthusiasts. Also, $20 off any Samsung SSD as well when you buy all 3 (at least at my MC).
I'm assuming this is for Feb and not Jan :p
I don't see the deal up yet, I'm at the Virginia Microcenter.
|
On February 01 2013 00:14 Torenhire wrote:Show nested quote +On January 31 2013 11:31 upperbound wrote:
For anyone thinking of building, ASRock z77 Extreme4 is $74.99 at Micro Center this month when you buy any i5 processor. You can actually get a 2500k and this mobo for under $200 for you overclocking enthusiasts. Also, $20 off any Samsung SSD as well when you buy all 3 (at least at my MC). I'm assuming this is for Feb and not Jan :p I don't see the deal up yet, I'm at the Virginia Microcenter.
Yes, it's for February, but the MA Micro Center started them on Monday.
|
Hm, I will keep an eye out, would make my build I'm planning a little easier. :D
Thanks for the heads up!
|
Ya I got my i5-3570k, extreme4, and samsung 840 120g ssd, for $189.99, $89.99(plus $10 MIR), and $75.99 from MC ...samsung 840 had a bad sign but manager said price was valid for me...
Question from previous build-post: My build so far is i5-3750k, asrock extreme4, and evga SC gtx660. Want to make sure my Antec ea-430W green 80+bronze atx12v v2.3 PSU can handle overclocking? Nothing too hardcore. If I should man-up and not be a cheap-ass (was gonna wait for SLI to upgrade PSU b/c am going to keep this one) then let me know. Haven't bought a CPU cooler yet so want to make sure.
edit:1 HDD, 1 SSD, no optical drive as of now
edit2: the newegg description isn't exactly right for mine. My box has the +12V1 @ 25A and +12V2 @ 25A.
|
Your Antec 430w will be more than fine. There's no way you're going to approach its limits without adding a second GPU.
|
|
|
|
|
|
|
|
EPT
