LN2Cooling.com Evaporator Under Dry Ice

Add Your Comments

I know some of you saw my recent dry ice runs on my Phenom II rig. In a few forum threads I discussed that for my next run I will have a different evaporator (pot) to play around with – and I assure you that promise was kept! Here you’ll see me take my Phenom II X6 1090T @ 5.625GHz with Dry Ice and this great pot:

I got in touch with a friend and he agreed to let me take one of his cooling pots for a spin. Aaron Schradin of LN2Cooling.com has lent me an excellent pot to play around with and review for all of you guys. This pot is usually meant for the purpose of running under LN2 and LHe4. This is seen in the many AMD world record breaking overclocking attempts, which can all be seen over at amdblackops.com. However I will be playing around with the dry ice performance of this pot, which has mostly been overlooked until now.

The new little toy in town is in fact not little at all. This pot is huge. Much larger than the previous aluminum cooling pot I have been benchmarking with. There is also a lot of mass to work with which in turn equals to better cooling (more stable temperatures). Combine that mass with high-quality copper and lots of surface area and you have yourself an excellent cooling pot.

The Pot

Pot Unboxing

Pot Unboxing

I was interested in the packing, a shipping tube made specifically to ensure this evaporator can be safely shipped anywhere.

Pot And Mounting Hardware

Pot And Mounting Hardware

Here’s a look at whats was included with my package. The hardware, mounting kit, and of course the copper base and aluminum reservoir.

The Base

The Base

Here is a closeup of the base. A little tarnished, but nothing a little bit of polish can’t clean up easily. Usually tarnished copper can be cleaned with ketchup as well – ketchup is mildly acidic which makes it well fit for the task as well as being readily available.

Probe Hole

Probe Hole

This pot comes with a drilled in temperature probe hole for to ensure the pot is cooling to maximum capacity.

Mirror Base And Hardware

Mirror Base And Hardware

Here is the base cleaned up. A perfect mirror. It also shows the mounting hardware close up. These springs tightened up will ensure just the right amount of pressure bearing down on the cpu.

Mounting Kit

Mounting Kit

The mounting kit comes with the ability to use this pot with all current sockets on Intel and AMD platforms.

Threading

Threading

The pot threading. It was important not to tighten the reservoir too much when screwing it on to the base or else it will become extremely difficult to remove.

Height

Height

The pot with both parts screwed together is around 16cm tall. That is about 6.3 inches.

Diameter

Diameter

The diameter of this monster is approximately 8.5cm. Once again that is about 3.3 inches.

Internals, Lots of Surface Area

Internals, Lots of Surface Area (Courtesy LN2Cooling.com)

Test System Setup

  • AMD Phenom II X6 1090T – Batch 1018EPAW
  • Gigabyte 890FXA-UD5 Rev 2.o
  • Mushkin Blackline 4GB DDR3 – Stock @ 1600MHz 6-8-6-24
  • Sapphire Radeon 5870 1GB GDDR5 – Overclocked @ 900MHz/1200MHz
  • PC Power And Cooling 750W PSU
  • LN2cooling.com LN2 Evaporator Cooling Pot
  • 20 Pounds Of Dry Ice
  • Plumbers Putty
  • Dielectric Grease

Setup Before The Run

Setup Before The Run

Shot Of The 1090T

Shot Of The 1090T

Crushed DICE And Chopstick

Crushed DICE And Chopstick

During The Run

During The Run

Slushie And Bubbles Mixture

Slushie And Bubbles Mixture

Frosty

Frosty

Frost Pattern

Frost Pattern

End Of Run

End Of Run

In the pics above, you can see the pot is filled with crushed DICE. I used a chopstick to stir the mixture. Following that you can see the setup during the run, the slushie Dry Ice mixture, the pot frosted over, and then a pretty cool frost pattern that formed on the pot. Finally at the end of the run, with the pot removed… Snow all around.

The Results

Benchmarks:

  • CPU-Z
  • SuperPi 1m
  • SuperPi 32m
  • wPrime 32m
  • wPrime 1024m
  • PiFast
  • 3DMark 06

For the results, I will be comparing the scores on this pot to the scores from my last run on my Duniek Aluminum pot.

CPU-Z  –  Core #1 5.579GHz @ 1.744VcoreValidation

Validation

Validation

For this CPU-Z validation I went all out overclocking my chip’s strongest core, core #1. As you can see I disabled all but 2 cores, and ran down a bunch of other settings to get this. Before settling on 5.579GHz, I was able to hit 5.624GHz but was not able to validate without crashing. Maybe I will make this on my next run. My previous run on the aluminum pot only netted me a max validation of 5.547GHz.

CPU-Z  –  All 6 Cores, 5.500GHz @ 1.696Vcore w/ 3000NB and 1000MHz 7-9-6-15 memory

5500MHz All 6 Cores

5500MHz All 6 Cores

During my last run, I was only able to push 5.4GHz all 6 cores. This goes to show how the increased mass and the copper material of this pot makes for more proficient cooling and coping with the load of all 6 cores pushed to 5.5GHz.

SuperPi 1m  –  12.625 seconds

SuperPi 1m

SuperPi 1m

My old time on the aluminum pot was 12.797s. SuperPi 1m is highly inefficient on Gigabyte boards for whatever reason in my personal experience (Editor’s Note: There isn’t much supporting evidence of this claim and suggest default RAM timings play more a factor than manufacturer, though Madshrimps did find a few boards performed slightly better in 1M in their recent testing. But also, the World Record for SPI 1M is on a Gigabyte board, thanks to Deanzo for the tip on that one), but I still managed to surpass my old score and net the top 1m DICE score for this processor on hwbot.org.

SuperPi 32m  –  12 minutes 58.219 seconds

SuperPi 32m

SuperPi 32m

As a result of this pot being able to hold down better temps under load, I was able to increase cpu speed in contrast to my previous run. I broke the 13 minute barrier for 32m on DICE with this processor.

wPrime 32m  –  4.765 seconds

wPrime 32m

wPrime 32m

Not much to say here. I’m thinking I could have fine tuned my setup for a higher score, but I settled at this time. For my next run I will get a faster time.

wPrime 1024m  –  151.25 seconds

wPrime 1024m

wPrime 1024m

I was astonished at the pots ability to handle the load of all 6 cores stressed. All extreme benchers know that 1024 takes a bit of work to pass at higher frequencies. I managed to pass at 5.125GHz. I was unable to yield any results over 5Ghz with the previous pot.

PiFast  –  20.50 seconds

PiFast

PiFast

Once again, another untuned run. However I surpassed all of my personal bests with this one.

3DMark 06  –  28306 3DMarks

3DMark06

3DMark06

For the first time in a long time, I decided to test out some 3D stuff again. I did this mainly to gauge how the evaporator handles 3D load. This is where I saw my biggest improvement from my old pot. During one of my 955 X4 BE runs, I found myself only able to run 06 at 4.520Ghz using the aluminum pot. I was able to push out and run 5.25GHz stable under the load of 3DMark06.

That is all for the results. However, here are the improvements I made with the new pot summarized into one chart:

Chart

Chart

Conclusion

This is an excellent pot for all types of extreme cooling. Remember that this pot is not primarily for DICE. It begins to shine once under LN2 and LHe. The 99% pure copper design, combined with the excellent internal surface area and the mass of the base makes for extremely proficient cooling. The install is pretty straightforward as well.

This pot helped me break all my personal best’s that I made on my previous aluminum pot. Half of these runs still have headroom to tune as this pot is very efficient. In the BIOS it read -75° C at almost all times, which is only 4° C above the actual temp of dry ice itself. This is next to impeccable cooling performance. Any extreme bencher would love this evaporator.

I would like to thank Aaron Schradin over at LN2cooling.com for lending me this pot to play with.

-Matthew Sembinelli AKA Slappa

My HWBot.org Profile

My Personal Blog

Disclosure: The author of this article is a friend of ln2cooling.com who are the makers of this evaporator, and has a subdomain on their website.

Leave a Reply

Your email address will not be published. Required fields are marked *

Discussion
  1. Yes, Intel chips are geared for better superpi performance:

    http://www.overclockers.com/forums/showthread.php?t=646030#SuperPi

    http://hwbot.org/rankings/worldrecords#SuperPi
    Great job. I would love to play with DICE someday. I do have a question though. Does SuperPi seriously favor Intel chips? I only have a Phenom 2 x2 (wife's rig) to play with so I can't make a fair comparison to my i5 in house. I only ask because it seems my i5 @ 3.8 is faster than the X6 @ 5.5 for the 1M test, I haven't run the 32M test yet.
    Very nice article and performance. The manufacturers website is disappointing though, most of the products are not is stock, very little info and hardly any pictures.