Review: TeamGroup T-FORCE XTREEM DDR5 8000c38

[LiquifyMods]

Having had the opportunity to play with Teamgroup's T-FORCE XTREEM DDR5 MEMORY 32GB (2x16GB) 8000c38 for the past 2 weeks, I delved deep into the realms of aesthetics, performance, and compatibility. Here's an honest overview, blending my personal experiences with the brochure insights:

Aesthetic Appeal and Performance Prowess:
My personal opinion is that these are perhaps the best looking Ram dimms on the market. The sleek, black design of these modules not only caught my eye but also promised a visual treat within my custom PC build. Additionally, their out-of-the-box performance was indeed commendable, delivering snappy responsiveness and seamless multitasking, exactly as advertised.

The Compatibility Conundrum:
However, the path to achieving optimal performance was riddled with challenges. My Intel i9 13900KS CPU paired with the Asus ROG z790 Apex motherboard initially struggled to handle the modules at their advertised speed of 8000c38. Despite expectations, stability issues surfaced, compelling me to throttle the speed down to 7800c38. While chips with better IMCs might fare better, my personal experience underscores the importance of CPU and motherboard compatibility and "silicon lottery", indicating that achieving the advertised speeds might (and will likely) require manual tuning.

Adaptive Tuning and Stability:
Despite the initial hurdles, I discovered that with manual tuning, the modules could indeed achieve the advertised 8000 speed. This experience highlights the adaptability of the T-FORCE XTREEM DDR5, showcasing its potential to unleash its full capabilities with the right adjustments. After taking some time to tweak the timing, sub-timings and voltages, I was able to stabilize the ram at 8000c36 for daily use. For benchmark use, it can go even higher and with tighter timings, but that will be a separate post in itself.

Excellent Engineering:
Behind the scenes, Teamgroup's engineering prowess shines through. The meticulously layered aluminum sandblasted fins, sturdy 2mm heat sink, and high-quality ICs with patented techniques attest to the brand's commitment to innovation and performance. These features not only enhance aesthetics but also ensure exceptional heat dissipation and reliability, even under demanding conditions.

Lifetime Assurance:
Lastly, Teamgroup's provision of a lifetime warranty speaks volumes about their confidence in the durability and reliability of their products. It's a reassuring commitment, offering peace of mind to users, including myself.
In conclusion the T-FORCE XTREEM DDR5 ram was a rollercoaster of aesthetics, performance, and adaptability. While the road to optimal performance might have had a few challenges, the end result was undeniably satisfying.
These dimms aren't just components; they change the way I look at my build and bring the performance as well.

[TASOS]

Very nice looking modules.

These are Hynix A-die , right ?

Have you kept any records-charts with the settings you used during your testing-reviewing ?

How far did you go with voltages ? any temp issues or instabilities ?

 

[LiquifyMods]

Yeap, they should be exclusively A-die. no I didn't keep records, but had no issue going up to 1.6v. My major issues is with my IMC, as it's holding me back and not allowing me to push the mem further up. Hopefully I'll be able to get a good 14900k or ks soon and test to see where I can get these.

Although, if it's frequency you're after, the 2x24GB kit may be better. 

[Mr. Fox]

Thanks for the review. I agree that these are nice modules. The 2x24GB  kit is much better than the 32GB (2x16GB) A-die Xtreem 8000 kit. I purchased that initially and returned it for a refund. It was not nearly as good as the 48GB kit.

I have the 8200 kit and these are the second variation of Hynix M-die. I did not take pictures of the IC after removing the heatsinks to install the waterblocks, but it is clearly printed on them. Removing the heatsinks was very easy using acetone to melt the adhesives. If you soak the modules for about 10 minutes the adhesives will turn loose and allow you to separate the heatsinks from the modules without using force. The foam pad and the thermal pad will be left intact. The round silver medalion has pins that align with holes in the metal, so don't try to pry off the medalion by itself. Gentle force lifting the heatsink off of the module will peel the medalion away from the upper portion of the heatsink without breaking the medalion.

Edited March 19 by Mr. Fox

[LiquifyMods]

4 minutes ago, Mr. Fox said:

Thanks for the review. I agree that these are nice modules. Much better than the 32GB (2x16GB) A-die Xtreem 8000 kit. I purchased that initially and returned it for a refund. It was not nearly as good as the 48GB kit.

I have the 8200 kit and these are the second variation of Hynix M-die. I did not take pictures of the IC after removing the heatsinks to install the waterblocks, but it is clearly printed on them. Removing the heatsinks was very easy using acetone to melt the adhesives. If you soak the modules for about 10 minutes the adhesives will turn loose and allow you to separate the heatsinks from the modules without using force. The foam pad and the thermal pad will be left intact. The round silver medalion has pins that align with holes in the metal, so don't try to pry off the medalion by itself. Gentle force lifting the heatsink off of the module will peel the medalion away from the upper portion of the heatsink without breaking the medalion.

Excellent insight on this! Thanks a lot! I haven't tried taking off the heatsyncs at this point, but this is great info to have for later! Cheers

[Mr. Fox]

Sure. I appreciate you taking the time to provide the review.

I edited my post above because my initial choice of words was not clear. The 32GB kit you did the review on are indeed A-die. The 48GB kits are the second variation of M-die (came after A-die, which came after the first M-die). In general, the second variation of M-die used on the 24GB modules overclocks higher than the A-die and requires less voltage than the A-die.

Edited March 19 by Mr. Fox

[LiquifyMods]

1 hour ago, Mr. Fox said:

Sure. I appreciate you taking the time to provide the review.

I edited my post above because my initial choice of words was not clear. The 32GB kit you did the review on are indeed A-die. The 48GB kits are the second variation of M-die (came after A-die, which came after the first M-die). In general, the second variation of M-die used on the 24GB modules overclocks higher than the A-die and requires less voltage than the A-die.

yeah yeah I got that. I am also looking at getting a 2x24GB kit soon (probably the ARGB version) and will hope for better results. But I'll probably need to change CPUs as well for a better IMC

[TASOS]

On 3/19/2024 at 7:36 PM, Mr. Fox said:

... the second variation of M-die used on the 24GB modules overclocks higher than the A-die and requires less voltage than the A-die.

I think , this is the key point.

A friend from the States was telling me the same thing about the Patriot extreme 5 kit , where the 8200 48gb kit (M-die) is performing better than the 8200 32gb kit (A-die)

But what about latencies ?

[LiquifyMods]

15 hours ago, TASOS said:

I think , this is the key point.

A friend from the States was telling me the same thing about the Patriot extreme 5 kit , where the 8200 48gb kit (M-die) is performing better than the 8200 32gb kit (A-die)

But what about latencies ?

I believe they're a bit better on A-Die, but I do not have the tools for extensive testing at this point. Here is a quick test I ran with my daily settings yesterday (CPU is basically stock with 0.02v undervolt and ram is set to xmp tweaked from bios and simply played with voltages and primary timings a bit. 

[mcmanus9442]

Latency is pretty much similar, but A die can also run MUCH tighter Trfc as opposed to M die. 

Edited March 21 by mcmanus9442

[TASOS]

10 hours ago, mcmanus9442 said:

Latency is pretty much similar, but A die can also run MUCH tighter Trfc as opposed to M die. 

What about tFaw and other performance secondary and tertiary timings ?

I am also trying to figure out ,

Why do people who bench aircooled tend to prefer the later versions of M-die sticks ?
Lower voltages (and temps) for the same speed and primary timings compared to A-die ?

Does an average good IMC have the same behaviour with a 32GB A-die kit and a 48GB M-die kit ?

 

p.s.
Sorry for asking more general questions in a review thread.

[LiquifyMods]

5 hours ago, TASOS said:

Why do people who bench aircooled tend to prefer the later versions of M-die sticks ?
Lower voltages (and temps) for the same speed and primary timings compared to A-die ?

Does an average good IMC have the same behaviour with a 32GB A-die kit and a 48GB M-die kit ?

Don't have all the answers you're looking for, but I did tighten a few things up a bit and managed a new personal best in SuperPi 32M

[Mr. Fox]

On 3/20/2024 at 12:29 PM, TASOS said:

I think , this is the key point.

A friend from the States was telling me the same thing about the Patriot extreme 5 kit , where the 8200 48gb kit (M-die) is performing better than the 8200 32gb kit (A-die)

But what about latencies ?

Sorry for the delayed reply. I just now saw your post. Here is my TeamGroup Xtreem 48GB 8200 kit bumped to 8400. Good latency numbers. The high temps were before I removed the heatsinks and install the waterblocks.

Edited March 26 by Mr. Fox

[LiquifyMods]

15 hours ago, Mr. Fox said:

Sorry for the delayed reply. I just now saw your post. Here is my TeamGroup Xtreem 48GB 8200 kit bumped to 8400. Good latency numbers. The high temps were before I removed the heatsinks and install the waterblocks.

Very nice!