Print this article

Team Xtreem Dark 2x1GB PC2-6400


Date
: 01/30/07 – 12:56:40 AM

Author
:

Category
: Memory

Page 1 : Index

Manufacturer:
Team Group Inc.

Price:
$230 USD (TankGuys.biz) / $280 CND (NCIX.com) / $230USD MSRP

There doesn't seem to be any shortage of new DDRII memory packages being un-leashed on the market by a seemingly unlimited amount of memory manufacturers. Today I am happy to introduce one of those manufacturers and yet another new face to the pages of Overclockers Online, Team Group Incorporated. Team Group was established in 1994 in the city of Taipei, Taiwan and has grown to be a leading memory producer in Taiwan. Team Group has used that production strength to expand its reach to various regions of the globe including distribution centers in Taiwan, Hong Kong, and the USA with no expectations to stop there.

In addition to production of SDRAM memory modules, Team Group also manufactures USB thumb drives and Flash memory products. The emphasis of all Team Group products is focused on quality and esthetically pleasing packages. This focus on innovative and elegant designs can really be seen in their impressive line of flash drives ranging from the Fusion Drive Plus to their unique and secure, Finger Print Drive. Today, however, I will be examining a product from their latest line of enthusiast DDRII memory modules, the recently released Team Xtreem Dark.

Team Xtreem Dark is an answer for the enthusiast on a budget providing impressive performance and solid overclocking without taking too much out of your wallet. Team Xtreem Dark was developed to compliment the Team Group high-end line of memory, known simply as Team Xtreem. With recent prices in the high-end overclocking segment rising out of control due to a lack of diversity in available ICs, Team Group extensively tested all available options and were able to create Team Xtreem Dark, a product that possesses plenty of ability with a more reasonable price tag. I was informed just before posting this review that Team Group has lowered the MSRP of the Team Dark series to $230USD on the PC2-6400 kit that I will be reviewing today and $219USD on the PC2-5300 2GB kit. Hopefully NCIX will lower prices another $20 or so in Canada to reflect these changes in MSRP.

Page 2 : Package & Contents

The Team Xtreem line of memory comes with silver heatspreaders but Team Group apparently decided that a change had to be made for the Team Xtreem Dark heatspreader. I think the difference is pretty obvious right away.

Team Group uses a very compact and secure plastic package for their memory. This package is slightly smaller than the standard blister pack used with many other manufacturers and I think it holds the modules a little more securely. The label on the front is actually a sticker that wraps around the entire face and secures to the tray at the back on both sides.

The label on the package contains all the pertinent information about the modules inside. There are actually two stickers that make it very obvious what the memory contained inside is rated to run at. Notice the lower left hand corner, Lifetime Warranty. We will continue examining the modules themselves as the specifications section will cover all of this information in greater detail.

Now that is a spectacular blue heatspreader. This is the signature difference between Team Xtreem and Team Xtreem Dark as I mentioned earlier. The anodized blue aluminum heatspreader is definitely the focal point of this memory.

The package contains only the two modules and a small booklet that briefly discusses the Team Group lifetime warranty and recommended installation procedure of the modules.

The modules heatspreaders also have a sticker on them that outlines the specifications of the memory. This is basically the exact same sticker that was found on the package with the part number, operating frequency, timings, and rated voltage along with a bar code.

The backside of the memory is identical to the front with the Xtreem logo on the left hand side and a sticker on the right just under the Team logo. The backside sticker is the serial number of the kit of memory instead of the information sticker, like on the front.

The heatspreaders are nothing fancy aside form the bright blue and raised Xtreem logo. They are a simple aluminum heatspreader that is not closed at the top and doesn't really have a lot of weight to them. Only recently did Team Group start sending out Xtreem modules with heatspreaders attached. Previously they actually came separately and could be attached at the users discretion.

Here we can see that the heatspreader is attached to the ICs on the modules with standard thick rubber-like thermal tape. I have a couple sets of Team Xtreem memory I purchased previously that came without heatspreaders attached. The heatspreaders with those modules came supplied with 3M 467MP adhesive thermal tape. I can only assume that Team Group is still using this thermal adhesive for these modules.

As mentioned, the heatspreaders simply attach to each side and do not interact with each other leaving the top open. I always mention that I prefer an open top on memory modules with heatspreaders of this design as it will allow air from a fan to actively cool not only the ICs, but also the PCB when placed above the modules.

There is definitely going to be no complaints about the looks of these modules but if there was one thing that I would like to see, it is a black PCB. Could you imagine how trick a black PCB would look with the bright blue aluminum heatspreaders? Given the memory's name of Team Xtreem Dark, I think a darker PCB would be more than fitting.

Page 3 : Specifications

Timings, voltages, PCB, warranty, etc..all falls under one category, the specifications. There are numbers, acronyms, and other nuances of a memory specification sheet that can make the average user completely confused. For the enthusiast in the crowd, the specification sheet is the menu before a feast. I definitely consider myself to be of the latter group being an absolute connoisseur that loves a healthy portion of memory. Regardless of what category you are in, there are some values that you will want to see on this page that describe the capabilities of this memory. I start off with a short except from the Team Group web site:

Team Group is clearly looking to position the new Team Xtreem Dark series in-between their already established Team Xtreem line and standard, non-enthusiast, memory. Team Group introduces this memory as a good piece of hardware for the average or new overclocker which means that I should see some overhead in the overclocking section. Before we get to overclocking though, here is how this memory breaks down at the rated frequency and timings.

We already knew this kit consisted of PC2-6400 modules that are rated at 4-4-4-8 and we can see here that the rated voltage is 2.0v-2.2v which should give us some room to play. The other standout specifications are the 8-layer PCB which helps tremendously with stability when voltages are increased and of course, the lifetime warranty offered by Team Group. The other thing that I would like to point out is that this memory is available in 2x512MB kits in addition to the 2x1GB kit that I am reviewing today.

I was informed by Team Group that the ICs used on these modules are manufactured by Promos but I'll be damned if I can match the markings on this IC up to anything on the Promos web site. Googling the information provided on the IC doesn't help either coming up blank. Looks like all I have is specifications of the modules them self to go on.

Page 4 : Installation

The installation of these memory modules is extremely straight forward and with no added height or width to these modules from an excessive heatsink, I will not be test fitting them with large CPU coolers.

It is obvious that there is no added height from the heatspreaders and they barely add any width to these modules from the above image. Even in a setup where dual-channel memory slots are side by side, these modules will have no problem fitting in.

This is how the AM2 testing setup looks with the modules installed and running. Notice the frivolous use of fans? I run a 120mm fan over the PWM area of the motherboard as well as the memory. This is always how this system is run, regardless of the memory it is running.

Moving over to the Intel platform, again, these modules have no problems being installed in the DIMM slots and getting ready for action.

I will be using the Hipro-Tech.com DDR Maximizer as I always do for this review and because the modules are no taller than standard DDRII memory, the Maximizer PCB fits in-between the Team Xtreem Dark without issue.

With everything plugged in and powered up, this is how the test bed looked throughout benchmarking and overclocking. I won't go over the entire list of hardware being used here but you can find that information in the Intel Overclocking section.

Page 5 : AMD Overclocking

The Team Xtreem Dark series is designed for entry level overclocking so naturally, I am going to see how far I can push these modules at reasonable voltages. Here is the full list of hardware used for the AM2 overclocking and benchmarking:

I will be using my typical blueprint for building stability at each of the overclocks listed below. This mixture of programs has a good track record with me for testing stability of memory. Keep in mind that further testing of a memory overclock is always a good idea for 24/7 use but this list of requirements is a great start for a 100% stable memory overclock.

• Dual 32M runs of Super Pi Mod 1.5 (ran at the same time)
• 2 hours of dual Prime95 using Orthos Beta on blend mode
• 1 hour of dual MemTest in Windows using 750MB/instance
• Multiple loops of each 3DMark 01 / 03 / 05
  

  As I said above, the stability testing above has passed at every overclock I have listed in the charts below. We start with a look at the overclocking of the Team Xtreem Dark at the tight timings of 3-3-3-8.

  
  3-3-3-8
  
  

  

  When I started overclocking these modules, I wasn't sure what to expect based on the fact that they were not the typical high-end overclocking memory I was use to seeing. At 3-3-3-8, I was very quickly surprised at how well these modules progressed. As you can see in the chart above, these Team Xtreem Dark have done incredibly well almost reaching DDR800 with only 2.29v. Normally this type of overclock is only found with much more expensive Micron D9 based modules and the values in this kit is growing quickly.

  DDR780 3-3-3-8 @ 2.29v

  

  
  4-4-4-8
  
  

  

  The good times keep rolling with this memory at 4-4-4-8. Much like the 3-3-3-8 overclocking, the results here are quite impressive for a kit of memory that isn't suppose to be for "high-end" overclocking. As you can see, there is plenty of headroom with these modules and the voltage required to reach these high frequencies is pretty impressive. I'm not so sure these modules are just for beginner overclockers, I'm far from a beginner and am quite pleased with these results.

  DDR860 4-4-4-8 @ 2.00v

  

  
  5-5-5-8
  
  The run of great overclocks ended at 4-4-4-8 I guess. It's quite obvious that there are no results in the 5-5-5-8 section and that is because I couldn't seem to get this memory stable at 5-5-5-8 at any higher clocks than 4-4-4-8. You can basically mirror the 5-5-5-8 results to the 4-4-4-8 results at the same voltage. Regardless of this lack of 5-5-5-8 headroom on the AM2 platform, getting to DDR950 at 4-4-4-8 with only 2.29v is plenty impressive.

  
  Page 6 : Intel Overclocking
  

  After looking at the overclocking results on the AM2 platform, I am very interested to see what will happen on my P5B-Dlx platform. I am most curious to see if I can get 5-5-5 timings to clock higher than 4-4-4. As always, we start with a quick look at the hardware used for the Intel overclocking and benchmarking:

  CPU: Intel C2D E6300 (L630A864)
  CPU Cooling: Chilly1 Single Stage Phase Change
  MB: Asus P5B-Dlx Wifi-AP Edition
  GPU: Biostar 7600GS 256MB DDRIII
  PSU: Silverstone Zeus 560W
  HD: Seagate SATAII 80GB 8MB NCQ
  OS: Windows XP SP2 (with all updates)
  Hipro-Tech.com DDR Maximizer v1.2

  Memory:

  TEAM Xtreem Dark 2x1GB PC2-6400 4-4-4-12

  Like you probably expected, stability testing remains the same on the Intel platform with the exact same programs being used for testing each overclock. The list of programs and their specifications will again be used for each and every overclock listed in the charts below. Here is that list of programs one more time:

• Dual 32M runs of Super Pi Mod 1.5 (ran at the same time)
• 2 hours of Prime 95 using Orthos Beta on blend mode
• 1 hour of dual MemTest in Windows using 750MB/instance
• Multiple loops of each 3DMark 01 / 03 / 05
  

  You will notice the first set of timings I am overclocking at is 3-4-3 instead of 3-3-3. This is because the P5B-Dlx performs better at 3-4-3 when the FSB is 400 or less so there really isn't a point of overclocking 3-3-3. For what it is worth, 3-3-3 usually clocks about 5-10MHz less than 3-4-3 at the same voltage.

  
  3-4-3-8
  
  

  

  The 3-4-3 results on the P5B-Dlx are very similar to the 3-3-3 overclocking numbers on the AM2 platform. DDR800 was reached though at 2.30v and that is just flat out redicules for a "budget" 2GB kit of memory. I have had plenty of more expensive Micron D9 based modules come through here and these Promos based Team Xtreem Dark are not very far behind them at 3-4-3.

  DDR800 3-4-3-8 @ 2.30v

  

  
  4-4-4-8
  
  

  

  It appears that the P5B-Dlx likes this memory a little bit better than my AM2 motherboard. The 4-4-4 clocks are even better than they were on AM2. Reaching a lofty DDR1000 with less than 2.30v is not only close to but actually as good as a few Micron D9 based modules I have recently reviewed. I really did not expect this kind of an overclock from the Team Xtreem Dark but it certainly is a nice surprise and you won't find me complaining at any time.

  DDR1000 4-4-4-8 @ 2.25v

  

  
  5-5-5-8
  
  

  

  Well, it's not the full slate of results that I normally have but certainly better than the AM2 platform. I basically stopped getting any further in clocking past 2.10v so it looks like this kit tops out at DDR1080 regardless of voltage. Unlike the AM2 platform, the P5B-Dlx seemed to run 5-5-5 pretty good but the limiting factor in stability wasn't Orthos or Super Pi, it was the 3D testing that would start to freeze the system with higher clocks. This isn't common and didn't show up at 3-4-3 or 4-4-4 so if you are doing stability testing at 5-5-5, don't forget to do some gaming for stability as that was the first sign of issues for each overclock.

  DDR1060 5-5-5-8 @ 2.00v

  

  
  Page 7 : AMD Benchmarks
  

  Now that we know what the memory is capable of in the overclocking department, it is time to see how those overclocks relate into performance. As I always do, I have decided on a few overclocks at different frequencies and timings to run benchmarks at.

  

  Because of the lack of 5-5-5 overclocking, I have stuck to 3-3-3 and 4-4-4 benchmarks. The timing sets and frequencies I have chosen should give some nice comparisons at different CPU speeds and should cover pretty much the entire range that this memory could run on the AM2 platform. We start off with a quick look at some memory bandwidth numbers.

  SiSoft Sandra 07 SP1 – Memory Bandwidth

  

  Lavalys Everest Ultimate 06 – Memory Bandwidth

  

  The bandwidth numbers seem to fall in line where they should with the highest numbers coming from the white bar in the graphs as the CPU speed is much higher for that result. With memory bandwidth, higher memory frequency will always provided higher testing results over a lower frequency with tighter timings, this is evident when comparing the grey and black bars. It will be interesting to see if the 3-3-3 result will outperform the 4-4-4 results in any other benchmarks.

  Lavalys Everest Ultimate 06 – Memory Latency

  

  I thought fore sure the black result was going to show a slight advantage in latency numbers with the tighter timings but I guess the gap in memory speed was too much to make up for the tighter timings.

  FutureMark PCMark 05 & 3DMark 01 SE

  
  

  Our PCMark 05 results sort of bunch up in the middle of the graph and show very little difference between the three results at roughly the same CPU frequency. Obviously PCMark 05 doesn't rely heavily on memory for its score. The same can be said for 3DMark 01 but there is a clear advantage for the black bar which lets us know that tight memory timings are marginally favored by 3DMark 01. I wonder if that translates into better 3D gaming performance for tighter timings as well?

  SuperPi Mod v/1.5 – 1M

  

  SuperPi Mod v/1.5 – 8M

  

  SuperPi Mod v/1.5 – 32M

  

  As small as the gaps are for the middle three results, there is still a pattern to be seen in SuperPi. Memory frequency has the smallest of advantages over tighter timings and the small memory difference between the blue and grey bars makes up for the difference in CPU frequency which is not what I was expecting.

  Gaming Benchmarks – Far Cry / Half-Life 2 / UT 2004

  

  The only thing I am really looking for in these results is if the 3-3-3 timings perform better than the 4-4-4 results did like in 3DMark 01 and sure enough, the smallest of margins is there. Despite a 60MHz CPU and 10MHz memory frequency handicap, the 3-3-3 results of the black bar still manage an average of 2FPS over the blue bar and even slightly more than that over the grey results. It just goes to show, speed isn't everything and the ability to clock this memory so high at 3-3-3 is definitely an advantage for the Team Xtreem Dark. Let's see how the Intel benchmarks shape up.

  
  Page 8 : Intel Benchmarks
  

  There is no reason to stop the benchmarks at AM2, so now we will get to some performance numbers on the P5B-Dlx which sports the Intel P965x chipset.

  

  Unlike the AM2 platform, I did manage to get 5-5-5 to clock a bit better than 4-4-4 so one set of results will be dedicated to the 5-5-5 timing. I will also be running benchmarks at the stock timings of 4-4-4-12 and some tighter timings of 3-4-3 and 4-4-4-8. The bottom three sets of results should make for an interesting comparison, let's take a look.

  SiSoft Sandra 07 SP1 – Memory Bandwidth

  

  Lavalys Everest Ultimate 06 – Memory Bandwidth

  

  The bandwidth numbers that we see here in both Everest and Sandra indicate that CPU speed plays a bigger role on memory bandwidth than anything else. Notice how the three bottom results (grey, black, and white) for Write bandwidth in Everest are all but identical? The Read results do show a slight variation but not a whole lot considering the large gap between memory frequencies of those three. The clear bandwidth winner is not so much DDR800 4-4-4 that the blue bar represents, it's the 2800MHz CPU frequency.

  Lavalys Everest Ultimate 06 – Memory Latency

  

  Where bandwidth doesn't show an advantage to memory frequency on this Intel platform, memory latency results certainly do. The lowest latency result comes from the DDR1080 scores despite running 5-5-5. The memory controller on the Intel platform is in the northbridge and not directly on the CPU die like on the AM2 platform. This results in varying memory performance and characteristics between the two platforms.

  FutureMark PCMark 05 & 3DMark 01 SE

  
  

  Again, I will focus on the bottom three results and point out that the difference in performance between them in PCMark 05 and 3DMark 01 is very minimal but leans towards higher memory frequencies. In 3DMark 01, the black bar seems to hold a slight advantage which plots the DDR972 4-4-4 performance. On the AM2 system, the 3DMark 01 results predicted the gaming performance, we will soon see if that holds true on the P5B-Dlx as well.

  Super Pi Mod v/1.5 – 1M

  

  Super Pi Mod v/1.5 – 8M

  

  Super Pi Mod v/1.5 – 32M

  

  There really isn't anything to discuss or point out with the SuperPi results. Memory frequency will always beat out timings and that is what has happened here. Of course, memory frequency comes in second place to CPU frequency as proved by the blue bar in the graphs, but that is pretty much common knowledge.

  Gaming Benchmarks – Far Cry / Half-Life 2 / UT 2004

  

  The results aren't conclusive with HL2 giving DDR1080 the win over the other two results at a CPU speed of 2520MHz but it appears that 4-4-4 around the DDR1000 marks seems to be the sweet spot in this particular environment. Obviously CPU speed is king, like it is in all results seen here today, but the memory can also play a role in performance and not always is faster better.

  
  Page 9 : Conclusion
  

  I wasn't sure how this review was going to go when I first heard that a set of the new Team Xtreem Dark were going to be coming my way. They are new and not the traditional Micron D9 based modules that make up all the overclocking DDRII I have tested so far. Needless to say, the overclocking sections were going to be a surprise no matter what happened there.

  

  As it turned out, the Team Xtreem Dark did really well for a budget 2GB kit. The overclocking put up some very impressive numbers with the memory surpassing the lofty standards of DDR800 CL3 and DDR1000 CL4 with a very conservative 2.30v on the Intel platform. Overclocking on the AM2 setup didn't go quite so smoothly but with a little effort I still managed impressive enough results at CL3 and Cl4.

  I was a bit caught off guard with the Team Xtreem Dark but with an amazing looking heatspreader, lifetime warranty, impressive overclocking, and a price I can live with, how could I complain? As I have said, some of the D9 based 2GB kits I have tested were not able to match the impressive overclocking of the Team Xtreem Dark 2GB kit that I reviewed here today and for exceeding all of my expectations, I am proud to give the PC2-6400 Team Xtreem Dark our Overclockers Online Budget Award:

  

  Advantages

  Lifetime Warranty
  Snappy looking heatspreader
  Plenty of overclocking headroom at low voltages
  A price tag lower than most other PC2-6400 2GB kits

  Disadvantages

  AM2 overclocking wasn't jaw-dropping like the Intel platform
  Stopped responding to voltage relatively quickly

  Overclockers Online would like to thank Robin from Team Group for this review opportunity.