Date: February 26, 2007
Author(s): Rob Williams
If you are on the lookout for an extreme overclocking kit but don’t want to make a huge withdrawal at the bank, then the DOMINATOR 9136 kit will be worth your consideration.
The memory market is unique. On the GPU side of things, it’s always bustling with new releases to drool over, but memory is far different. When we moved up to DDR2, the only thing to get excited over were faster speeds. So what can memory companies do to get people excited about their product? Release a great looking and efficient, heatspreader of course. Corsair caught the attention of many last year with the launch of their DOMINATOR heatspreader. At the same time, they announced their top end kit, the PC2-8888 with very tight timings of 4-4-4. That retained the top spot in their product line up until this past CES, where they announced their PC2-10000 kit.
The kit we are taking a look at today is placed in between the two, although it has less tight timings than the 8888. Because of this, the 8888 kit still retains a higher spot and should prove faster in all respects when compared to the PC2-9136 kit that we have on hand.
Before we jump into the specifics, we should touch on what makes the DOMINATOR so special. Heatspreaders have a simple purpose… to dissipate heat. This is primarily beneficial when overclocking, since stock speeds tend to not heat up the memory that much, unless purposely stressing them for hours on end. Overclocking is a big hobby for many, which is why heatspreaders caught on so fast. However, all spreaders on the market have one thing in common… they touch the memory chips and nothing else.
Go ahead, bookmark this review, shut down your PC, haul out your DDR2 and see what I mean. I’ll wait. *twiddles thumbs*
Great, now that we got that out of the way, we can move on. With heavy overclocking, heat plays a huge role, so the faster and more efficient heat can be dissipated, the better. This is where the unique design of the DOMINATOR comes into play. As you can see in the diagram below, not only do the FBGA chips have a heatsink on them, but also the the tip of the PCB. Then layered on top of these is the actual DOMINATOR heatspreader.
This is all paired with the fact that the modules use a custom PCB which was designed specifically for performance and heat transfer. Essentially what this should result in is better overclocks without the use of a fan. But as most people know, a fan will always be required for huge overclocks. It’s similar for CPUs. You wouldn’t dream of overclocking with a passive cooler; You’d be asking for trouble.
Along with the release of the PC2-8888 kit, Corsair also launched their Dominator Airflow fan, which is essentially three small quiet fans paired together to maximize airflow coverage over the entire area. These are not required to run stock speeds, but are highly recommended for overclocking for the same reason mentioned above. Sadly, this fan is only included with the 8888 and 10000 models. You can purchase it separately though, for $20. That’s fine though, as the 9136 kit retails for around $430 while the 8888 kit is closer to $600.
But, enough about the modules that we are not dealing with today. Shortly after the 8888s were released, Corsair followed-up with a higher clocked kit with loose timings, so that it could be priced better for those who don’t want to spend more than a house payment on a new kit of ram.
What we are dealing with are presumably Micron D9GMH chips, while the higher kits likely use the higher binned D9GKX. Does this mean that this kit won’t be as overclocking friendly? PC2-9136 equates to 571MHz frequencies, aka DDR2-1142. Being that high already, DDR2-1200 is not outside the realm of reason with a little extra voltage. Speaking of… stock voltages are 2.1v, not bad at all given the frequency. The timings are 5-5-5-15, which seems loose but isn’t when dealing with these frequencies. To hit 4-4-4 at anything 600MHz and over, you will require a -lot- of voltage, which isn’t safe by any means.
You are probably interested in packaging… so here we are. Given the fact that these spreaders are huge, so is the plastic wrapped around them.
These are great looking spreaders any way you look at it. I have high respect for anyone that slides away from the incredibly dull green PCB and uses black instead. For an enthusiast part, it adds a lot, and suits the spreaders themselves perfectly.
Close up to the sticker we see revealed the model code, frequency, timings and revision number. The only thing that is missing here is the required voltage. If you use the EPP feature, that will be automatically set for you inside the BIOS.
Looking top-down, you can see how the heatspreader comes together. It may double as a comb if you are in a bind.
Lastly, here is the module when compared to OCZs latest Flex. The Flex is by far the taller module, thanks to the fact that it accepts tubing for water cooling.
On to overclocking!
Is there even a need to overclock such a fast kit of ram? Of course there is, are you kidding? The overclocking headroom with this kit actually impressed me quite a bit, given the fact that they are not the highest binned that Corsair offers. However, my first goal was to tweak the timings down to 4-4-4, which proved 100% stable at 565MHz, or DDR2-1130… more on par with their PC2-8888 kit. This required a healthy 2.4v, which is still in the confines of being safe, especially with the Dominator fan.
Using the same timings as stock, I managed a surprising stable overclock of 635MHz, aka DDR2-1270, aka PC2-10160. I am not talking about a max overclock here, but max stable overclock, as you can see from the picture below. This required a slight bump in the voltage, at 2.5v. This is the max voltage that the eVGA 680i board will allow, so chances are good there is still room with going higher. DDR2-1300 stable does not sound foolish if you have the capability for 2.6 or even 2.7v. Anything above that is a tad ludicrous though, unless you have a fetish for burning out perfectly good modules.
That said, here are the five stable settings I used (CL-tRCD-tRP-tRAS-tRC):
While most of the settings above used tRC 13, the max overclock of DDR2-1270 used 16 as anything lower would bring on errors in MemTest for Windows. tRC 16 was a dream though, as you can see in the photo above, where the memory had a 1000% coverage error free.
Throughout all of our benchmarks regardless of what we are reviewing, testing is done in a clean and stand-alone version of Windows XP Professional with SP2. Prior to testing, these conditions are met:
The testing rig used for today’s benchmarking is as follows:
Most of the results listed deal with how all of our settings scaled, although there will be direct comparisons to other kits of ram as well.
First up is one of my favorite benchmarking tools, SiSoftware’s Sandra XI. All of the results here are buffered, so the CPU does come into play with these scores. In general, there are not huge bandwidth differences between frequencies, but the CPU plays a much larger role. Tighter timings can also make a larger difference than higher frequencies, as you can see when comparing the bottom two results. Even though the second last setting has a full 100MHz on the bottom setting, the extra loose timings didn’t help anything at all.
Great results here either way. Our top overclock almost reached 7,000 and would have, if the CPU was closer to 3.0GHz.
Everest is another tool I use regularly, for similar reasons to Sandra. It’s a quick and easy way to see how your memory overclocks scale to each other. Once again, great results all around. One interesting thing to note though, are the bottom two results again. In our Sandra tests, we saw that the extra frequency compared to tighter timings didn’t really have an advantage. Here however, our DDR2-800 with 4-4-4 timings outperforms the DDR2-1000 with 5-5-5 timings.
The Write results are also apparently heavily CPU dependant. The bottom four results were all done with a clock speed of 2.4GHz and the results are spot on. This chart below perfectly shows just how beneficial overclocks can be on your latency. Moving from the already speedy stock speed to our maximum shaved off near 5ns.
In our Sciencemark tests, we can see again how beneficial tighter timings are over higher frequencies. The bottom two results prove that well, with DDR2-800 actually providing better bandwidth over DDR2-1000 thanks to it’s 4-4-4 timings. Our top overclock helped up break through the 6,000 mark, with stock speed still having a nice result.
Super Pi is the only benchmark we run in our memory reviews that doesn’t directly affect the RAM itself. Instead, we are looking to see how beneficial faster memory is when it comes to hardcore CPU intensive tasks, suck as calculating Pi to 8 Million digits.
The bottom four results to me are the most interesting, simply because they were all run with the identical clock speed. The only thing that was changed was the memory frequency or timings. DDR2-1000 4-4-4 proved faster than DDR2-1066 4-4-4 which could be due to a number of reasons. Perhaps at the higher clock speed, some secondary timings were affected. Those timings in question are not available from within the BIOS though, so I am unable to state that fact for certain.
Since this is my first RAM review in a while, I am re-evaluating how we can make these reviews better overall. If you have recommendations of benchmarks you’d like us to consider or different methods of testing, please let us know and we will look into them.
That said, starting in this review, we will be comparing our overclocks of the memory being reviewed alongside other recently evaluated sticks. These graphs include benchmarks with each kit of ram at DDR2-1000 4-4-4-12-13 2.1v along with each kits own top overclock. While the DDR2-1000 results should not vary much, the top end overclocks will, given that each kit will top out differently.
For reference, here are the top overclocks for each kit of ram included:
Additional kits will be included as time passes. I had other kits on hand, such as Kingstons HyperX PC2-8000, but they did not give realistic results. I assume they are not fully compatible with the 680i motherboard.
In our first comparison, EVEREST, only the top overclocks where used. It’s no surprise that since the Dominator 9136 had the top overclock, it also has the top result. It scales just as expected.
The Dominator showed its true colors here, claiming the top spot for our highest overclock. The 500MHz settings proved a little different though, with the OCZ VX2 easily taking the crown.
Not much is a surprise here. The Dominator proved the best overclocked time, thanks to the slightly higher CPU frequency. As far as 500MHz goes, the VX2 once again beat out the other kits by a smidgen.
As with most top-end memory, this kit is not inexpensive. It retails between $425 and $450 on various websites. Despite its high price, it still falls in line with other similar kits on the market. The OCZ Flex PC2-9200 kit for example costs around $40 more than this Corsair one. However, there are many kits (Corsair included) that can be had at PC2-8500 speeds for over $100 less. One thing is for sure, the kit we had on our test bench today is amazing. It offers plenty of overclocking ability, as was evidenced earlier. DDR2-1270 stable is one nice feat.
I don’t doubt for a minute that there is more headroom here as well, if you are able to go beyond 2.5v. Most huge overclocks require between 2.7v – 2.8v, so DDR2-1300 stable would not be a surprise to me, with even higher (but less stable) overclocks possible as well. As it stands, the fact that 635MHz was stable at a still modest 2.5v impressed me.
The only real downside is that this kit does not include the airflow fan, but luckily they can be purchased separately for $20. I have come to enjoy using this as opposed to a usual 120mm, for the simple fact that it’s easier to mount. I had originally thought that having three mini-fans constantly going would be noisy, but it’s not. I can’t even hear it overtop my water cooling fan, so I recommend that accessory to anyone, regardless of whether you have this exact kit or not.
That said, price won’t be taken too heavily into consideration here as it would if there were equivalent kits out there for a much lower price. At this speed range, this kit is much lower priced than OCZs Flex, and I managed to clock it higher. In the end, I am impressed. If you don’t mind paying a premium for some of the fastest memory on the market, this is a fantastic kit. If you are more of a novice overclocker or an enthusiast who would like to save money, there are many other budget-conscious kits out there that would better suit you.
If you have a comment you wish to make on this review, feel free to head on into our forums! There is no need to register in order to reply to such threads.
Copyright © 2005-2020 Techgage Networks Inc. - All Rights Reserved.