Overclock.net › Forums › Cooling › Air Cooling › CRYORIG Fan & Cooler Club; R1 Ult, R1 Uni, C1, H5, XF140 & XT140 and all things Cryorig
New Posts  All Forums:Forum Nav:

CRYORIG Fan & Cooler Club; R1 Ult, R1 Uni, C1, H5, XF140 & XT140 and all things Cryorig - Page 186

post #1851 of 3698
I have been wondering a lot about a good test methodology performed on 2011-3 as opposed to 1150 or 1151.

Lets say you have a 4790K. It's cores are concentrated in the center. Thus you get high thermal density in the CPU. Those CPU cores will get limited contact with heat-pipes from an Air cooler. A CLC though, will not have such issue, since the liquid inside is not as restricted area wise as a heat-pipe.

On X99 however, you get 6-8 cores very well distributed around. The CPU die is also larger. Thus giving a lot more area in which the heat-pipes can draw heat. This will not majorly impact a CLC, since liquid is not as affected as a heat-pipe location wise.

So - my logic is - on LGA 2011 / 2011-3 - R1 Ultimate, Dark Rock Pro3, NH-D14/15, Silver Arrow IB/SB and their E edditions, TC14PE and etc, will perform better on the Enthusiast socket than they do on the mainstream socket - which sadly is what 99% of the reviewers use.

P.S.
Sadly, nobody will sponsor me with a X99 and any CPU that fits in that socket to test it myself. tongue.gif If someone has a spare - I can use a new X99 system biggrin.gif
Edited by Shneiky - 11/7/15 at 10:21am
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
post #1852 of 3698
I use a 5820k, or a 2600k. I avoid haswells and Ivy's as not many people delid, and the difference in temps even between undelided ones are insane. Its down to luck how thick the silicone glue is on your specific processor..

So very much useless to test. And testing on a delided processor well, that would yield many more issues. Like risk of killing it from constant mount/demount of coolers, people comparing to their processors which are not delided and calling results fake etc.

And you are right, CR1 does even better as it has more heatpipes so they catch the outskirts which are way cooler on consumer platform than on HEDT. smile.gif
post #1853 of 3698
Thread Starter 
There is validity to solme of your logic. the size area of chips on the CPU PCB can and does make a differnce, but size of the IHS has no or almost no effect on cooling. Except for direct contact heatpipe bases and aluminum bases, the solid copper of the base transfers heat over the area of the heatpipes very efficiently ..

But the thermal conductivity of water versus copper is extremely different, even if it is moving constantly over the surface.
Copper is 385.0 W/m K
Aluminum 205.0 W/m K
Water . is . 00.6 W/m K
*Higher is better* biggrin.gif
Heatpipes are very efficient. thumb.gif
post #1854 of 3698
Thread Starter 
Quote:
Originally Posted by Kutalion View Post

I use a 5820k, or a 2600k. I avoid haswells and Ivy's as not many people delid, and the difference in temps even between undelided ones are insane. Its down to luck how thick the silicone glue is on your specific processor..

So very much useless to test. And testing on a delided processor well, that would yield many more issues. Like risk of killing it from constant mount/demount of coolers, people comparing to their processors which are not delided and calling results fake etc.

And you are right, CR1 does even better as it has more heatpipes so they catch the outskirts which are way cooler on consumer platform than on HEDT. smile.gif
I am going through the dilemma with my new 6700K. Delidding will give better heat transfer to IHS and therefor to cooler/s being used on it .. therefor being a much better heat source for testing. But I can't afford to replace it if it dies. frown.gif Which means I will probably keep the serious cooking of coolers to my old i7 920.

At least on a delidded CPU (or i7-920) the heat source is being directly moved to cooler base.

Honestly, even same model of overclocked CPUs using same cooler should not be compared. But a series of accurate tests of several different coolers done on one CPU, then another CPU give us data we can compare. While their temps will be different there should be an accurate correlation between each CPU using same cooler from lows to highs.

Add in the lack of monitoring or accurate monitoring of cooler intake air temp during testing and most test results have little meaning.

One of best testing setups I seen is for comparing cooler performance is "Project Jalapeno"
http://www.thelab.gr/topic/88302-codename-project-jalapeno/
Edited by doyll - 11/7/15 at 11:14am
post #1855 of 3698
@Kutalion,

Yes, I am also using a 2700K and finding very consistent results. Sadly enough, I was in a bachelor programme in Computer Sciences and now I am in my last year of Art&Technology bachelor (specializing in VFX) so I am barely making ends meet and getting new hardware is completely our of the question. biggrin.gif Being a bachelor for 7-8 years is crewed-up. tongue.gif

@Doyll,

I completely agree with you. Metals have way higher thermal conductivity than water.

But - and this is a big "BUT", as we saw on Ivy, Haswell and Skylake as opposed to the Ivy-E and Haswell-E - the best coolers for the mainstream platform is not the one that can deal with the largest amount of heat - but the ones that can take heat away faster. Or in other words - the consumer platform, which is chips with high thermal density, although overall - lower heat volume - benefit more from the said CLCs. The CLCs do not actually use plain water, but a coolant liquid, that can absorb smaller amounts of heat very rapidly.

The main point I am trying to make here is - the liquid in CLCs is, of course , a liquid - a matter which molecules are loosely arranged (with no regular arrangement) - thus small amounts of heat can travel fast. (same is applicable to vapor fluid)

In a metal - the molecules are densely packed and neatly organized. This makes large quantities of heat travel very efficiently, but at a slower pace.

So we have a LGA 115* (something from Ivy to Skylake) which are much more susceptible to the speed from which you can take out the heat.
And we have LGA 2011 / 2011-3 which is much more dependent on the overall volume of heat you can get out.

Please, feel free to fix my physics if I screwed up anywhere tongue.gif

This brings me to the original point - large air coolers kick even more ass on HEDT platforms thumb.gif

All of this is in the logic that most reviewers actually use the consumer platform for benches - which gives the CLCs an advantage.
Edited by Shneiky - 11/7/15 at 11:14am
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
post #1856 of 3698
I get cancer every time i see somebody use a consumer i7 Ivy/Haswell for benches.
post #1857 of 3698
Quote:
Originally Posted by Kutalion View Post

I get cancer every time i see somebody use a consumer i7 Ivy/Haswell for benches.

Feel the same bro

I'm going to order a NHD15 end year, and switch to open air bench probably
紅魔館
(12 items)
 
秋山 澪
(11 items)
 
For Sale: Crucial 2x8GB DDR4 2400 / 1TB 2.5" HDD
$40 (USD) or best offer
CPUMotherboardGraphicsRAM
i7 5820K ASUS X99 Deluxe Zotac GTX 1080 AMP! Crucial Ballistix Sport 4x4G DDR4-2400 
Hard DriveHard DriveOptical DriveCooling
WD Black 4TB Samsung 850 EVO 1TB Dell CH30N BD-Read Noctua NH-D15 
CoolingOSPowerCase
Noctua NF-A14 PWM x6 Windows 10 Enterprise EVGA 750W G2 Phanteks Enthoo Luxe 
CPUMotherboardGraphicsGraphics
i7 7700K P870KM1 Nvidia GeForce GTX 1080 Nvidia GeForce GTX 1080 
RAMHard DriveHard DriveOS
Gskill 3000MHz 16GB F4-3000C16D-16GRS  Samsung 850 EVO 500GB Samsung 850 Pro 500GB Windows 10 Enterprise 
MonitorPowerCase
AUO B173QTN01.0 2560x1440p G-sync  Eurocom 780W Clevo P870KM1 
  hide details  
Reply
紅魔館
(12 items)
 
秋山 澪
(11 items)
 
For Sale: Crucial 2x8GB DDR4 2400 / 1TB 2.5" HDD
$40 (USD) or best offer
CPUMotherboardGraphicsRAM
i7 5820K ASUS X99 Deluxe Zotac GTX 1080 AMP! Crucial Ballistix Sport 4x4G DDR4-2400 
Hard DriveHard DriveOptical DriveCooling
WD Black 4TB Samsung 850 EVO 1TB Dell CH30N BD-Read Noctua NH-D15 
CoolingOSPowerCase
Noctua NF-A14 PWM x6 Windows 10 Enterprise EVGA 750W G2 Phanteks Enthoo Luxe 
CPUMotherboardGraphicsGraphics
i7 7700K P870KM1 Nvidia GeForce GTX 1080 Nvidia GeForce GTX 1080 
RAMHard DriveHard DriveOS
Gskill 3000MHz 16GB F4-3000C16D-16GRS  Samsung 850 EVO 500GB Samsung 850 Pro 500GB Windows 10 Enterprise 
MonitorPowerCase
AUO B173QTN01.0 2560x1440p G-sync  Eurocom 780W Clevo P870KM1 
  hide details  
Reply
post #1858 of 3698
Thread Starter 
Quote:
Originally Posted by Shneiky View Post

@Kutalion,

Yes, I am also using a 2700K and finding very consistent results. Sadly enough, I was in a bachelor programme in Computer Sciences and now I am in my last year of Art&Technology bachelor (specializing in VFX) so I am barely making ends meet and getting new hardware is completely our of the question. biggrin.gif Being a bachelor for 7-8 years is crewed-up. tongue.gif

@Doyll,

I completely agree with you. Metals have way higher thermal conductivity than water.

But - and this is a big "BUT", as we saw on Ivy, Haswell and Skylake as opposed to the Ivy-E and Haswell-E - the best coolers for the mainstream platform is not the one that can deal with the largest amount of heat - but the ones that can take heat away faster. Or in other words - the consumer platform, which is chips with high thermal density, although overall - lower heat volume - benefit more from the said CLCs. The CLCs do not actually use plain water, but a coolant liquid, that can absorb smaller amounts of heat very rapidly.

The main point I am trying to make here is - the liquid in CLCs is, of course , a liquid - a matter which molecules are loosely arranged (with no regular arrangement) - thus small amounts of heat can travel fast. (same is applicable to vapor fluid)

In a metal - the molecules are densely packed and neatly organized. This makes large quantities of heat travel very efficiently, but at a slower pace.

So we have a LGA 115* (something from Ivy to Skylake) which are much more susceptible to the speed from which you can take out the heat.
And we have LGA 2011 / 2011-3 which is much more dependent on the overall volume of heat you can get out.

Please, feel free to fix my physics if I screwed up anywhere tongue.gif

This brings me to the original point - large air coolers kick even more ass on HEDT platforms thumb.gif

All of this is in the logic that most reviewers actually use the consumer platform for benches - which gives the CLCs an advantage.
AFAIK the liquid in CLCs is the same as AIOs and loops (and heatpipes as well); H2O with maybe a small amount of inhibitor added to stop corrosion.

I have not had to privilege of using any of the newer H2O loops, AIOs or CLCs so can't say how well heatpipe bases and waterblocks compare under equal test conditions. I do know some heatpipe bases work better on some CPUs better than others .. and heatpipe direction / orientation can also effect how well heat is transferred out of CPU.

Problem is the same as other testing. Procedures are not nearly accurate enough for much if any meaningful analysis' to be made.
post #1859 of 3698
Quote:
Originally Posted by doyll View Post

AFAIK the liquid in CLCs is the same as AIOs and loops (and heatpipes as well); H2O with maybe a small amount of inhibitor added to stop corrosion.

I have not had to privilege of using any of the newer H2O loops, AIOs or CLCs so can't say how well heatpipe bases and waterblocks compare under equal test conditions. I do know some heatpipe bases work better on some CPUs better than others .. and heatpipe direction / orientation can also effect how well heat is transferred out of CPU.

Problem is the same as other testing. Procedures are not nearly accurate enough for much if any meaningful analysis' to be made.

Hey doyll, what case do you recommend if I were to use the NH D15? It has to house 2 5.25 and a minimum of 4 hdd and a space for 120mm clc for the graphics card.

Thanks in advance.
紅魔館
(12 items)
 
秋山 澪
(11 items)
 
For Sale: Crucial 2x8GB DDR4 2400 / 1TB 2.5" HDD
$40 (USD) or best offer
CPUMotherboardGraphicsRAM
i7 5820K ASUS X99 Deluxe Zotac GTX 1080 AMP! Crucial Ballistix Sport 4x4G DDR4-2400 
Hard DriveHard DriveOptical DriveCooling
WD Black 4TB Samsung 850 EVO 1TB Dell CH30N BD-Read Noctua NH-D15 
CoolingOSPowerCase
Noctua NF-A14 PWM x6 Windows 10 Enterprise EVGA 750W G2 Phanteks Enthoo Luxe 
CPUMotherboardGraphicsGraphics
i7 7700K P870KM1 Nvidia GeForce GTX 1080 Nvidia GeForce GTX 1080 
RAMHard DriveHard DriveOS
Gskill 3000MHz 16GB F4-3000C16D-16GRS  Samsung 850 EVO 500GB Samsung 850 Pro 500GB Windows 10 Enterprise 
MonitorPowerCase
AUO B173QTN01.0 2560x1440p G-sync  Eurocom 780W Clevo P870KM1 
  hide details  
Reply
紅魔館
(12 items)
 
秋山 澪
(11 items)
 
For Sale: Crucial 2x8GB DDR4 2400 / 1TB 2.5" HDD
$40 (USD) or best offer
CPUMotherboardGraphicsRAM
i7 5820K ASUS X99 Deluxe Zotac GTX 1080 AMP! Crucial Ballistix Sport 4x4G DDR4-2400 
Hard DriveHard DriveOptical DriveCooling
WD Black 4TB Samsung 850 EVO 1TB Dell CH30N BD-Read Noctua NH-D15 
CoolingOSPowerCase
Noctua NF-A14 PWM x6 Windows 10 Enterprise EVGA 750W G2 Phanteks Enthoo Luxe 
CPUMotherboardGraphicsGraphics
i7 7700K P870KM1 Nvidia GeForce GTX 1080 Nvidia GeForce GTX 1080 
RAMHard DriveHard DriveOS
Gskill 3000MHz 16GB F4-3000C16D-16GRS  Samsung 850 EVO 500GB Samsung 850 Pro 500GB Windows 10 Enterprise 
MonitorPowerCase
AUO B173QTN01.0 2560x1440p G-sync  Eurocom 780W Clevo P870KM1 
  hide details  
Reply
post #1860 of 3698
@doyll,

You might be right about the substance used in CLCs and AIOs, but heat-pipes work differently.

What I know - at least in Cooler Master heat-pipes and in Sapphire heat-pipes for their Vapor X GPU coolers - they use either a liquid with very low boiling points (30-40-50 C - don't really know, it is an engineering secret in the end) or they use a gas with very high liquefying temp (again same 40-50C, secret again).

Heat-pipes will not be efficient at all if they use plain water. The use of a gas with high liquefying temp or a liquid with low boiling temp provides the efficiency we see today. I saw same engineering documents some years ago - cant really remember where and when. Only Cooler Master and Sapphire were that proud of their exact heat-pipe technology to showoff some of their engineering.

And you are definately right about testing procedures. The ultimate cooler performance procedure from Frozen CPU using a hotplate is the best procedure to get the ultimate performance of a cooler. But that is completely off when compared to a practical application with a certain CPU.
Edited by Shneiky - 11/7/15 at 11:45am
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
Work Horse
(8 items)
 
  
CPUMotherboardGraphicsRAM
2700K Z77 Extreme 6 Asus GTX 650 TI 2 GB Geil Black Dragon 16 GB 
Hard DriveCoolingPowerCase
Samsung 840 Evo Cryorig R1 Ultimate be quiet! L7 500W H440 New Edition (Solid Side Panel) 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Air Cooling
Overclock.net › Forums › Cooling › Air Cooling › CRYORIG Fan & Cooler Club; R1 Ult, R1 Uni, C1, H5, XF140 & XT140 and all things Cryorig