Mushkin HP3 10666 2x2GB Review

ccokeman - 2008-08-06 07:07:30 in Memory
Category: Memory
Reviewed by: ccokeman   
Reviewed on: November 10, 2008
Price: $182.99


While two gigabytes of memory was sufficient enough to run your Windows XP based computer, two gigabytes just is not quite enough for Windows Vista. So what do you do to fix the problems with slow loading programs and having those programs constantly paging out to the swap file on your hard drive? More system memory is usually one of the solutions. With DDR2, the costs involved really make this a no-brainer. I have seen standard 4 gigabyte kits available as low as 50 bucks after rebate if you can catch a deal. More realistically, 60 dollars is about the low point on DDR2. DDR3 on the other hand is not quite that inexpensive. A nice performance 2 x 1 GB set with decent timings will run you around $150 and up. A 4GB set of performance DDR3 will still take a chunk out of your computer budget. But you have to pay to play ,the question is how fast do you want to go?

With the cost of performance DDR3 coming down as its use becomes more widespread, it was time we showed what a 4GB set of DDR3 can do. The Mushkin HP3 10666 modules are part number 995562 and includes two, two gigabyte modules with timings of 7-7-7-20 at a frequency of 1333MHz(666FSB) using just 1.7 to 1.8 volts.

Closer Look:

If you have never seen Mushkin packaging before, it comes without all the flash that some kits carry. The logo is simple and understated and rides right under your radar. The front carries the modules in plain sight with the "Get More" slogan as well as the company logo. The rear view carries the installation instructions and some troubleshooting tips, should you have a problem with the modules.



Part number 995562 is the Mushkin HP3 10666 DDR3 2x2 gigabyte set of memory. This kit is designed to run at a frequency of 1333MHz with latencies of 7-7-7-20 with a modest 1.7 to 1.8 volts to the modules. The modules use Mushkin's proprietary Frostbyte Technology heatspreaders much the same as most of the Mushkin performance lineup. The heatspreaders are blue signifying that these modules are part of the HP line up. These modules are not rated for the extreme speeds and voltages of the Ascent lineup, so the Frostbyte technology will do just fine.



The Frostbyte heatspreaders had a little movement in them so i decided to try and slide them off, just to see what was underneath the blue anodized aluminum. The modules used for this set of memory are made by Elpida and carry the part number J1108BASE -DJ-E. After trolling Elpida's website, I found the documentation on these module. They are rated for 1333MHz at 1.5v with timings of 9-9-9. The increase in voltage to 1.7-1.8 is needed to run these modules at latencies of 7-7-7 at 1333MHz.



Now that we know a little more about the modules, I think it's high time to find out what kind of performance they offer.




Part number
2 x 2GB 996652
1333 MHz
240 Pins



When it comes time to purchase your new memory modules, most people look to review sites to get a good idea on the performance capabilities of the memory they want to buy. Why, you ask? So they don't have to go through the endless buy it and return or sell it routine to find the set of their dreams. Hey we do it for you! How? By testing the memory with a series of benchmarks that show some of the capabilities of the system memory. Synthetic benchmarks as well as real gameplay are used to show the capabilities. Also, there are comparisons to other performance modules just so that this is not a one-sided affair. That just would not do and offers up only the knowledge of what the featured product can do.

Testing Setup:

Comparison Modules:


CPU-Z: This application shows us the settings that we have chosen in the BIOS. Items shown in this application include CPU speed and bus settings, motherboard manufacturer, BIOS revisions, memory timings, and SPD chip information.


Task Manager: We use this utility to show physical memory, kernel memory, page file, and processor usage.



Overclocked settings:

Trying to get the most from these modules was an interesting experience. First off, the modules require 1.7 and 1.8 volts to work at the factory specified timings and clock speed of 1333MHz. Fair enough, decent timings at the baseline voltage threshold. But OCC is not about the baseline number it's a starting point. The memory chips on these modules carry a specification of 9-9-9 at 1333MHz 1.5 volts. Of course its going to take some more juice to get the tighter timings. When running memory benchmarks on an Nforce motherboard allows a couple of things you cant do with an Intel based board such as running the memory unlinked from the CPU allowing for really nice screen shots, but at the cost of reduced or no increase in performance. The best performance is gained by running linked and synced with the CPU FSB. That's what I have done here. Running at 9-8-8-20 unlinked I was able to push only to 787MHz (1574Mhz) at 1.85 volts. Not really the kind of overclock I was looking for especially running unlinked. So back to the drawing board. I reduced the timings to 9-9-9-24 P1 and P2 enabled and started from scratch. Higher volts on the northbridge, lower volts on the northbridge it all was for naught. Finally I went the clean slate approach and again started working my way up. 1400Mhz , no problem, 1500MHz, no problem, 1600MHz, a small hiccup, 1700 getting close after so more tweaking and finally the point I could go no further 867MHz 9-9-9-28 at 1.55 volts. Yes 1.55 volts! It shocked me to to see where I had to go to get the modules to overclock.



The benchmarks used in this review include the following:




PcMark Vantage: With this benchmark, I will be running the system suite, as well as the memory test suite. The measurement for the system suite will be the total score. The measurement for memory performance is the total memory score. A comparison will be made of the performance at DDR3 1333 for a comparison point, and the highest achievable speed for the Mushkin HP3 10666 2x2GB kit.








SiSoftware Sandra XII: In this program, I will be running the following benchmarks: Cache and Memory, Memory Bandwidth, and Memory Latency. All benchmarks will be at default timings. Higher is better in all tests, except for Memory Latency, where lower is better.






In the PCMark Vantage 1333MHz testing, the Mushkin HP3 modules did not finish at the top, but were far from the bottom of the list. In the Vantage memory testing though, it was a strong second, finishing behind only the CSX modules. In the Sandra testing, the scores at the 1333MHz level were just about identical, given that the speed and latencies were again almost identical. In the Cache and Memory testing, the HP3 modules bested all of the modules, except the offering from CSX.



Company of Heroes is a real time strategy game set during World War II. The object is to occupy and control the ground you capture, while forcing the opponents to capitulate. We will use the in-game performance test to measure the performance of the system.


The settings used in this test are listed below:













Higher is Better


The first two resolutions were really unfavorable towards the Mushkin HP3 set. Without being able to tighten the latencies I think a little something was left on the table. However at the top two resolutions, the performance matches that of the 2x1GB modules.



Mushkin's slogan is "Get More". Did I? Why, yes, yes I did. At stock speeds, the Mushkin modules consistently performed almost identically the same as most of the 2x1GB sets. This, while running looser latencies to do it. Most of the 2x1GB kits are able to reduce the latencies to 6-5-5-16 at 1333MHz, but I was unable to make this happen on the HP3 10666 2x2GB set. But in reality, the addition of two gigabytes of additional system memory makes this a non-issue for day-to-day use. 'Feel' is something that is hard to explain, but most of you know just what I am talking about. Programs respond that little bit quicker, games load just that little bit faster, all things that you can feel but are not easily measured. Adding more system memory does this for you. If you run Vista, many of you know the amount of resources that are eaten up by all the background services. Running with just two gigabytes, means that eventually the requested memory is going to have to be paged out to the swap file and this of course takes longer to respond than having the physical memory available.

When it came time to push these modules to find out where the performance bandwagon stopped, I was able to get the HP3 10666 modules to 1734MHz (867 MHz). This represents a 201MHz improvement over the stock speeds. But to do this, I had to reduce(yes reduce) the voltage to the modules, to 1.55 volts and loosen the latencies to 9-9-9-28 with a 2t command rate. This is in fact, better than I was able to get on the Mushkin HP3 10666 2 x 1 gigabyte kit by a large margin. Running at the Jedec specification for the modules is possible with the default 1.5 volts and 9-9-9 latencies, in fact this is where the best overclock occurred. When pushed, the Frostbyte heatspreaders do their job shedding the heat from the modules. Running a fan over any DDR3 modules, running greater than 1.5 to 1.6 volts, does come highly recommended to preserve the modules and prevent them from self-destructing from heat. When removing the modules, I did find that the heatspreader had loosened up from the modules, presenting me with an opportunity to see what was underneath the heatspreader. Having the heatspreader come loose is not typical of the Mushkin build quality. All in all, the Mushkin HP3 10666 2x2GB set of memory delivered very respectable performance and compared favorably against 2x1GB kits that ran at tighter latencies. The price point for this set is less than $200 bucks, so if you need a set of Mushkin quality memory, you really can't go wrong with the latest HP3 10666 2x2GB kit!