Table of Contents
Today we are evaluating Team Group’s T-Force XTREEM 4133 MHz DDR4 memory modules. The XTREEM line of memory from Team Group doesn’t have any flashy RGB or wild color schemes. What it does have is rock-solid performance and a legacy of legendary overclocking results. Exceptional build quality, high XMP speeds, and performance are its major hallmarks. Join us as we take a look at these features and investigate the XTREEM memory’s overclocking potential.
Specifications and Features
Team Group’s T-Force XTREEM lineup is a performance-oriented series of DDR4 memory modules. Instead of RGB, they’ve chosen to outfit this series of memory with an award-winning, high-efficiency aluminum heat spreader on this and a lot of other memory configurations. With kits composed of 4GBx2, 8GBx2, 8GBx4, 16GBx4, the XTREEM series has something to fit every need.
The 8GBx2 sets are by far the most popular choice for overclockers and gamers, so we will direct our focus there. Starting out with the highest binned SKU is the XTREEM 4500 @ $196.99. The 4500 MHz kit runs an astonishingly tight set of timings. At a low 18-20-20-44 they are among the tightest, and thus highest performing, of all 4500 MHz kits on the market today. At the lower end of the spectrum, they offer the XTREEM 3466 @ $109.99. Interestingly enough, both the 4500 MHz and 3466 MHz editions of this memory run the same exact primary timing profile of 18-20-20-44.
Team Group offers a plethora of different options, but the sweet spot seems to be the XTREME 4133 @ $137.99. What makes the 4133 MHz kit unique is that it’s the only one running over 3733 MHz that comes with a tight timing profile of 18-18-18-38, while all of the others are rated for 18-20-20-44. In the table below we examine the particular details of the memory being evaluated today.
|Team Group T-Force XTREEM 4133|
|Capacity||16GB (2 x 8GB)|
|Speed Spec||PC4 33000|
|Rated Frequency||DDR4 2066MHz (4133MHz Effective)|
When purchasing DDR4 memory the main factors to consider, other than memory size or physical features, are the operating frequency and timings. XMP is a memory profile stored inside the actual memory, which allows the user to easily apply the rated frequency and timings. This kit of memory has a very high-performance XMP rating of 4133 CL18-18-18. Note the rating is high enough that many users might have compatibility issues with certain motherboards. The mid-range and budget motherboards often have issues above 3600 MHz, so that should be a factor to consider when purchasing memory.
|Team Group T-Force XTREEM|
Moment of Glory
On January 13th, 2017, Thai Overclocker, “Audigy”, submitted a historic DDR4 frequency world record of 5260.8 Mhz. Audigy then went on to break his own record again and submitted a higher frequency of 5280 Mhz.
High-Quality IC Chips
TEAMGROUP insists on using high-quality original IC chips which were selected and passed through various rigorous internal tests. Team Group equips their memory to meet and exceed the demands of high-performance computers
Unique Trench Design
The TEAMGROUP design team built a heat spreader with an all-new exterior design concept. The heat sink is built by top-notch aluminum extrusion process and CNC machining. The unique trench design can increase the radiating area to improve radiating efficiency by up to 10%. The clean, sleek lines are able to magnify its high-speed specification.
For those interested, here is a closer look at the specifications. Below is a screenshot of Thaiphoon Burner, which is a wonderful free tool that allows one to read the Serial Presence Detect (SPD) firmware of the DRAM. The SPD information is critical in determining how the stick will perform and how the computer will recognize it.
As the Thaiphoon Burner screenshot shows, this specific set of memory is composed of Samsung B-DIE ICs. This memory uses a total of 8 ICs, all of which are located on only one side. As you might know, Samsung B-die has become synonymous with high frequency and tight timings.
Please note that while Thaiphoon Burner shows the PCB revision as A2 (8 layers), we have visually confirmed that the PCB design is in fact A0. This A0 or A2 designation refers to how the ICs are arranged on the PCB itself in accordance with modern DDR4 PCB standards. Team Group confirmed for us that this memory is utilizing a newly designed 10-layer PCB for maximum performance and reliability.
Team Group aims to change the marketplace standard for high-end memory with aggressively competitive pricing. Consumers who are on a tight budget aren’t necessarily looking for a high-end product with attractive packaging, but that’s exactly what they get. Team group does not disappoint. The memory comes in a basic full-color box with a glossy image of the memory that hides beneath. While it’s not the most premium packaging we have seen for DDR4, it certainly exceeds expectations for the price point. The front of the package proudly displays that this memory was awarded the iF Design Award 2018, Computex D&I Award 2017, and the Good Design Award 2017.
Opening the box, owners are greeted with the memory safely nestled in the prototypical dual plastic clam-shell packaging. Team group has included stickers and a brief owners manual inside the package.
Once out of it’s packaging, the overall feel in the hand is quite memorable. The final fit and finish of this memory come together naturally and nothing feels out of place. Hefting this memory in your hands give you the feeling of quality.
While there is no RGB on this memory, we want to point out the other physical characteristics, such as the integrated heat spreader. Team Group has equipped each memory module with a thick and beefy-feeling, aluminum heat spreader. Even to a casual observer, it’s clear that Team Group went to great lengths to set the standard for RGB-less memory. The T-Force XTREEM heat sink is by far one of the nicest we have had the pleasure of reviewing.
The heat spreader, or heatsink, doesn’t just give the memory style points, it also helps to keep the memory running cool even under the most demanding circumstances. The main body is extruded aluminum, but to dress that up Team Group has applied an anodized and brushed aluminum piece to the lower half.
The thickness of the aluminum that composes the heat sink is about 2.3mm, and the total weight of one memory module is 93 grams. The overall maximum dimensions are 134mm long, 48mm tall, and 9mm thick.
The memory pairs nicely with an all-black motherboard build such as our ASRock Z390 ITX motherboard.
Testing and Overclocking
The overall objective is to evaluate the memory under a variety of different conditions. To accomplish this task, we will turn to benchmark programs to examine the performance of the memory under various conditions. Our approach is to start out by fully testing the XMP profile. Once we have established that the XMP profiles are working on the test system, then the real fun begins as we evaluate the memory from an overclocking perspective.
We will examine the overclocking potential without excessive voltage. According to the XMP 2.0 certifications, the absolute maximum allowable voltage is 1.50 V VDDR. Thus, our overclocking endeavors will be conducted with less than 1.50 V. However, later in this review, we will examine what happens if we push this memory to the extreme limits with more than 2.0 V.
Below are the test system and resulting memory speeds that will be used to evaluate the memory and run the benchmarks.
|CPU||Intel i9 9900k@ 5.0 GHz (4.7 GHz Cache)|
|Cooler||NZXT Kraken X62 280mm AIO|
|Motherboard||ASRock Z390 Phantom Gaming-ITX/AC|
|Graphics Card||MSI R9 290X Lightning|
|Solid State Drive||Team Group L5 LITE 3D SSD|
|Power Supply||Enermax RevoBron 700W|
|Operating System||Windows 10 x64|
|Memory Speeds Compared|
|XMP ~ 4133 MHz CL18-18-18-38 – 1.40 V|
|Test Case 1 ~ 4500 MHz CL18-18-18-38 – 1.40 V|
|Test Case 2 ~ 4133 MHz CL16-15-15-38 – 1.40 V|
|Test Case 3 ~ 4500 MHz CL16-15-15-38 – 1.50 V|
As is the case with all overclocking adventures, your results may vary, so proceed only if you assume all risk. To view and examine all of the various memory profiles we use two primary tools which include AIDA64 and ASRock Timing Configurator. AIDA64 is a powerful system diagnostic and benchmarking tool, which can be purchased for a reasonable price. Next, we will be using the ASRock Timing Configurator, which is a free piece of software that allows users of all major motherboard brands to see the primary, secondary, and tertiary timings of Intel-based systems.
Below is the XMP profile. This particular kit of memory comes with an XMP profile of 4133 MHz CL18-18-18-38 @ 1.40 V. Most of the kits we review, and indeed most of the kits available on the market today, come with an XMP voltage of 1.35 V. The manufacturer decided to set 1.40 V in order to make the rating stable. While it’s not common with modern DDR4, some modules come with even higher XMP rated voltage of up to 1.50 V.
The improvements in XMP profile speeds is greatly attributed to modern manufacturing processes and memory PCB layouts; however, it might not be attainable on all motherboards. The profile is intended to be a one-click overclock, but it likely only applies to enthusiast-grade motherboards. If you are building a computer based on a budget motherboard, don’t be surprised if you cannot achieve stability with this XMP.
Memory overclocking is all about pushing the frequency higher and lowering the timings. One hurdle for memory overclocking is the motherboard. The distance between the CPU and the memory modules has a direct relationship to the overclocking potential of the memory. Motherboards with a larger distance between the CPU and the memory will have greatly reduced overclocking potential compared to motherboards with a shorter distance. Therefore, motherboards with only two dual, in-line memory modules (DIMMs) will have a better likelihood of increased memory overclocking potential than motherboards with four DIMMs, because the distance is inherently shorter.
Once the XMP profile has been successfully tested, we can dive into overclocking. Our methodology is to set my maximum working voltage of 1.50 V and see what could be accomplished, then lower the voltage to find the stability point. The motherboard being used is the ASRock z390 Phantom Gaming-ITX, which is an entry-level priced motherboard that has elite-level memory overclocking capabilities.
For the first test case, only the operating frequency was increased with no other settings adjusted. With a voltage increase from 1.40 V to 1.45 V, the frequency was able to be increased by an astonishing 366 MHz for 4500 MHz effective DDR speed. Below is the resulting first overclock test case.
For the second overclock test case, the same small voltage increase of 1.40 V to 1.45 V was used. Contrary to the first result, this time only the memory timings were decreased to provide an alternative example of what types of overclocks are possible. The operating frequency was held at the XMP 4133 MHz, but the primary timings were decreased as far as possible while still maintaining relative stability. Below is the resulting second overclock test case of 4133 MHz with CL15-16-16.
For the third and final overclocking test case, the goal was to achieve the best overclock possible while still staying within the maximum allowable voltage of 1.50 V according to XMP 2.0 specifications. Finding the optimal overclock for a given criteria, such as voltage, is not an exact science and requires a high degree of knowledge and patience, as well as the right combination of equipment. However, the results can be quite impressive. The resulting overclock was a 366MHz increase in frequency beyond XMP, with tighter timings (decreased) as well. Given the test equipment and the voltage constraint, this is a very remarkable overclock.
First up, we used AIDA64 Cache and Memory Benchmark. The graph below shows that overclocking the memory beyond XMP had a positive impact for all test cases. Simply lowering the primary timings showed a negligible impact on relative performance. To have a noteworthy gain in performance, the frequency must be increased, as in test cases 1 and 3.
Next up is Geekbench 4, and it has proven itself to be an excellent tool for determining the real-world performance of the system being tested. This type of benchmark is purely 2D calculation-based and there is no graphical processing element, so it’s a great analytical tool to evaluate memory performance.
Geekbench 4 tells a similar story to AIDA64 in terms of memory frequency scaling. Across the board, the integer and floating-point tests showed an insignificant performance gain from all overclocking test cases. The memory and crypto scores showed that primarily frequency had the biggest effect on the score.
Next, we will examine the performance using a few of the memory benchmark tests offered within the SiSoftware Sandra suite of benchmarks. The flagship product, known as Sandra, is a powerful suite of many different benchmarks to evaluate computer performance of all major components, including the processor, graphics, memory, and disk.
Some benchmarks tests don’t show much performance gain from memory overclocking, and the SI Sandra Transactional Throughput is one of them. Here we can see that overclocking the memory showed a noticeable gain in the memory bandwidth. As is often the case, overclocking the memory frequency often has the biggest gain in performance compared to tighter timings.
Memory overclocking and relative performance are directly related to the tasks being performed. For mathematical tasks such as encoding, Geekbench 4 shows us that memory overclocking plays a fairly important role. However, when we look at gaming performance, as is the case with UL Benchmarks, we see that memory overclocking has no noteworthy effect.
In this next section, we will examine the XMP profile performance of recently review memory. All of the memory compared here has been tested on the same exact z390 test bench and with the same exact operating system. Due to this fact, we are able to make some direct performance comparisons of recently reviewed memory. For the purposes of continuity, we will examine only the XMP profile and exclude any overclocked comparisons.
Below, in the AIDA64 test results, we observed that the T-Force XTREEM was faster in all tests when compared to all other compared memory modules.
Next, we will look at the XMP performance using the Geekbench 4 benchmark program. In the results below, we can see that for all of the compared memory, integer and floating-point tests showed an insignificant performance gain from the different XMP ratings. The memory and crypto scores give us a better picture of the speed differences. However, taking the average into account we conclude that Geekbench 4 is only marginally affected by big differences in XMP rating.
As most benchmarks and real-world tests will show, DDR4 frequency is often the most important factor when it comes to overall system performance. Other than for competitive benchmarking purposes, making the memory run tighter often has very little effect on the overall system performance.
Certain benchmark programs, such as Geekbench 3, have a direct relationship between memory overclocking and the overall score. For competitive overclockers, memory overclocking is critical for the overall score. Here we take a look at what can be accomplished by taking overclocking to an extreme level. With potentially destructive voltages of 2.10+ V, the memory comes alive and allows for truly astonishing timings and frequency. The gold standard for Samsung B-die within the extreme overclock community has been 4000MHz with CL12. However, the new B-die IC standards are changing.
Here we have a frequency of 4133 MHz that could be accomplished in Geekbench3 using CL12-12-12 timings. The secondary and tertiary timings are also exceptionally tighter when compared to XMP.
The memory was able to pass Geekbench 3 at slightly higher frequencies of about 4180 MHz; however, the core and cache speeds were altered in the process making a comparison to XMP speeds impossible. The next memory frequency required to keep the core and cache consistent was 4205 MHz, which was not possible with this memory. There was no form of exotic cooling used for this result.
In the graph below we can see that extreme memory overclocking has a huge benefit on the memory score. Looking closer at the other test we see that the floating-point test showed relatively no increase from extreme memory overclocking. What matters to competitive overclockers is the overall score. In this case, we can see that the extreme overclocking gave us a scant 5.5% gain over XMP.
In the competitive arena, where every last bit of speed is critical, an overall score gain of close to 5% is excellent. However, to an average user or gamer, the risk factor might not outweigh the considerable danger involved.
Team Group offers up a performance-minded series of memory that is serious about system compatibility and overclocking. For those who don’t want RGB, we cannot think of a better kit of memory. The T-Force XTREEM comes with an impressive heat sink, which was built to handle the most demanding heat loads possible for DDR4. From the sleek black color to the clean and modern design, we feel that this module would be a nice fit in just about any build. Our only concern is that the heat sink is very tall and may interfere with some aftermarket air coolers.
When it comes to overclocking, the T-Force XTREEM exceeded our expectations in every way. Were able to increase the memory frequency from 4133 MHz to 4500 MHz with no additional voltage. Once we finally set the maximum XMP voltage of 1.50 V, we were able to considerably lower the primary timings for a resulting stable daily overclock of 4500 MHz CL16-15-15. The overclocking meets or exceeds the best B-Die we have reviewed up to this point.
You might be thinking that exceptional design and best-in-class overclocking would come with a premium price tag, but you’d be wrong. At just $159.99 for the modules we tested, the T-Force XTREEM 4133 falls in the midrange price point for similar XMP ratings. If you are looking for top B-die overclocking potential at incredibly competitive pricing, then look no further, because this is as good as it gets.
David Miller – mllrkllr88