What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

Corsair Dominator GTX2 PC3-18000 CL8 Memory Review

Status
Not open for further replies.

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
<center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/corsair_logo-1.png" alt="Corsair Dominator GTX2">


<b>Corsair Dominator GTX2 PC3-18000 CL8 Memory Review</b></center>


<b>Price:</b> $200 USD/Per 2GB Stick <a href="https://shop.corsair.com/store/item_view.aspx?id=1235506">Corsair Online Store</a>
<b>Manufacturer Product Page:</b> <a href="https://shop.corsair.com/store/item_view.aspx?id=1235506">Corsair</a>
<b>TechWIKI Info:</b> <a href="http://techwiki.hardwarecanucks.com/product/1OTEyNjMxMDcxMDY2NDA/Corsair-Dominator-GTX/">Corsair Dominator GTX2 - TechWIKI</a>
<b>Manufacturer's Part Number:</b> CMGTX2
<b>Warranty:</b> Lifetime Warranty




<p style="text-align: justify;">The snow has recently fallen here on the toilet bowl that is south western Ontario. So while it may have finally turned into winter outside; inside things are just starting to heat up. Well, not so much in a literal sense since the Dominator heat sinks are keeping things cool. But rest assured, the performance and frequencies coming from these Corsair Dominator GTX2 modules are sure to get the blood boiling. Today we look to tame the slow cold pace of a winter's storm with some high clocks, tight timings, and everything in-between.

It is no secret that Corsair has been pushing the memory envelope since their popular Dominator line of memory debuted during the DDR2 frequency peak. The Dominator name was upgraded with a very strong GT acronym when we looked at the PC3-15000 Dominator GT last winter which was almost a year ago already. Since then, the Dominator GT line has seen some updates with modules reaching the lofty clocks of DDR3-2000 at timings as tight as 7-8-7. Of course, Corsair hasn't stopped there going completely overboard introducing the GTX1 and GTX2 DDR3 modules.

The CMGTX1 memory is designed for P55 platforms running up to DDR3-2400 at timings of 9-11-9. The CMGTX2 modules are specified for operation on either P55 or X58 platforms at DDR3-2250 8-8-8, and AMD 790FX platforms at DDR3-1900 6-6-6. Needless to say, both modules are the highest clocked DDR3 sticks Corsair offers for either P55, X58, or AMD platforms. Today we will be looking at the Corsair Dominator GTX2 modules. But unlike memory reviews of past, we will be taking the overclocker's express route with these sticks focusing primarily on overclocking and all but ignoring stability testing. This CMGTX2 memory is built for benchmarking, not running Prime95 or LinX...so we are simply going to do what it is designed for today.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/index-2.jpg" alt="Corsair Dominator GTX2"></center>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Specifications

Memory Specifications

<p style="text-align: justify;">We start with a quick look at the specifications. The Corsair Dominator GTX2 modules are the 'bigger little brother' to the Corsair Dominator GTX1 sticks. I say bigger little brother because it is a matter of opinion as to which kit you might want to call the top of the specifications sheet. We'll discuss this in a moment. First, have a look at the rap sheet that accompanies these Corsair flagship modules.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/specs-1.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">First thing we have to discuss is the fact that these modules come in individual pieces. There are no kits, so you simply purchase however many modules you need; whether it is two for P55, or three for X58. This is a bit of a foreign concept as we have long been used to the idea of buying memory in kits. Selling the modules this way alleviates the need for different kits for each platform but we like the decision Corsair has gone with here since it just makes sense to us.

We just mentioned that there was some hierarchical confusion as to which Dominator GTX memory module is at the top of the Corsair pile. The reason we say this is because the GTX1 modules -while rated for much higher frequencies- have very loose timings and also don't have the flexible specifications of the GTX2 modules we are looking at here today. You can see in the specs above, these modules are rated for DDR3-2250 at 8-8-8-24 timings with 1.65v for the P55 and X58 platforms. They are also rated for a very lofty DDR3-1900 6-6-6-18 at 1.65v on the AMD 790FX platform. That makes these sticks a powerhouse on pretty much any modern platform.

The GTX1 modules on the other hand have ungodly high memory clock ratings of DDR3-2400, but only for P55 and with very lax 9-11-9-27 timings. The validity of memory clocks that high for daily use can also be questioned. While we are on the subject, justifying the need for even DDR3-2250 that these GTX2 modules are rated for can be a difficult argument. It is for this reason that we are stripping down our usual memory review for more of a report on overclocking from a benchmarking point of view here today. We will re-visit this concept in our methodology section shortly.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/specs-3.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">Before we look at the CMXAF2 GTL, we offer a diagram from Corsair that outlines the DHX heat sinks on these modules. Nothing has changed in a while with these heat sinks so there is nothing new if you are already familiar with them. Another quick point we want to make is that despite not removing the heat sinks, we are fairly confident that the IC's in use on these modules are Elpida Hyper . Photos in the next section will confirm this but with specifications like these, it goes without saying that Elpida Hyper at the controls.
</p></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/specs-2.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">This information about the CMXAF2 GTL fan is coming straight from the Corsair web site.

Product Description
The CMXAF2 GTL fan assembly is a fan kit made especially for the GTX 2 memory modules. It has extended legs to give you the necessary clearance to accommodate the high-performance DHX+ fins.

Key Features
  • Get the most out of your memory modules by keeping them cool
  • 2 x 60mm fans
  • Compatible with all Corsair Memory products
</p><p style="text-align: justify;">The one thing we will mention is that this fan comes with taller mounting tabs because of the taller heat sink fins on the GTX modules. Our earlier fan that came with a kit of Dominator GT's would not fit the new GTX modules as the mounting tabs were too short, so keep that in mind if upgrading. Now that we know exactly what we are working with today we can take a look at the modules themselves.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Package & Memory Overview

Package & Memory Overview

<p style="text-align: justify;">The Package & Contents section has two parts. We have the Corsair Dominator GTX2 modules themselves, as well as the CMXAF2 GTL fan that can be purchased separately. We'll start with a look at the packages they come in then have a quick peek at each piece of hardware.</p><center>
package-1.jpg
package-2.jpg
package-3.jpg
</center><p style="text-align: justify;">In the specifications section we covered the fact that these modules are sold individually. Therefore it only makes sense to package them individually, and that is what we see here. We received three module packages and one fan assembly package from Corsair. Both the modules and the fan are very neatly and secured wrapped in their appropriate boxes. Enclosed in anti-static sealed bags, smothered in an air pillow, the Dominator GTX2 modules have virtually no way to be damaged in shipping.</p><center>
modules-1.jpg
modules-2.jpg
</center><p style="text-align: justify;">We very briefly touched on the height of the new CMXAF2 GTL fan assembly. Aside from the length of the side mounting tabs; it is virtually identical to what we received with our Dominator GT triple channel kit of memory almost a year ago. The fans move a lot of air without a lot of noise and the mounting is really easy with this fan so we give it the two thumbs up. Personally I still like using a 120mm fan sitting over the memory on an angle as it will then push air over the NB location, but for those with adequate cooling around their motherboard, the CMXAF2 GTL is a great unit.</p><center>
modules-3.jpg
modules-4.jpg
</center><p style="text-align: justify;">The Dominator GTX2 modules utilize the Corsair DHX heat sink but come accented with tell-tale taller cooling fins on top. The cooling fins are rosso corso red as are all the color accents on the modules. If you are familiar with the Dominator GT, or even Dominator memory, you will notice that there really aren't any differences. This is a positive thing in our view because the cooling system has worked well in the past. At the same time, for typical use in a daily system, the DHX heat sinks are a bit overkill. When it comes to cooling though, overkill is never a bad thing in our minds.</p><center>
modules-5.jpg
modules-6.jpg
</center><p style="text-align: justify;">We mentioned that the only real difference is the cooling fins at the top for these modules compared to the other lines of Dominator memory from Corsair, but the fins are removable and can actually be purchased at the Corsair online store. So for those wanting to dress up your GT's to GTX spec, you can. Keep in mind, your GT supplied fan likely won't fit, but we can't confirm that so you will want to make sure yourself.</p><center>
modules-7.jpg
modules-8.jpg
</center><p style="text-align: justify;">On the left we have the first set of Dominator GTX2 modules we received. On the right, we have the second set we received. The macro photos taken make it pretty clear: the first set of modules had very little to no contact between the IC's and the heat sinks. The second set (photo on the right) looks a lot better but still isn't perfect. The center module is pretty much perfect, but the other two modules have one side that is not sitting perfectly flush to the IC's. In all the Dominator modules we have seen over the years, this is the first time we have seen anything of this nature. Normally this wouldn't be an issue as sometimes things happen in manufacturing but when we are looking at their absolute best modules at $200/stick; you would like to see perfect quality control. Perhaps we are asking too much, but we don't think so.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Install & Test Setups

Install & Test Setups

<p style="text-align: justify;">We are going to have a quick look at potential installation issues with the GTX2 modules. As noted, the top heat sink cooling fins are taller than standard Dominator modules and on some motherboards with large processor coolers in specific orientations, there are conflicts when using the slots closest the CPU.</p><center>
install-1.jpg
install-2.jpg
</center><p style="text-align: justify;">As we already knew based on our dominator and other 'tall' memory heat sink experience, the first slot in both the EVGA X58 and EVGA P55 Classified boards will not work with these modules in the first slot. With that said, the industry standard is to optimize the second set of slots for primary use. That doesn't alleviate the issue for those that want to run all four slots on P55 or six on X58, but even in those situations you can turn the TRUE and run it in the other orientation. Either way, the height of the heat sinks are going to play a role in a portion of peoples setups, so be aware of their height.</p><center>
install-3.jpg
install-4.jpg
</center><p style="text-align: justify;">In the first photo above we can see that installation of the KingpinCooling.com F1 EE liquid nitrogen cooling pot doesn't interfere at all with memory modules. We even have space for a second layer of insulation around the pot as you will see in some photos below. In the second photo we have shown the difference in height of the two different Dominator heat sinks. The regular Dominator heat sinks in the foreground are about 11mm shorter than the GTX modules. Time now to take a look at what we are going to be cooking with today...I apologize in advance for the mess.</p><center><table cellpadding="0" cellspacing="0" width="735px"><tr><td align="left">
setup-1.jpg
</td><td align="left">
setup-2.jpg
</td><td align="right">
setup-3.jpg
</a></td><td align="right">
setup-4.jpg
</a></td></tr></table><br /><table border="0" bgcolor="#666666" cellpadding="5" cellspacing="1" width="735px"><tr><td colspan="4"><b><font color="#ffffff">P55 Test Platform:</font></b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Motherboard:</b></td><td align="left" bgcolor="#ececec" width="75%">EVGA P55 Classified 200</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor:</b></td><td align="left" bgcolor="#ececec" width="75%">Intel i5 670 (934B282)<br />Intel i7 860 (925B478)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Chilly1 SS (Tuned by Ruffus)<br>KingpinCooling.com F1 EE & LN2</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Thermal Paste:</b></td><td align="left" bgcolor="#ececec" width="75%">Arctic Silver Ceramique</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>PCH Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>NF200 Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>PWM Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Memory:</b></td><td align="left" bgcolor="#ececec" width="75%">Corsair Dominator GTX2 2x2GB PC3-18000 8-8-8-24 (CMGTX2)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Power Supply:</b></td><td align="left" bgcolor="#ececec" width="75%">Corsair HX1000W</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Video Card:</b></td><td align="left" bgcolor="#ececec" width="75%">Diamond HD5870</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Additional Fans:</b></td><td align="left" bgcolor="#ececec" width="75%">Scythe Ultra Kaze 120MM 2000RPM 87.6CFM (DFS123812L-2000)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Hard Drives:</b></td><td align="left" bgcolor="#ececec" width="75%">Seagate 7200.9 80GB SATAII 8MB cache</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>OS:</b></td><td align="left" bgcolor="#ececec" width="75%">Windows XP SP3<br />Windows Vista SP2</td></tr></table></center><p style="text-align: justify;">Our P55 setup is centered around the EVGA P55 Classified 200 that we reviewed not so long ago. Now this platform has its issues when it comes to overclocking memory. The word on the street is that the EVGA FTW is better for memory clocks but some people have no problem with the Classified and Lynnfield processors...we do. Regardless, it is all we have for memory overclocking so we will make do with what we have. You should also notice the i5 Clarkdale processors listed above, these are used primarily for the fun overclocking seen at the end of the review. Let's now see what we have cooked up in the way of X58 overclocking.</p><center><table cellpadding="0" cellspacing="0" width="735px"><tr><td align="left">
setup-5.jpg
</td><td align="left">
setup-6.jpg
</td><td align="right">
setup-7.jpg
</a></td><td align="right">
setup-8.jpg
</a></td></tr></table><br /><table border="0" bgcolor="#666666" cellpadding="5" cellspacing="1" width="735px"><tr><td colspan="4"><b><font color="#ffffff">X58 Test Platform:</font></b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Motherboard:</b></td><td align="left" bgcolor="#ececec" width="75%">EVGA X58 Classified</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor:</b></td><td align="left" bgcolor="#ececec" width="75%">Intel Xeon w3540 (3845B010)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Chilly1 SS (Tuned by Ruffus)<br>KingpinCooling.com F1 EE & LN2</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Thermal Paste:</b></td><td align="left" bgcolor="#ececec" width="75%">Arctic Silver Ceramique</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>NB Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>SB Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>PWM Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Memory:</b></td><td align="left" bgcolor="#ececec" width="75%">Corsair Dominator GTX2 3x2GB PC3-18000 8-8-8-24 (CMGTX2)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Power Supply:</b></td><td align="left" bgcolor="#ececec" width="75%">Corsair HX1000W</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Video Card:</b></td><td align="left" bgcolor="#ececec" width="75%">2 x MSI GTS250 1024MB</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Video Card Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock<br />KingpinCooling.com Tek9 4.0 Slim & LN2</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Additional Fans:</b></td><td align="left" bgcolor="#ececec" width="75%">Scythe Ultra Kaze 120MM 2000RPM 87.6CFM (DFS123812L-2000)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Hard Drives:</b></td><td align="left" bgcolor="#ececec" width="75%">Seagate 7200.9 80GB SATAII 8MB cache</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>OS:</b></td><td align="left" bgcolor="#ececec" width="75%">Windows XP SP3<br />Windows Vista SP2</td></tr></table></center><p style="text-align: justify;">The X58 platform utilizes our trusty, and beaten, EVGA X58 Classified motherboard. This motherboard has been to WWII, Vietnam, and two deployments to Iraq. It takes a beating and always comes back. It has also been the guinea pig to a number of volt-mods in attempts to lower cold bugs. Needless to say, ever since the day it showed up, it has been the primary X58 motherboard used in this benching lab. The choice of GTS 250's for GPU's might seem odd with such a powerful setup but we had access to them so we figured we would have some fun and go gold cup hunting at HWBot.org for the review. Those results will show up in the Extreme Overclocking section.

Now that we are familiar with the modules and the rest of the supporting cast, we will go over our overclocking methodology and explain in depth exactly how we are going to test these modules today.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Overclocking Methodology

Overclocking Methodology

<p style="text-align: justify;">We are trying something new in this review of the Corsair Dominator GTX2 modules. We figured that since the memory was already specified to run at the internal memory controller (IMC) punishing frequency of DDR3-2250, that trying to push beyond those clocks for 24/7 stability was profitless. In memory reviews of the past, we have demonstrated that memory overclocks are of the least important factor in overall system performance using traditional programs as the measuring stick. After DDR3-1600, for most anything but competition type benchmarking, memory really doesn't play a massive role in performance.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/method-1.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">With this in mind, we are going to limit the overclock stability testing to that of single Super Pi 32M. This is in no way, shape, or form intended to mimic 24/7 daily operation stability. This methodology for testing stability is directed more towards how the memory will be used in a benchmarking environment. This is personally the first thing I do with any kit of new memory in order to understand how it scales with voltage and timings. Single 32M Super Pi stable is an excellent gauge for what memory can handle as far as benchmarking is concerned. <img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/method-2.png" alt="Corsair Dominator GTX2" style="float: right; margin: 4px 0px 4px 7px"> Generally 3D Mark stability is the same or better. It is quick and relatively consistent in dictating exactly what a kit of memory is capable of.

The charts on the next two pages are straight forward enough that only a couple seconds of examination will be required to understand the information contained within them, but we would still like to outline how we go about overclocking. As the example chart to the right outlines, we will be testing the memory at <b>1.65v</b>, <b>1.70v</b>, and <b>1.75v</b> vDIMM. We wanted to test at 1.80v as well but there was concern for damaging the memory above 1.75v and wanted to ensure we didn't degrade the memory at any point during the review process. Not only do we test the memory at the three voltages, but we will also be testing various timing sets. Together, this gives us a great idea of how the memory scales with both timing, and voltages.

The cooling on the processor for all of our single 32M Super Pi overclock testing in the next two sections came by way of a single stage phase change CPU cooler, the Chilly1 unit listed in the previous section. This helps eliminate CPU limits in the memory overclocking; we will however, likely still see limits other than the memory on both setups. The OS used for these sections will be Windows XP SP3. The last section is an all out Extreme Overclocking bonanza of sorts. We basically used this kit for various sessions with the CPU and GPU's under LN2 and will outline each setup in that section. That should cover it, so let's move on and see how the Corsair Dominator GTX2 modules shake down.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
P55 Overclocking

P55 Overclocking

<p style="text-align: justify;">The phase change is down to about -30C, the timings set, the OS setup...there is nothing left to do but see how the memory clocks.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/p55_clocking-1.png" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">Unfortunately, this discussion is going to revolve around caveats starting with the top red result. Obviously the memory clocks didn't increase, but this is absolutely not a limitation of the memory. In fact, the voltage of 1.65v listed is rather misleading, this motherboard completely hates us. With newer BIOS's the memory clocks have absolutely gotten worse with our sample and the i7 860 processor we have paired with it. Check the screen shots below for the 1100MHz 8-8-8 result, it was actually achieved at 1.56v!

At CL6 and CL7, these sticks really clock well except at 7-7-6. We fought with this timing combination for a long time but just couldn't get 7-7-6 to scale with volts much at all. As we can see, it starts out actually clocking higher than the 7-8-7 timing set but fails to improve with voltage. 7-8-7 clocks right up to a very healthy 1086MHz, almost equal with 8-8-8. DDR3-2172 at 7-8-7 with only 1.75v is very nice indeed. So far the Dominator GTX2 modules have performed as expected.

As good as the results are, we have to absolutely bring this up. The numbers achieved here are fantastic, don't get us wrong, but they are also on par with any other PC3-16000+ kit of memory that uses Elpida Hyper IC's. Some triple channel kits that cost as much or significantly less than two sticks of these Dominator GTX2 modules will overclock just as well. The other thing we want to say is that we did do some testing on our first set of memory received with the poor heat sink application. Before sending them back we achieved much better results than is the case with these modules. These particular sticks do not like to run 6-6-5 where as our last kit clocked 6-6-5 almost as well as 6-7-6. In addition, the other pair clocked 7-7-6 better than this pair clocks 7-8-7. That is the probably the greater reason for our disappointment. Despite the high price tag, the Dominator GTX modules still appear to have a wide range of capabilities while we hoped they would be binned to near-perfection. It really was a night and day difference between the second set of modules we received and the first. The last thing you want to be hoping for with a $200 stick of ram is whether or not it is going to be great...they should all be better than great in our opinion. Like everything in overclocking, however, it comes down to luck of the draw.

Here are the screen shots of the runs at 1.75v...</p><center><table align="center" bgcolor="#666666" cellpadding="5" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="25%">6-7-5 Overclock @ 1.75v
click for full size...
p55_clocking-2.png
<br>CMGTX2 @ 950MHz 6-7-5-18 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">7-7-6 Overclock @ 1.75v
click for full size...
p55_clocking-3.png
<br>CMGTX2 @ 1059MHz 7-7-6-20 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">7-8-7 Overclock @ 1.75v
click for full size...
p55_clocking-4.png
<br>CMGTX2 @ 1086MHz 7-8-7-20 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">8-8-8 Overclock @ 1.56v
click for full size...
p55_clocking-5.png
<br>CMGTX2 @ 1098MHz 8-8-8-20 1T<br>@ 1.563v under load</td></tr></table></center>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
X58 Overclocking

X58 Overclocking

<p style="text-align: justify;">We have already seen our first set of numbers on the P55 platform and we expect to see very similar results on X58. The advantage we have over here is the lack of a limit on 8-8-8 clocks...well, until the IMC on this poor abused W3540 gives out anyway.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/x58_clocking-1.png" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">As it turns out, we are still being limited by the system...this memory loves 8-8-8 timings. At one point in the life of this processor, it was capable of running DDR3-2400 but time and benching abuse has clearly taken its toll. The Dominator GTX2 modules had no problem out-running the uncore capabilities of the IMC while still being stuck at 1.65v. When we tried to lower volts at these clocks we got instability so it looks like we would have had to bump vDIMM very shortly anyway. As it stands, we don't have a processor/setup-P55 or X58-that can keep up with this memory at 8-8-8 timings.

Looking at 7-7-6 and 7-8-7, we can see an almost identical pattern as the P55 setup. The two timing sets clock about the same with a slight advantage for the 7-8-7 setup at 1.75v. It is a clear cut case of these particular modules not liking tRCD 7. As mentioned in the P55 overclocking section, our first set of modules from Corsair loved 7-7-6 so the variation is quite apparent.

On the whole though, the X58 clocks appear to be slightly lower than P55 which is expected given the fact that we have three modules cooking opposed to just two, but the clocks are equally impressive. Just to provide an example of what we were discussing in the P55 section about our first set of modules being stronger, here is a screen shot of what the first kit could do: DDR3-2184 // 7-7-6-20 @ 1.75v. As you can see, the difference between the sets is pretty substantial with this timing set. Keep in mind the context though; we are looking at this memory from a benchmarking perspective so 30MHz at an important timing set like 7-7-6 can play a crucial role. For daily use, the numbers above are ridiculous. Granted, 24/7 stable overclocks are going to be lower by anywhere from 20-30MHz...possibly even a bit more depending on the setup.

Here are the screens from the results above at 1.75v vDIMM...</p><center><table align="center" bgcolor="#666666" cellpadding="5" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="25%">6-7-6 Overclock @ 1.75v
click for full size...
x58_clocking-2.png
<br>CMGTX2 @ 942MHz 6-7-6-18 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">7-7-6 Overclock @ 1.75v
click for full size...
x58_clocking-3.png
<br>CMGTX2 @ 1062MHz 7-7-6-20 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">7-8-7 Overclock @ 1.75v
click for full size...
x58_clocking-4.png
<br>CMGTX2 @ 1072MHz 7-8-7-20 1T<br>@ 1.751v under load</td><td align="center" valign="top" bgcolor="#ececec" width="25%">8-8-8 Overclock @ 1.65v
click for full size...
x58_clocking-5.png
<br>CMGTX2 @ 1173MHz 8-8-8-20 1T<br>@ 1.652v under load</td></tr></table></center><p style="text-align: justify;">We will now turn these modules loose on some real benchmarking...after all, this is what they are meant for. As mentioned earlier in the test setups section, we have a pair of MSI GTS 250's that we will do some Gold Cup hunting for on HWBot.org as well as a Clarkdale processor that is going to knock your socks off, let's take a look.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Extreme Benchmarking

Extreme Benchmarking

<p style="text-align: justify;">Alright, it is time to have some fun. The first item on the agenda is some gold cups for these MSI GTS 250's.</p><center>
xoc-1.jpg
xoc-2.jpg
</center><p style="text-align: justify;">We talked earlier about how a pair of GTS 250's seem like an odd GPU combo to put under LN2 for an overclocking exhibition, but they happened to be laying around while we were benching with these modules so we thought; why not make them dance? Unfortunately they had cold bugs around -30C so even with volt mods and an LN2 pot strapped on...there is only so much we could do. The setup for all the results with these cards is the X58 platform. We have a ton of results though so we will keep the yapping to a minimum and get to the scores.</p><center><table align="center" bgcolor="#666666" cellpadding="10" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 01 - Multi GTS 250
click for full size...
xoc-3.png
<br>CMGTX2 @ 1097 7-8-7-20 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 03 - Single GTS 250
click for full size...
xoc-4.png
<br>CMGTX2 @ 1147 8-8-8-24 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 03 - Multi GTS 250
click for full size...
xoc-5.png
<br>CMGTX2 @ 1147 8-8-8-20 1T</td></tr><tr><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 05 - Single GTS 250
click for full size...
xoc-6.png
<br>CMGTX2 @ 1147 8-8-8-24 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 05 - Multi GTS 250
click for full size...
xoc-7.png
<br>CMGTX2 @ 1147 8-8-8-20 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 06 - Single GTS 250
click for full size...
xoc-8.png
<br>CMGTX2 @ 1082 7-8-7-20 1T</td></tr><tr><td align="center" valign="top" bgcolor="#ececec" width="33%">3DMark 06 - Multi GTS 250
click for full size...
xoc-9.png
<br>CMGTX2 @ 1067 7-8-7-20 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">AM3 - Single GTS 250
click for full size...
xoc-10.png
<br>CMGTX2 @ 1167 8-8-8-24 1T</td><td align="center" valign="top" bgcolor="#ececec" width="33%">AM3 - Multi GTS 250
click for full size...
xoc-11.png
<br>CMGTX2 @ 1147 8-8-8-20 1T</td></tr></table></center><p style="text-align: justify;">When it was all said and done, our Corsair Dominator GTX modules helped power these GTS 250's to a total of 6 x
hwbot_gold-1.png
, 1 x
hwbot_silver-1.png
, and 1 x
hwbot_bronze-1.png
at HWBot.org. That is not bad for a couple of quick sessions. The modules ran great and were 100% predictable relying on the 32M Super Pi testing done earlier. We will now bring the temps down a little bit more and let the Dominator GTX2 sticks loose on our EVGA P55 Classified 200 Clarkdale setup for some 6GHz+ overclocking fun.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/xoc-12.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">As was just mentioned, the platform is our EVGA P55 Classified equipped with a very hot little i5 670 Clarkdale. High clocks were the order of the day and the setup did not disappoint. Matched up with an HD5870, we tossed in a couple 3D results to compliment the regular 2D numbers that this setup excels at.</p><center><table align="center" bgcolor="#666666" cellpadding="10" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="50%">3DMark 03 - Single HD5870
click for full size...
xoc-13.png
<br>CMGTX2 @ 713 6-6-6-18 1T</td><td align="center" valign="top" bgcolor="#ececec" width="50%">3DMark 05 - Single HD5870
click for full size...
xoc-14.png
<br>CMGTX2 @ 713 6-6-6-18 1T</td></tr><tr><td align="center" valign="top" bgcolor="#ececec" width="50%">Pifast @ 6720MHz
click for full size...
xoc-15.png
<br>CMGTX2 @ 746 6-6-6-18 1T</td><td align="center" valign="top" bgcolor="#ececec" width="50%">32M @ 6248MHz
click for full size...
xoc-16.png
<br>CMGTX2 @ 925 6-7-6-18 1T</td></tr><tr><td align="center" colspan="2" valign="top" bgcolor="#ececec" width="100%">SuperPi 1M @ 6734MHz
click for full size...
xoc-17.png
<br>CMGTX2 @ 748 6-6-6-18 1T</td></tr></table></center><p style="text-align: justify;">This wasn't a gold hunting expedition like with the GTS 250's above; just some good old fashioned down home benchmarking. The Clarkdale platform -this processor anyway- isn't very conducive to memory overclocking but even at nice tight timings of 6-6-6, the Dominator GTX modules hold their own. The 32M run above does show off the memory to a certain extent, running 925MHz at 6-7-6...just ignore the CPU-Z memory tab, it doesn't report CAS Latency correctly with Clarkdale yet.

Overall, there are a couple global top 20's and a gold cup for the 32M 670 category...not too shabby for just fooling around.</p>
 
Last edited by a moderator:

3oh6

Well-known member
Joined
Mar 18, 2007
Messages
1,049
Location
Edmonton, AB
Conclusion

Conclusion

<p style="text-align: justify;">The fact that we can now buy memory that is specified to run at DDR3-2250 with 8-8-8 timings and only 1.65v is absolutely astonishing. A year ago that idea would have been downright ludicrous had you tried to tell us that. The fact of the matter is that memory is now capable of frequencies that far exceed most setups and as we saw with the EVGA P55 Classified setup, anything over DDR3-2200 just wouldn't run for us. Obviously this means we cannot run the Dominator GTX2 modules at even spec. This isn't typical of all P55 platforms but clearly was the case for us and goes to highlight that setups which can really push these modules could be a rarity.

Furthermore, our 8-8-8 overclocking results on the X58 platform were limited by the CPU being used as it couldn't handle the uncore any higher than DDR3-2346, and our CPU was at sub-zero temperatures. It really is a nice problem to have; memory that can literally outrun your setup, but it is something we want to stress. There is going to be some effort in order to get these modules to be running at spec with certain setups. At the same time, many users will pop them in and be running at DDR3-2250 or higher within minutes. It is just the nature of the game when it comes to high-end memory.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/corsair/domgtx2/conclusion-1.jpg" alt="Corsair Dominator GTX2"></center><p style="text-align: justify;">Our review today revolved around the Dominator GTX2's capability to perform in a competitive benchmarking environment. This requires a different set of criteria to designate pass or failure. On one hand, the modules easily made our setups look pathetic, even at vDIMM as low as 1.56v while the overclocking of these modules just seems limitless at the 8-8-8 timing set. On the other hand, we unfortunately didn't get the performance we wanted at 7-7-6 or 6-6-5 from our second set of memory we received from Corsair while our first set absolutely knocked our socks off.

We really hoped that the price tag of $200/stick would eliminate the saying that "in overclocking, it always comes down to luck of the draw". That obviously wasn't the case. It turned out to be a good thing that our first kit had the odd heat sink application as we were able to find out there was some variation amongst the modules with the Dominator GTX2 label. There are no guarantees, but it is evident that even with the worse of two different sets of modules, we ended up with a very strong Elpida Hyper based set of memory and that is what you are definitely going to end up with when purchasing sticks from the Dominator GTX2 series.</p>
<b>Pros:</b>
  • Corsair Dominator heritage with DHX technology at the heart of the cooling
  • Aesthetics that are still at the very top when it comes to memory in our opinions
  • Excellent overclocking that we have come to expect from Elpida Hyper based DDR3 memory
  • We are confident in saying that these modules at 8-8-8 will outrun almost any X58 system's memory clocking capability
  • Sold in single modules, keeps things nice and simple, we like this very much


<b>Cons:</b>
  • The $200/stick price tag has to go first on this list
  • With this type of price premium, we wanted to see something substantially better in the overclocking...but were limited by the CPU in some cases.
  • There clearly is a bit of an issue with the heat sink application as one kit illustrated rather blatantly

<p style="text-align: justify;">We really wanted to give the Dominator GTX2 modules a Dam Good award but we simply can't justify the $200/stick price tag for memory that really didn't outperform any other top bin Elpida Hyper based modules that can cost $200 less than a set of three GTX2 modules. That much of a price difference is just too much for us without some kind of guaranteed benefit.</p>

<p style="text-align: justify;"><center><b><i>Hardware Canucks would like to thank Corsair for making this review possible and supplying the memory used today.</i></b></center></p>
 
Last edited by a moderator:
Status
Not open for further replies.
Top