老当益壮！新疆军区用八一杠步枪打靶

TEAMGROUP is a memory, AIO and SSD manufacturer based out of Taiwan and founded in 1997. They sell memory under the T-FORCE brand, and when their contact person reached out to me wondering if I was interested in taking a look at their memory sticks, I jumped at the chance; this follows a review for I did for them back in May testing their DELTA 7200MT/s DDR5 memory kit.

First a disclosure, TEAMGROUP sent me this XTREEM DDR5 32GB (2x16GB) 6000MT/s CL30 Black kit sample to keep, without any review pre-approval.

Sayan Sen contributed to this feature, and also provided the benchmark graphics.

Specifications

First off, here are the full specifications of this memory kit.

Introduction

The T-FORCE memory was benchmarked in the following system:

• Cooler Master MasterBox NR200P MAX

• ASRock Z790 PG-ITX/TB4

• Intel Core i7-14700K with Thermal Grizzly Carbonaut Pad
• Corsair Vengeance 2x16GB 6000MT/s CL30 (XMP Profile)
• TEAMGROUP T-FORCE DELTA 2x16GB 7600MT/s CL36
• ASRock AMD Radeon RX 7900 XTX Phantom Gaming
• Kingston Fury Renegade SSD

The ASRock Z790 PG-ITX/TB4 motherboard had BIOS version 15.01 at the time of testing, and I reset BIOS to default settings and only enabled the Intel XMP 3.0 profile with all of the tested memory.

Windows 11 was up to date with July Patch Tuesday build 26100.4652 (KB5062553) ?at the time of testing and I ensured I had minimal programs running in the background with the exception of AMD Adrenaline, Razer Central, and Microsoft Defender active in the system tray.

Benchmarks

For our benchmarks, UL Solutions provided us with Professional (commercial use) licenses for 3DMark, and Procyon; and a copy of AIDA64 Engineer was provided to us by Aida64.com. In addition, HWiNFO provided us with a commercial license that let us confirm that this kit is Hynix SK, M-die, Single Rank memory.

We start with AIDA64 Cache & Memory Benchmark for all four kits to get a measure of the full capabilities of each of them:

T-FORCE XTREEM DDR5-6000 CL30:

Corsair VENGENCE DDR5-6000 CL30:

T-FORCE XTREEM DDR5-7600 CL36:

T-FORCE XTREEM DDR5-7200 CL34:

Following the purely synthetic tests, we next move to workloads that are more representative of typical tasks like gaming, AI, productivity, and other everyday usages.

For reference, in the charts below, the memory kits are sorted by different color shades of blue, the higher the frequency the deeper the shade.

Starting with 3DMark, we tested the physics test only and not the entire suite since the CPU is what is necessary for processing the game physics, logic, as well as GPU draw calls for the graphics card. A CPU-heavy test like the 3DMark Physics test helps to gauge the CPU's true gaming capability.

The XTREEM DDR5-6000 CL30 kit does exceptionally well on 3DMark DirectX 11 especially in Sky Diver—which also makes sense, given that it is lighter graphically and thus the CPU has more to do here than on Fire Strike, relatively speaking.

On Sky Diver, we see a 19% better performance than the Corsair VENGEANCE memory that is specced similar. On Fire Strike too the XTREEM 6000 CL30 was the fastest.

In DirectX 12, things do not change much as TeamGroup's 6000 CL30 kit keeps up with the faster memories. However the performance differences are not as prominent as were in the case of DX 11 since the older API was more single-thread bound and thus ran into CPU bottlenecks far more easily.

The XTREEM 6000 C30 RAM uses SK Hynix M-Die and that could explain this really great showing. The AIDA64 memory and cache synthetic benchmark above did not hint at something like this, that we would see such a big gap in real-world performance, and this has definitely surprised us, in a good way.

7-Zip decompression was the only test the new T-FORCE memory lost in, though as you can see above, the scores are all quite close to one another. In decompression however the XTREEM 6000 CL30 kit trumps the CORSAIR VENGEANCE by a good margin..

Next we checked out AIDA64's AES, Zlib and PhotoWorxx benchmarks as we wanted to see check performance in things like encryption (AES), data compression (Zlib), and image processing (PhotoWorxx).

AIDA64 showed almost no difference in the AES test as all kits were in the margin of error territory. This surprised us as we thought the RAM differences would reflect better in an encryption benchmark. That is evidently not the case as it seems the processor is more important.

It's a similar story for Zlib too, which was another surprising result given that Z-Zip earlier showed a significant difference. Thus, it is clearly not a case of one-size-fits-all as the compression algorithm on Zlib is much less sensitive to memory than the one in the 7-zip benchmark.

PhotoWorxx was where we saw the differences between the kits. Both the 7200 and the 7600 modules were much faster than the 6000 ones.

We tested browsing performance using Speedometer 3.0. Speedometer provides a value and also a range showing the highest and lowest scores as indicated in the chart above by the two set of scores for each browser.

Microsoft Edge showed the most response to faster memory speeds but Chrome also liked the new XTREEM 6000 RAM. Mozilla's Firefox too, which was generally quite uncaring of memory speed and latencies in case of the other kits, seemed to love these XTREEM Black 6000 CL30 modules.

Next up, we did some productivity testing with UL's Procyon suite of benchmarks.

First up, we have the Office test and the TeamGroup 6000 C30 kit does an amazing job here as it even outpaces the faster memories.

We also ran Computer Vision, which is an AI inference benchmark and saw identical figures with each of the RAM kit. Each of them put up around 192 points. We used the WinML API and float32 precision as it is more memory-heavy than float16.

Finally we have Geekbench AI and once more, the XTREEM 6000 CL30 put on a great show especially in the case of the Quantized metric. However, it must be noted that we updated Geekbench AI from 1.2.0 to 1.4.0 and although Geekbench warns that scores cannot be compared between major versions, what the margin of error is between point versions is unknown.

Pricing

Conclusion

Coming into the review, I did not expect to see such a big difference in the kits' performances; since both the TeamGroup and the Corsair are rated at 6000 CL30, I thought they would be trading blows in most cases. However as we see, that is clearly not the case as the XTREEM Black memory punches well above its weight.

Currently the kit of Corsair Vengence on Amazon costs $30 more since the XTREEM memory is ~$106; in fact there's a note that the Corsair kit is frequently returned, which is not good.

I find it hard to find any faults at all with this memory kit and it is a 10 out of 10 for sure considering the value and the performance. The only thing that may be works against it is the lack of RGB lighting but that is simply not enough to deduct a point in our book given its outstanding value for money.

TEAMGROUP gets a thumbs up from me for their T-FORCE memory, they installed without any issues and from the multiple times I powered on the system, the ASRock motherboard did not have to recalculate the timings.

As an Amazon Associate we earn from qualifying purchases.