Mushkin Redline Ascent 3x2GB PC3-12800 C6 Memory Review

[3oh6]

Mushkin Redline Ascent 3x2GB XP3-12800 C6 Memory Review

<CENTER>

Mushkin Redline Ascent 3x2GB XP3-12800 C6 Memory Review</CENTER>



Price: $226 USD Mushkin Online Store
Manufacturer Product Page: Mushkin, INC.
Manufacturer's Part Number: 998692
TechWiki Info: Mushkin Redline Ascent 3x2GB XP3-12800 C6
Warranty: Limited Lifetime Warranty



Mushkin is a well respected name amongst enthusiasts and veteran system builders, but the Mushkin crowd has always appeared to be a bit of a niche market. That niche group of Mushkin users seemed to shrink over the past couple product cycles with very little forum activity or hoopla around the Mushkin name. With low voltage i7 memory kits being all the rage, Mushkin decided it was time to change things up and bring to the track a familiar decal on the door, but something new under the hood.

The Rocky Mountain based memory manufacturer Mushkin has a following of avid users and fans like any other. They have a great support team with some recently added fresh faces that are working hard to get the Mushkin name back in the mouths of enthusiasts. The best way to accomplish this is to release some high end, high quality products. The name Redline is synonymous with Mushkin and its users. The Redline series of memory products has represented the best memory Mushkin has produced. Nothing has changed with the recent crop of Redline candidates and we would like to introduce one of those candidates to you right now.

Today we are going to be spending our afternoon with a quiet but subtly strong young fellow. He comes from a humble background without a lot of money. Despite this lack of fortune, he still has a strong heritage in his name, a heritage that goes back many generations of successful family members. He also brings something to the table with triple channel memory kits that no one else has done. This is of course a Cas Latency rating of 6 at a frequency of 800MHz or DDR3-1600. This young man's name is Mushkin, Mushkin Redline Ascent 3x2GB PC3-12800 6-7-6-18 to be more precise, also known as 998692. As we said, his family name of Redline goes back generations of high performing memory. The name Redline is also associated with overclocking and this is something we will definitely be testing over the course of the next few pages. So please, just sit back, relax and enjoy your time as we present the story of the Mushkin Redline Ascent 3x2GB PC3-12800 memory kit.
<CENTER>

</CENTER>

 

[3oh6]

Package & Memory Overview

Package & Memory Overview

<p style="text-align: justify;">We are about to have a look at these new Redline Ascent modules, but be warned, the images very much resemble the color of the actual modules. We have gone to great lengths to ensure white balance and color representation is accurate. These modules are definitely Redline, but the color isn't pure red, the photos are telling the truth with a hint of fuchsia in the color of the heat sinks.</p><center>

</center><p style="text-align: justify;">Mushkin has managed to retain the traditional molded plastic blister pack for presenting the Mushkin Redline Ascent memory. At first we were a bit surprised as we would have figured that three sticks of these beefy heat sinks would be too much weight for the plastic package but it turned out to hold the memory quite well. The memory fits in nice and tight and the package actually felt like it wanted to hold together. For shipping, the package will still need to be wrapped in bubble wrap though as there certainly is the possibility of the molded plastic opening up in the middle through flex. This could potentially release a module with ease as again, these Ascent heat sinks really add a lot of weight to each stick.</p><center>

</center><p style="text-align: justify;">As secure as the memory is in the plastic package, we still would prefer to see something more secure. OCZ has gone with molded plastic packages for its triple channel kits, but by placing them inside a box, OCZ prevents the molded plastic from opening. Corsair has done the same with its triple channel memory. We would like to see Mushkin do the same and evolve their package. With pretty much all memory being shipped to end users these days, the odds of someone receiving a kit with a module loose and scratched up may not be good, but at least plausible.</p><center>

</center><p style="text-align: justify;">If this is the first time seeing the Ascent heat sinks from Mushkin, it is a bit of an experience. Yes the modules do look as thick and heavy as they appear. When you first grab a stick of this memory it feels like you grabbed a lead weight. The vapor chambers on each side and large aluminum heat sinks definitely make these modules one of the heavier sticks of memory out there. We can clearly see a split in the heat sink at the top of the module in the middle as this is where the two pieces of the heat sink meet. Essentially, each side is its own piece with half of the cooling fins at the top being part of one side each.</p><center>

</center><p style="text-align: justify;">The two clips are nice and tight to the heat sinks providing ample pressure to each side of the modules. Along the sides of the modules run ribbed fins that increase surface area which helps to dissipate heat. From above, we can see four cooling fins running the length of the heat sink providing even more surface area for the heat to be transferred from the module to the ambient air. We can also see these modules are much thicker than a standard module. The height isn't too far out of spec, but the width is. In fact, Mushkin advises that these modules are not designed for six module use and users looking to use six sticks of this memory should consider the standard Redline heat sinks. We will of course be checking for this during the installation section.</p><center>

</center><p style="text-align: justify;">The first image above is outlining the fact that we have two solid pieces to these heat sinks. The top cooling fins closest to us in this photo are a solid piece with the other side of the heat sink. The side of the heat sink we can see is a single piece with the top cooling fins furthest away from us out of focus. The point is, these are actually solid pieces of aluminum that go from the IC surface, to the top cooling fins. This will ensure excellent thermal transfer as there are no joints, or connections unlike the latest Corsair Dominator memory modules.

The second photo is obviously a macro shot of the underside of the heat sinks. We won't be removing the heat sinks as Mushkin advised against such a task due to the thermal adhesive being designed to take IC's with it. We will show some detailed diagrams of the Ascent heat sink and give an idea of exactly what we are looking at, but in essence, the IC's are connected to a single large heat pipe that runs the length of the modules. This large, flat heat pipe is essentially a vapor chamber of sorts that absorbs the heat from the IC's and spreads it out the length of the chamber, transferring the heat to the other side where the aluminum heat sinks move the heat to the ambient air for dissipation. The Ascent heat sinks are definitely one of a kind and do a marvelous job in theory, but the one downside is the added overall thickness to the module.</p><center>

</center><p style="text-align: justify;">The markings on the edge of the memory indicate a Brainpower produced PCB with six layers. These PCB's are quite common on entry level to mid range DDR3 memory modules and have been used with good success as far back as September of 2007 that I can recall seeing them. In fact, these same PCB's were used in the last Mushkin Ascent memory we looked at here at Hardware Canucks back in July of last year. Having been used with modules that long and no major revisions or issues arising indicates a great track record for the PCB's used on the Mushkin Redline Ascent modules we are looking at today.</p>

 

[3oh6]

Specifications

Specifications

<p style="text-align: justify;">Obviously the photos we just looked at outline the primary specifications of the Redline Ascent memory we are looking at today. Here is a complete run down of the operating specifications for these modules from Mushkin.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/specs-1.png" alt=""></center><p style="text-align: justify;">We mentioned in the opening that these Mushkin Ascent modules, along with their standard Redline counterparts are the first triple channel DDR3 parts that have been released for operation at DDR3-1600 with a CAS Latency timing of 6. The overall timing set is 6-7-6-18 with the standard 1.65v for operation with the X58 chipset and Intel i7 processors. Mushkin, however, also has a 6-6-5-18 set of memory available on the market as well, at a substantially higher price of course. The real question that everyone in the forums is asking about these modules is, "what IC's do they have on them?". Unfortunately we are not going to be able to take the heat sinks off to tell. Mushkin, however, was quite forthcoming with the information with this quote:</p><p style="text-align: justify;">"...we used a new proprietary IC design made from a strange, glowing material gleaned from a meteorite that fell in the parking lot behind our office. The whole thing weighs about 300lb so I figure we can make a couple hundred more kits. It has a few minor side effects (headaches, nausea, zombification) but we figure the performance is worth the risks. We do recommend that users not taunt the modules, as this has been found to result in some...issues*.</p><p style="text-align: justify;">*We are looking for new employees for shipping/receiving."</p><p style="text-align: justify;">Well, we may not be closer to knowing what the ICs are, what we can do is say which IC's they are not, based on their shape.</p><center>

</center><p style="text-align: justify;">We already saw a pretty good photo of the IC's in the previous section when looking at the heat sinks underbelly. Through the use of macro photography, we get another here from the bottom of the memory. Under the heat sinks we can see a virtually square IC mounted to the PCB. This indicates that the memory is definitely not the much raved about Elpida Hyper memory that is on the high end PC3-16000 7-8-7 kits. Some speculation was that any CL6 PC3-12800 module was going to be this expensive new IC. The price of this kit is also an indication that it is not Elpida Hyper aside from the shape as Hyper have beveled edges to the ICs. These have perfectly straight edges with 90 degree corners. Our best guess at the IC's on these modules, based on the regular Redline memory of the same operating specifications, is that they are indeed Elpida; but DJ1108BBSE DJ-F. A couple of forum users were able to remove the regular Redline heat sinks and identify the IC's. The overclocking results will certainly be interesting for us as we have no prior experience with this IC, if in fact that is what are on these modules for ICs.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/996625/specs-5.png" alt=""></center><p style="text-align: justify;">We really don't have a great grasp on the eVCI technology behind the Ascent heat sinks, so we have presented a diagram from the Mushkin site that outlines just what is going on inside the long flat heat pipes behind the Ascent heat sinks. Again, our best description of the process is similar to that of a vapor chamber. The interior of the heat pipes simply act as a distribution channel to evenly spread the heat out over a larger surface area, thus allowing dissipation of the heat to be easier. The heat pipe simply takes the heat from the relatively small surface area of an IC and spreads it out over a larger surface area that is then transferred more efficiently to the outside heat sink for dissipation into the ambient air. In theory it makes sense, but at these low voltages, heat might not be the biggest determining factor of a modules performance. As with any heat sink on memory, we question the validity of it even being there. No matter how you look at it, they are definitely cool looking and beefy modules.</p>

 

[3oh6]

Memory Installation & Test Setup

Memory Installation & Test Setup

<p style="text-align: justify;">We already know Mushkin has said these heat sinks are not designed for 6 module use, but they also say they only run 800MHz at 6-7-6 timings...and we all know we are going to prove that wrong. We also want to check and see how these modules do with the wing span of the Thermalright Ultra-120 eXtreme. In recent motherboard reviews, we have found the TRUE to cause problems with the Corsair Dominator heat sinks when in the first slot due to height. The Mushkin Ascent heat sinks are taller than the memory PCB, but are they going to be too tall?</p><center>

</center><p style="text-align: justify;">The answer is definitely a no. The Mushkin Ascent heat sink has absolutely no problem fitting under the cooling fins of a Thermalright Ultra-120 or its variants. We have plenty of room to play and even install the modules with the heat sink in place on the EVGA X58 3X SLI motherboard. The real question was whether or not we could manage to get 6 modules installed and as the second photo outlines, that is going to be impossible on these i7 X58 motherboards. The DIMM slots are just too tight to save space having a second pair added from Intel platforms of past. This is definitely something to keep in mind when deciding to buy the Redline Ascent or simply the standard Redline memory modules as upgrading to 12GB's will not be possible with the Ascent heat sinks.</p><center>

</center><p style="text-align: justify;">Despite the fact that these modules should fit in pretty much any motherboard under a TRUE because they are not too tall, we still thought we should throw the sticks in another motherboard. We happened to have the Gigabyte GA-EX58-UD4P laying around so we mounted the TRUE and installed the Mushkin Redline Ascent's. At first everything looked to be working well, then we noticed something and took the second photo. The heat pipe on the TRUE is actually contacting the top of the Ascent heat sink. This is more a by-product of the motherboard layout more than anything as you can see the DIMM slot is extremely close to the CPU socket. Also keep in mind that the outside slots are the primary memory slots so the odds of using a single kit in the blue slots is rather slim...but still something to keep in mind when looking at these modules.</p><center><table><tr><td>

</td><td>

</td><td>

</td></tr></table></center><p style="text-align: justify;">A quick check from the XMP profiles on the Gigabyte motherboard provide an interesting result with everything loading up fine, but tRAS showing up as 20 when CPU-Z and CPU Tweaker clearly identify the proper tRAS of 18. On the EVGA X58 3X SLI, the XMP profile came up perfect with no anomalies. Needless to say it is clearly the Gigabyte board having a little fun with the XMP profiles. Overall these modules are pretty effortless to work with. The XMP profile asks for 1.65v for vDIMM and 1.40v for VTT which is definitely enough for any CPU to be able to run these modules at their rated frequency. At DDR3-1600, we didn't find 6-7-6 timings to really push the memory controller so we feel most if not all users really shouldn't have any issue running these modules at specification.</p>

Test Setup

<center><table cellpadding="0" cellspacing="0" width="735px"><tr><td align="left" colspan="2">

</td><td align="right" colspan="2">

</td></tr></table><br /><table border="0" bgcolor="#666666" cellpadding="5" cellspacing="1" width="735px"><tr><td colspan="4"><b><font color="#ffffff">Test Platform:</font></b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Memory:</b></td><td align="left" bgcolor="#ececec" width="75%"><b>Mushkin Redline Ascent 3x2GB PC3-12800 6-7-6-18 (998692)</b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Motherboard:</b></td><td align="left" bgcolor="#ececec" width="75%">EVGA X58 3X SLI</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor:</b></td><td align="left" bgcolor="#ececec" width="75%">Intel Core i7 965 Extreme Edition<br>Intel Core i7 920 D0</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Thermalright Ultra-120 eXtreme-1366<br>2 x Scythe Ultra Kaze 120MM 3000RPM 133.6CFM (DFS123812H-3000)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Thermal Paste:</b></td><td align="left" bgcolor="#ececec" width="75%">Arctic Cooling MX-2</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>North Bridge Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>South Bridge 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>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%">Gigabyte GTX260OC 216SP (NVIDIA GeForce 185.85 WHQL)</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 Vista SP1 (with all updates)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Ambient Temperature:</b></td><td align="left" bgcolor="#ececec" width="75%">23C ~ 25C</td></tr></table></center><p style="text-align: justify;">We have made a couple changes to the testing platform from the last memory review. Instead of running the DFI X58-T3eH8 or the EVGA X58 3XSLI Classified, we are running the EVGA X58 3X SLI motherboard. The Classified will make an appearance a little later on as it is primarily a sub-zero benching board these days, but for all of the stability overclocking and benchmarks, the EVGA X58 3X SLI is going to be running the show. We have also downgraded the GPU to a single GTX260 from a GTX295. No real reasoning behind this aside from perhaps to see if the 3D benchmarks have any influence by a more typical video card being used instead of the ultra high-end GTX295.

The last item of note is the fact that we used two processors during this review. Typically we like to stick to one, but our 965 decided to drop a memory ratio mid-way through the review process. This meant that after overclocking the memory with the i7 965 processor, we wouldn't be able to run the XMP profile any longer for the benchmarks or the XMP stability testing as we lost the 2:12 memory ratio. This is where the i7 920 processor stepped in. In the stability testing you will notice the XMP profile testing on the EVGA X58 3X SLI motherboard is done with the 920, while the rest of the overclocking is done with the 965. Keep in mind, however, that all of the benchmarks will be completed with the i7 920 processor.</p>

 

[3oh6]

Stability Testing & Overclocking

Stability Testing & Overclocking

<p style="text-align: justify;">Hopefully you have all heard my spiel before about memory stability, but in case you haven't, here it is again in all its boring glory.</p><b>Stability Testing Methodology</b><p style="text-align: justify;">Memory stability, what constitutes stable? What is not considered stable? These questions get hotly debated in enthusiast forums all over the internet like little brush fires on the fringe of an inferno that play havoc with forest fire crews. Everyone has their own opinion about stability, especially when it comes to memory stability. For some, stable means they can do whatever it is on their computer without it crashing, blue-screening, or restarting; whether that means gaming or just surfing the internet. To this user, stable means simply using the computer as they normally would.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/setup-5.jpg" alt="Corsair Dominator 3x2GB PC3-12800 C8"></center><p style="text-align: justify;">Of course, that is not what we would call stable and we do a lot more thorough testing before labeling a memory frequency/timing stable. One of the toughest types of programs on a system has been found to be distributed computing projects such as Rosetta@Home, Folding@Home, World Community Grid, and more. Running 24/7 crunching for one of these great causes is a sure way to find holes in a system if there is truly some instability, unfortunately it takes a considerable amount of time to use them for stability testing so we use the list of programs below to all but guarantee the system to be 24/7 distributed computing stable:</p>

• 8 threads of 32M SuperPi Mod 1.5 (ran at the same time using Hyper Pi 0.99b)
• 2 hours of 8 threads of Prime95 using Prime95 v25.7 Stress Testing Blend
• 2 hours of HCI MemTest Pro in Windows using all available memory
• 6 loops of 3DMark Vantage ran back to back

<p style="text-align: justify;">As you can see, stability testing can mean any number of things to any number of people. We obviously aren't going to satisfy everyone's needs and some we are going to over satisfy with our testing methodology. The one item we haven't discussed is the role that the processor plays in memory clocking. With the i7 memory controller being on the CPU die, the strength of said memory controller is going to determine the strength of a memory overclock. With this kit of Mushkin Redline Ascent memory where tight timings are the primary role, the memory controller's strength shouldn't come into play as higher frequencies tend to be what really pushes the CPU. Needless to say, we don't anticipate any problems with this memory at the stock clocks despite the extremely tight timings.</p>

<b>Specification Stability Testing</b>
<p style="text-align: justify;">Before we get to any overclocking, we first want to quickly test this memory at the specified frequency and timings. With the included XMP profiles, this was a breeze on the EVGA X58 3X SLI motherboard. We simply went into the BIOS and enabled the XMP profile. We zipped through the stability testing and quickly tried the stock settings with a little lower voltage as we knew the motherboard was already under-volting the memory. We were pleasantly surprised by what we found, blown away to be honest. To double check our findings to ensure nothing strange was going on, we moved the kit over to the Classified and tested the XMP profiles there, but with the reduced voltage we were finding stable on the regular EVGA X58 board. Have a look at both screen shots, pay particular attention to the voltages listed in E-LEET. These voltages are what is set in the BIOS, and don't forget, the regular X58 3X SLI under volts memory under load by 0.05v.</p>
Click for full size screenshot...
<center><a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/stable_spec-1.png" target="_blank">

</a> <a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/stable_spec-2.png" target="_blank">

</a></center><p style="text-align: justify;">Now we understand that this kit of memory is binned for DDR3-1600 operation at 6-7-6-18 with 1.65v, but being able to run through our entire suite of stability testing at just 1.45v and 1.50v is absolutely incredible in our opinions. Memory guru's all agree that DDR3-1600 6-7-6 with only 1.65v is quite impressive, wait until they see the memory run at much less voltage than that. Obviously this is just a single kit but even still. Please don't hesitate to join us in the forums to show off your kits of this memory at stock settings with low voltage.

XMP testing on both EVGA boards obviously went as easy as could be. Mushkin has a good XMP profile that sets VTT at 1.4v which isn't too high, but definitely more than enough for any setup. We really don't anticipate any issues for users of this kit simply looking to run the XMP profile. With the knowledge that stock settings run fine at 1.50v, we couldn't wait to get this kit clocking, let's take a look at the results.</p>

<b>Stability Overclocking:</b><p style="text-align: justify;">With such tight timings at a relatively low voltage, we weren't quite sure how this kit was going to handle higher clocks and looser timings. We also weren't sure how far this kit would go with the tight timings of 6-7-6, but based on the stock clocks at 1.50v testing we did, we were certain we had some head room. Needless to say, we have a bit of a mixed bag of results on our hands. There is plenty to explain so let's start with the numbers and go through them one at a time.</p><center>

</center><p style="text-align: justify;">This is definitely one of the more odd kits we have looked at here at Hardware Canucks. We start off with the overclocking of the rated timings and voltage. Keeping in mind that our EVGA X58 3X SLI droops vDIMM under load by 0.05v, we set voltage to 1.70v so load vDIMM was right at 1.65v. With the specified timings of 6-7-6-18 1T, we managed a very impressive overclock of 868MHz or DDR3-1736. This was completely expected as we found the XMP profile to run easily at 1.50v. What caught us off guard was the fact that adding voltage, did absolutely nothing for clocking at 6-7-6. So despite the impressive overclock at 1.65v, the overall clocking at this timing set was disappointing because voltage didn't help this kit. We quickly found out what was limiting a frequency increase, tRCD.

You will notice the grey bar, our second timing set, having only a slight variation in timing changes for a very substantial frequency increase. Loosening up tRCD to 8 instead of 7, lifted the wall at 868MHz that we found and let us run this kit right up to 963MHz or DDR3-1926, at 6-8-6-20 1T! We have to admit that this is the first time we have run any memory at 6-8-6, but 963MHz is a very high frequency to be running CAS 6 through our stability testing. Clearly these IC's have some distinct characteristics. They have an almost brick wall at a certain frequency at 6-7-6 timings, but simply opening up tRCD to 8 unleashes some incredible head room for a rather well priced kit of memory.

Our overclocking headroom slowly dried up as timings were further loosened but we did manage to hit 1000MHz or DDR3-2000 stable at 8-9-8. Although this is nice, the sweet spot for this memory is in the 950MHz range with timings of 6-8-6. We are actually quite anxious to start seeing others users results to determine if our kit was simply a bit of a freak at 6-8-6, or if other kits will find similar walls at 6-7-6 broken down by opening up tRCD to 8. Overall this Mushkin Redline Ascent memory caught us off guard with this type of scaling, and lack thereof in a sense.</p>
Click for full size screenshots...
<center><a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/oc-2.png" target="_blank">

</a> <a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/oc-3.png" target="_blank">

</a> <a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/oc-4.png" target="_blank">

</a> <a href="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/oc-5.png" target="_blank">

</a></center><p style="text-align: justify;">So if you do grab a set of this memory, or the standard Redline modules of the same specifications, be sure to stop by our memory section in the forums for the discussion thread on this memory to share your results.</p>

 

[3oh6]

Benchmark Methodology

Benchmark Methodology

<p style="text-align: justify;">With an XMP profile alive and kicking, we will have no problem running this kit at the same settings anyone else will be able to with a motherboard that works with the XMP profile.<img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/method-1.png" alt="Just a sample graph outlining where the results will be coming from in the up-coming benchmarks" style="float: right; margin: 4px 0px 4px 7px"> This will give us a truly "stock" set of numbers for a very typical setup that consists of an i7 920 processor and single GTX 260 video card all at stock settings. As far as build budgets go, this memory fits right into that type of build and should be a fairly useful set of numbers for people to compare with. These <b>stock</b> results will appear in <b>blue</b> in our result graphs. The other three results will be a mixture of our stable overclocks with this memory, with one particular comparison being the focus.

The <b>grey</b> set of results are going to be our overclocked setup at the rated timings of 6-7-6-18 and rated voltage of 1.65v. The best part of this overclock is that with the CPU multiplier at 21X, we get a CPU frequency of 3645MHz. This is good because our overclocked setup at 6-8-6-20 which reached that impressive 963MHz (DDR3-1926) can run almost the same CPU frequency with a 19X CPU multiplier. These results will be the first <b>red</b> bars in the charts. This is going to give us a chance to directly compare the benefit of running DDR3-1736 at 6-7-6 to DDR3-1926 at 6-8-6. We have obviously never done this comparison before so we are excited to see the results.

The second set of <b>red</b> results will simply be our highest overclock of DDR3-2000 at 8-9-8-24. We will drop the CPU multiplier to 18X so we will have the CPU running at 3600MHz. Again, we are simply trying to get some interesting comparisons between the various memory overclocks we were able to achieve at similar CPU clocks. Here is a rundown of how we setup the OS for each set of results to provide an even playing field.

1. Windows Vista x64 w/SP1 is installed using a full format
2. Intel Chipset drivers and accessory hardware drivers (audio, network, GPU) are installed followed by a defragment and a reboot
3. At time of benchmarks the latest drivers were downloaded from their official web sites as the latest drivers, most notable, NVIDIA GeForce Release 185.85 WHQL
4. Programs and games are then installed followed by another defragment
5. Windows updates are then completed installing all available updates followed by a defragment
6. Benchmarks are each ran three times after a clean reboot for every iteration of the benchmark unless otherwise stated, the results are then averaged

</p>

 

[3oh6]

Memory Benchmarks

Memory Benchmarks

Everest Ultimate v4.60<p style="text-align: justify;"><i>Everest Ultimate is the most useful tool for any and all bench markers or overclockers. With the ability to read most voltage, temperature, and fan sensors on almost every motherboard available, Everest provides the ability to customize the outputs in a number of forms for display on your desktop. In addition to this, the memory benchmarking provides a useful tool of measuring the changes to your memory sub-system when tweaking to measure the differences. Unfortunately with the i7 processors, the results aren't always consistent and we can receive variations as much as 1000MB/s at any given time. Because of this we use multiple runs and drop any of the "high" scores from the averages.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/mem_bench-1.png" alt=""></center><p style="text-align: justify;">We are off to an interesting start here in the pure bandwidth section. The last three configurations look like they are going to give us some interesting results throughout the benchmarking pages. Our 6-8-6 setup appears to be the overall winner with the DDR3-2000 8-9-8 configuration coming in a close second. It is always a little too early to tell how things are going to pan out after just the Everest bandwidth numbers but our prediction is this pattern will continue throughout testing.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/mem_bench-2.png" alt=""></center><p style="text-align: justify;">Lavalys Everest latency test clearly favors memory frequency because even with the lower CPU frequency, the DDR3-2000 setup takes a decisive win here. Overall though, the last three configurations all show very good latency here.</p>

SiSoft Sandra 2009.SP2<p style="text-align: justify;"><i>SiSoft Sandra is a popular and well used benchmark in the industry but not really a friend of serious benchmarkers. The results SiSoft Sandra produces have been suspect at times basing the numbers it comes up with on system specs and not actual testing. The latest version of Sandra seems to be one of the few programs that appear to calculate memory bandwidth consistently so we decided to include it in today’s benchmarks. Like we have always said with SiSoft Sandra though, take these results for what they are and nothing more.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/mem_bench-3.png" alt=""></center><p style="text-align: justify;">SiSoft Sandra looks to be playing the frequency game with perfect scaling based on the speed of the memory in the setups disregarding any kind of timings. At the same time, the latency numbers seem to disagree.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/mem_bench-4.png" alt=""></center><p style="text-align: justify;">As we can see, the perfect scaling seems to have stopped with the DDR3-2000 setup coming in a nano second behind the 6-8-6 setup. Obviously the latency test in Sandra weights memory timings a little heavier than the bandwidth test.</p>

ScienceMark v2<p style="text-align: justify;"><i>ScienceMark is an almost ancient benchmark utility at this point in time and hasn't seen an update in a long time. It is, however, still a favorite for accurately calculating bandwidth on even the newest chipsets.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/mem_bench-5.png" alt=""></center><p style="text-align: justify;">ScienceMark again plays the definitive role in the bandwidth game and shows that while the DDR3-2000 setup does come out on top, it isn't by much. From the results we've seen here, we have to say that the 6-8-6 setup really has disappointed in its performance. With a bit of work on the secondary timings and some RTL adjustments, this timing set could really scream. We still can't get over the fact that we are running 100% stable at CL6 with only 1.725v under load with this kit at these kind of clocks. If you asked me if I expected to get close to 1000MHz at CL6 with this kit when I first got it I would have laughed...laughed hard.</p>

 

[3oh6]

System Benchmarks

System Benchmarks

SuperPi Mod v1.5<p style="text-align: justify;"><i>When running the 32M benchmark of SPi, we are calculating Pi to 32 million digits and timing the process. Obviously more CPU power helps in this intense calculation, but the memory sub-system also plays an important role, as does the operating system. SPi 32M has been a favorite amongst benchmarks for these very reasons and is admittedly the favorite benchmark of this reviewer.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-1.png" alt=""></center><p style="text-align: justify;">Well look what we have here, a solid 7 second drop going from DDR3-1736 6-7-6 to DDR3-1924 6-8-6. We were expecting those two setups to be neck and neck through 32M but it turns out that tRCD of 8 didn't hurt performance as much as we expected and the higher memory clocks cleaned the floor of the 6-7-6 setup. We can also see that the lower CPU clock of 3600MHz didn't hold back the DDR3-2000 8-9-8 result that much. It still managed to come within a whisker of the 6-7-6 setup, well maybe not a whisker, but two seconds isn't too bad considering it was almost 50MHz slower on CPU clocks.</p>

PCMark Vantage<p style="text-align: justify;"><i>The latest iteration of the popular system benchmark is PCMark Vantage from the Futuremark crew. The PCMark series has always been a great way to either test specific areas of a system or to get a general over view of how your system is performing. For our results, we simply run the memory benchmark suite which involves a wide range of tests on primarily associated with media management and testing.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-2.png" alt=""></center><p style="text-align: justify;">This was the first time we ran the memory suite in PCMark Vantage. Normally we run the standard full suite but figured, this is memory we are testing, why not just test the memory? The results really don't mean too much too us aside from the fact that the test likes memory frequency. We see a pattern of results very similar to that of SiSoft Sandra bandwidth.</p>

DivX Converter v7.1<p style="text-align: justify;"><i>Next up is a real life benchmark where we simply time a common task done on the computer. Encoding DVDs for viewing on the computer or other devices is an increasingly important task that the personal computer has taken on. We will take a VOB rip of the movie Office Space, and convert it into DivX using the default 720P setting of the new DivX converter v7.1.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-3.png" alt=""></center><p style="text-align: justify;">Obviously CPU clocks mean more in DivX encoding than memory frequency. Our last setup at DDR3-2000 gets beat pretty badly by the two CL6 setups, and it is obviously from the lower CPU frequency. The higher memory frequency of the 6-8-6 setup did manage to gain 17 seconds over the 6-7-6 setup, but that is 17 seconds over the course of a 21 minute task which equates to about a 1.4% gain in performance. If you are doing a lot of DivX encoding, don't look at memory frequency for performance gains.</p>

Lame Front End<p style="text-align: justify;"><i>Un-like the DivX conversion we just looked at, Lame Front End is not multi-threaded and only utilizes a single core of a processor. This will obviously limit performance but we should still recognize significant time savings going from the stock settings to the overclocked results. We will be encoding a WAV rip of the Blackalicious album, Blazing Arrow and converting it to MP3 using the VBR 0 quality preset.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-4.png" alt=""></center><p style="text-align: justify;">Lame Front End shows less gains from higher memory frequency than DivX encoding did. Our two CL6 setups are a dead heat, and the DDR3-2000 setup again falls a couple seconds off the pace, because of the lower CPU clock.</p>

Photoshop CS4<p style="text-align: justify;"><i>Adobe Photoshop CS4 is fully x64 compliant and ready and able to use every single CPU cycle our processor has available including the implementation of GPU support utilizing the GTX 280 in our test system. It is just a shame it can't fully utilize all 8 threads of the i7 processor yet. We have changed our Photoshop benchmark to more of a standardized test configured by DriverHeaven.net. Their Photoshop benchmark utilizes 15 filters and effects on an uncompressed 109MB .JPG image that will test not only the CPU but also the memory subsystem of our test bench. Each portion of the benchmark is timed and added together for a final time that is compared below.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-5.png" alt=""></center>

WinRAR 3.80<p style="text-align: justify;"><i>We all know what WinRAR is and does. It is a compression and decompression tool that has a built in benchmark, a way to tell just how fast a system can do this programs given task. We simply run the benchmark up to 500MB processed and time how long it takes.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/sys_bench-6.png" alt=""></center><p style="text-align: justify;">These last couple of benchmarks are showing the same exact pattern that we saw in SPi 32M as well as DivX and Lame Front End. In most if not all of the real world benchmarks, CPU frequency just plays more of a role than memory clocks or timings. Sure the higher frequency of memory helps, but it can't even make up for a 50MHz difference in CPU frequency. We can also see that tighter timings really don't take much to overcome with memory frequency on the i7 platform.

The setups we looked at today were definitely interesting for us, and they also provided some valuable information that we haven't seen before. I mean, we have seen the same type of factors influence results, but these setups gave us another perspective to them. Overall, we are quite pleased with how things turned out and pleased with our choice in configurations to compare.</p>

 

[3oh6]

3D/Gaming Benchmarks

3D/Gaming Benchmarks

Futuremark 3DMark Vantage / 06<p style="text-align: justify;"><i>We have forced ourselves to step up to 3DMark Vantage results for all reviews because the public demands it. 3DMark Vantage is the newest in a long line of 3D benchmarking software from Futuremark and is the most elaborate to date. Featuring multiple presets for various system configurations, Vantage is the culmination of all 3DMarks past relying on system and GPU power for its results. We will stick to the Performance preset as it seems to be the most popular at this point in time. 3DMark 06 is the previous iteration of this successful 3D benchmark suite.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/3d_bench-1.png" alt=""></center><p style="text-align: justify;">#d and gaming benchmarks really are a bit of a wash in memory reviews but we still like to show them, just in case it is your first time reading a Hardware Canucks review. Basically, you are going to find little to absolutely no effect on games with high-end memory, such as the Mushkin Redline Ascent modules we are looking at today. Even at the stock settings, these modules provide more than enough bandwidth that increasing the frequency is not going to improve gaming performance. The 3DMark 06 and Vantage results show this, as will the rest of the gaming benchmarks on this page.</p>

Crysis - Sphere benchmark<p style="text-align: justify;"><i>We all know what Crysis is and how much it beats up systems but we wanted to add it to the gaming benchmarks to see how system changes can improve performance on a mid-level system. Detail levels are all set to Very High with the resolution at 1680x1050 with no AA. We ran the benchmarks with a demo of the Sphere level in DX9 and 64-bit. The game looks great with this setup and plays just well enough to keep us happy.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/3d_bench-2.png" alt=""></center><p style="text-align: justify;">Crysis is so GPU bound that even the CPU frequency jumps we make between the various setups doesn't influence results all that much. These are the same type of results we see with a GTX295 as well so it isn't just the GTX260 limiting the influence the system has on performance.</p>

FarCry 2<p style="text-align: justify;"><i>Another fall release of this past silly season Far Cry 2 has some beautiful scenery but does lack that buttery smooth game play in places. A lot of moaning and groaning has occurred with FarCry 2 but acceptable frame rates are much easier to achieve than Crysis and the game play is plenty smooth enough to enjoy. We were really able to crank up the settings with this benchmark on this setup.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/3d_bench-3.png" alt=""></center><p style="text-align: justify;">Far Cry 2 usually has some impact on its frame rates by adjusting the system settings but clearly that isn't the case with a single GTX260. Our average frame rates are all within 2 FPS of each other, and the min/max FPS are almost identical across the board.</p>

Left 4 Dead<p style="text-align: justify;"><i>The newest game in our testing sweet, Left 4 Dead was just added after we were asked to include a Source powered game in our memory benchmarks. Being based on the Source engine, there is definitely a chance that system performance will heavily influence the results. We used FRAPs to measure frame per second on a custom time demo of the rooftop level.</i></p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/3d_bench-4.png" alt=""></center><p style="text-align: justify;">Our most system influenced game on the schedule, Left 4 Dead, doesn't even bat an eye at the changes we make to CPU or memory frequency. As with the other games, we basically have results identical through all four setups.</p>

 

[3oh6]

Extreme Overclocking & Benching

Extreme Overclocking & Benching

<center><table cellpadding="0" cellspacing="0" width="735px"><tr><td align="left" colspan="2">

</td><td align="right" colspan="2">

</td></tr></table><br /><table border="0" bgcolor="#666666" cellpadding="5" cellspacing="1" width="735px"><tr><td colspan="4"><b><font color="#ffffff">Test Platform:</font></b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Memory:</b></td><td align="left" bgcolor="#ececec" width="75%"><b>Mushkin Redline Ascent 3x2GB PC3-12800 6-7-6-18 (998692)</b></td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Motherboard:</b></td><td align="left" bgcolor="#ececec" width="75%">EVGA X58 3X SLI Classified</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor:</b></td><td align="left" bgcolor="#ececec" width="75%">Intel i7 920 D0 (3845B026)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Processor Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">MMouse Rev 3 CU LN2/Dry Ice Pot<br>w/Liquid Nitrogen</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>North Bridge Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">Stock</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>South Bridge 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>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%">Gigabyte 9800GT OC 1GB (NVIDIA GeForce 185.85 WHQL)</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Video Card Cooling:</b></td><td align="left" bgcolor="#ececec" width="75%">K|ngp|n Tek9 4.0<br>w/Liquid Nitrogen</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 3000RPM 133.6CFM (DFS123812H-3000)</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 Vista SP1 / Windows XP Pro SP2</td></tr><tr><td align="center" bgcolor="#cc9999" width="25%"><b>Ambient Temperature:</b></td><td align="left" bgcolor="#ececec" width="75%">23C ~ 25C</td></tr></table></center><p style="text-align: justify;">We happened to have a Gigabyte 9800GT OC 1GB card laying around and decided we should put it through its paces under sub-zero cooling. What a better way to bench with a kit of memory than to take a rather pedestrian video card and turn them into the monster of the mid-range with the help of a little liquid nitrogen, AKA LN2.

Photos of the setup are above for your browsing, along with a rundown of the complete setup used here. We have a couple substitutions from the standard testing we just finished looking at. You can also see we prefer kneaded rubber eraser for our insulation method. It has worked wonders from day one on both the motherboard and video cards alike. Feel free to ask any questions you have about this type of benching or the setup used in our discussion thread in the forums. For now, let's just let the screen shots do the talking and keep the chatter to a minimum. We start off with the 2D benching first, then onto single card 3D benchmarks...enjoy.</p><table align="center" bgcolor="#666666" cellpadding="10" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="50%">SuperPi 1M
click for full size...
</td><td align="center" valign="top" bgcolor="#ececec" width="50%">SuperPi 32M
click for full size...
</td></tr></table><p style="text-align: justify;">The key component in 32M SPi on the i7 platform is the memory. We saw the effects in the standard benching, and while there are no comparisons here, the Mushkin held up it's end of the bargain in this 32M run. As it turned out, I managed a personal best and improved on my previous i7 920 WR 32M time. The memory plays a vital role in a time like this and the Mushkin Redline Ascent really pushed the limits running a wild 250 base clock and 1000MHz 7-8-7-20 1T rather easily, only requiring 1.77v for vDIMM. This is quite a respectable number, especially for 32M SPi. The 1M isn't too bad either, a little low on efficiency but the clocks are nice. These were the only 2D benchmarks I wanted to run as the primary focus of the benching today is the 3D results with the 9800GT, so let's take one i7 920 WR and have a look at the initial single card runs next.</p><center><img src="http://images.hardwarecanucks.com/image/3oh6/mushkin/998692/xsetup-9.jpg" alt=""></center><p style="text-align: justify;">The memory overclocks for these 3D results aren't quite as impressive as we just saw, but still another very solid showing by the Mushkin Redline Ascents. I wanted to run 6-8-6, which we already know the Ascents are more than capable of, but the motherboard absolutely wouldn't run those tight timings at the high base clock I was running at with the i7 920. It actually seems to be a CPU limit of sorts that I haven't figured out yet, so 7-8-7 timings will have to do.</p><table align="center" bgcolor="#666666" cellpadding="10" cellspacing="1" width="90%"><tr><td align="center" valign="top" bgcolor="#ececec" width="50%">3DMark 03
click for full size...
</td><td align="center" valign="top" bgcolor="#ececec" width="50%">3DMark 05
click for full size...
</td></tr><tr><td align="center" valign="top" bgcolor="#ececec" width="50%">3Dmark 06
click for full size...
</td><td align="center" valign="top" bgcolor="#ececec" colspan="2" width="50%">3Dmark Vantage
click for full size...
</td></tr></table><p style="text-align: justify;">Some of the above results are tops at HWBot.org, some aren't despite higher clocks. Unfortunately there is a fair bit of cheating going on over there right now. Some 3D tweaks that aren't allowed get used on these mid-range cards because Futuremark ORB links aren't required. Either way, almost 20K for a single 9800GT is mighty impressive in 06, and with an 05 run like that, I think we can call it a good day.</p>