Testing:

Ambient temps have been all over the spectrum lately due to the near country wide heat wave. Outside temps have been in the 90s here in the Seattle where it's normally not very humid. The test room was kept to 24 C as best as possible. Readings were recorded using the sensors provided by the Koolance RP-1000 as well as the NVidia Panel. Medium and High speed settings were used to determine the best and most efficient results. Testing water blocks is pretty easy these days. You basically need a good processor or multi-core processor program to properly load each core to 100% for best results. CPU Burn and Prime 95 are still my two current favorites since they are both very small utilities that do a perfect job. For the Fuzion GFX 7900 GT, Fear and Battlefield 2142 were used to keep the GPU loaded. Since, the Fuzion GFX Uni-Sink is intended for a Geforce 8800 GTX (which I unfortunately do not have on hand), memory sinks from Koolance were used to make sure that the memory doesn't limit the overclocking results.

D-TEK Fuzion GFX Idle

Thermaltake ND3 TMG Idle

After a good 12 hour "burn-in", the system was allowed to idle for 2 hours to check both the GPU and CPU temps. With the cooling system set to medium speed, the Fuzion GFX sensor reported an excellent 30 C and 25 C with high settings. I've seen temps close to the 35 C mark before, but nothing quite as low as 25 C. This proves that the GFX cooler should be very, very capable of cooling any graphics card from the slowest all the way up to the top 8800 series cards. Under a graphics intensive load, the GPU finally reached 35 C going with a medium system speed.

Overclocking the card was less than perfect mainly due to the ambient temp and the use of only aluminum heat sinks which isn't the fault of the D-TEK kit. The memory modules just got too hot, while the GPU stayed very cool. The GFX cooler sensor reported 40 C at load with the system speed at high and 45 C when set to medium. A maximum GPU and Memory overclock was reachable, but was kept short of stable due to the memory getting too warm. However, a few less megahertz results in a miniscule performance gain at best. When 10 to 20MHz of frequency is involved, it really doesn't matter unless you are a complete OC and benchmark fanatic. The GPU core frequency is where the most performance comes from especially when you need those extra frames.

Typical Load Temps

Dual Core Load Temps

The Fuzion CPU sensor reported some also pretty decent results as well. The two best blocks in the test drawer are the Apogee GT and Koolance CPU-330. It will be nice being able to add Danger Den's current best cooler mainly to completely round off the entire round of heavy weights in future tests. Using a program called SpeedFan, it's just one of a few ways to see what each individual core sensor is reporting internally. Core 0 is the main core doing all the work when the program it's performing can't use the other cores. I'm not surprised to see that the blocks are very close when idle and under typical desktop load. There isn't much in the way of heat to dissipate since only one core is doing all the work. Browsing web pages, listening to music, or playing some of the games that don't support multi-core processors, won't get any of the water blocks warm enough to matter.

When a full load was placed on all four cores, the differences between each block is revealed. The Fuzion does surprisingly well for its simple design and stays close to the rest of the competition. The top benefiting factor for these temp results is the Koolance KIT-1050SL which includes a triple 120mm 1kw capable radiator. When you put that kind of cooling support to any set of blocks, you'll undoubtedly see great performance. 

One thing to keep in mind during all the testing, processor temps taken only for the active cores. For instance, if you run a single instance of Orthos, the program tests two cores at a time and tends to show the results accordingly. Regular graphics software or games, tend to utilize just one core out of the bunch. Occasionally, you will see two cores start to heat up which can effect the over all results which is why only the active core temps were taken. The results say it all for the D-TEK Fuzion and Fuzion GFX water blocks.

Conclusion:

Performance is surprisingly good, especially in regards to the GFX block. My first impression was based on the mounting method. A more secure block seemingly transfers heat better and the GFX mounting method is less than "toight" (tight). And being that it can tilt if a good amount of pressure is placed on the block, it didn't seem possible to do well. Alas, I was very wrong. In fact, the block is very efficient for its mounting configuration. Maybe the base plate is much thinner than any other blocks tested to date, or the large water movement helps, or perhaps the internal and external design make it more efficient. It's most likely all the above. One thing is for certain, the Fuzion CPU and Fuzion GFX water blocks are very, very capable and perform well with good water flow. The blocks keep up with any prospective competition in performance and design.

The footprint of the Fuzion and GFX both seem a little bit slimmer than previous generation blocks. They seem to require less space and aren't as obtrusive. The mounting method of the Fuzion CPU block isn't the most efficient, but it it is at least achievable. The components needed to mount it are very small and a pain to hunt down if you drop one. Size improvements of some of the mounting components could be a worth while venture. It wouldn't have to be much of a size increase, but something a little bigger could work easier. At the same time, I also understand that they probably want to keep components small and less noticeable when installed.

My thoughts on the D-TEK Fuzion and Fuzion GFX Uni-Sink Combo can be summed up in saying this is excellent first round performance and very good design as well. The foremost visual things about the blocks are their unique materials used in the fabrication process. They are much lighter than many of the CPU cooling blocks to come across the test bench. This is a very, very helpful feature when installing the GFX block on a bare graphics card processor. There have been casualties in the past when the cooling blocks were just too much for the video card GPU to handle which resulted in a damaged card. The GFX and its mounting procedure eliminate a great deal of concern in this regard and proved to be a successful design. You have to love the Delrin construction making it that much lighter to work with. The water cooling market just became more competitive thanks to Sharka Computers and D-Tek Customs.

Got News? Send 'em in!