If you are looking for supercomputer performance at reasonable prices, you have come to the right place. Liquid Nitrogen Overclocking sells units that are more than twice as fast as the top-of-the-line systems that are available today. We do this by overclocking hardware, testing it rigorously, and documenting the performance of each system. If you want a system that is faster than anything that could be created by the mainstream manufacturers for the next several years, this is the place to be!

One thing we did was to "search around" online for information about overclocking. After a bit of googling, we were immediately left with two distinct impressions:

1. There are a great many discussion boards loaded with technical discussion, more acronyms than you can "shake a stick at", and many overclocking clubs that have some bizzare benchmarking competitions.

2. There is literally nobody selling units that are stable and fully built!

Most of the discussion boards you visit (some of which are exclusive and do not allow people entry without a referal from an existing member) are loaded with metrics that demonstrate remarkable "bursts" of speed achieved by clock frequencies we can only dream of at the present time.

What the discussion board owners don't tell you: These incredible speeds were very fleeting, sometimes lasting only as long as the benchmark application was recording the results!

We posed the following couple of questions on perhaps the most visited worldwide overclocking discussion board in February, 2010:

The replies to this were a bit shocking. Not only were these short bursts of speed "acceptable", the mentality was that "anything else" was a waste! Imagine that! A computer being used only for a benchmark that ran for seconds or minutes! Didn't anyone think that maybe overclcoking should be aimed at providing 24x7 high performance that was reliable?





Clearly they were being at least partly humorous, indicating how much they enjoyed their hobby. But, as we had already discovered from my foray at a previous discussion board, these competitions to "outgun" each other are a rather heated (no pun intended) pursuit. We were amused that our perception of the "practical side" of overclocking was not being fully embraced. In fact, it seemed there was little interest in building a system that could be used for any serious length of time.

It was not easy to find overclocking enthusuiasts who were receptive to the idea of running a benchmark for several hours. Apparently some of the systems they were using were basically dismantled CPUs sitting on a workbench, spread out to allow for ease of "changing cooling methods" as various tests were being performed.



You can see from the image above, Buckeye is cranking the core down from -36.5 Celcius on its way to -196 Celcius. Compare the condition of the wingnuts at the top of the insulated tube in this picture with the frosted liquid nitrogen run shown below.



The first overclocker who seemed to have more than a moderate interest in the checkers db6 benchmark that I created was the intrepid rickss69 from the Overlockaholics discussion board. Rick and I first started chatting when he showed me a picture of his "ice cold" liquid water cooling "solution". I peered at the images he sent me. It looked to be no more than a bucket of icewater on the desk with some sort of suction hose drawing the fluid out into some kind of chamber. He told me he had to "fill the bucket" up several times during my "long" benchmark run!



Our feelings were mixed upon seeing the picture above. On the one hand, we were happy to see such dedication from someone that surely must have known we were new to this (in early 2010). On the other hand, it seemed like there might not be a practical way to sustain high clock speeds for any reasonable duration. This latter thought was a bit depressing, since we were really hoping that "breaking 4 GHz" might be possible.

The first person to surpass what was in my mind a "magic number" was Justin Ashburn. When he sent in his first benchmark result of 5 hours 30 minutes (recall my fastest system took 8 hours 57 minutes) I was amazed. Then Justin told me that the result was just from his "normal" 3 GHz Intel Core 2 Quad Q 9650 chip, and again I was floored. He ran the benchmark later with his air-cooled system, and reported an overclock speed of 4.05 GHz for the duration of his 4 hour 11 minute benchmark run.

I had to re-read his results. 4 hours and 11 minutes? Was it possible for his system's performance to be better than double the speed of my 2.4 GHz AMD Opteron? His 4.05 GHz was "better than" 4.8 GHz on my AMD scale. In fact, if I do the math properly, it would be closer to 5.14 GHz on my machine's architecture! Wow! More importantly, he achieved better than 4 GHz, a number I thought was too unattainable for a long run, yet his system ran at this clip for hours. Success!

While I was still dealing with being stunned at this new "#1 result" for the benchmark run, Buckeye turned in an even more amazing result. His 5 GHz AMD 965BE C3 Deneb with the cascade cooling solution broke 4 hours with time to spare. In fact, he beat Justin's time by nearly half an hour! Was this possible? Then Buckeye beat his own time with an amazing 3 hour and 36 minute result, and it dawned on me for the first time: long term overclocking stability is not just a possibility, it is a reality!

At about this same time, Kristoffer from Norway shot to the top of the charts with a time a little faster than Buckeye's, albeit with a slightly slower 4.39 GHz overclock on the better Intel Core i7-920 architecture. Then gold ole' Rick showed up, and took top honors with not just one setup, but 4 amazing configurations that each ran in 3 hours 30 minutes or faster!

Rick took this one step further in two significant areas. He created systems capable of 4.9 GHz or better for prolonged periods of time, and he built systems that were visually appealing, integrated into a manageable chasis, and were completely self contained. This met all of my "wish list" criteria. These were boxes that were easy enough for any end user to appreciate and use, and, more importantly, they delivered significant improvements over top-of-the-line hardware.

In July of 2010, Rick decided to focus more on "bench rigs" capable of high overclocks for short bursts of time, as he re-entered the world of competitive hobbyist overclocking projects. Buckeye has stepped up to fill his shoes, and he even surpassed some of Rick's stable overclocked results.

By August 2010, Buckeye had come up with the blueprint for the first real 5.0 GHz design. After a great deal of research & development, we were able to roll out the "Trinity Lightning" model, featuring 6 cores of the famous Intel Gulftown overclocked to a perfect 5.0 GHz! Buckeye is always very eager to experiment and explore new overclocking techniques, and he has a grounded, pragmatic approach to building stable systems. It is a great honor for me to have him on the Liquid Nitrogen Overclocking team.