GIGABYTE Replaces DIMMs with CAMM2 on the AORUS Z890 Tachyon ICE Motherboard

[AleksandarK]

GIGABYTE has just rolled out its refreshed Z890 AORUS Tachyon ICE CAMM2 motherboard, designed to give overclockers and power users a serious edge. The biggest change is support for CAMM2 DDR5 memory modules, which offer cleaner signaling and much higher speeds compared to traditional DIMMs. At first glance, the new board looks a lot like the previous Tachyon ICE model. It still features the distinctive 90-degree rotated CPU socket and top-mounted memory slots. But instead of multiple DIMM slots, you now get a single CAMM2 connector. That slot comes with a toolless heatsink for better cooling—no fiddly screws required, just snap it on and you're ready to push your memory harder. If you're into extreme memory overclocking, the original Z890 AORUS Tachyon ICE recently set a world record at DDR5-12752 thanks to GIGABYTE's in-house expert, HiCookie, using V-Color's Manta Xfinity RGB modules. That achievement shows just how far this platform can be pushed. We recently reported on G.Skill offering a DDR5-10000, 64 GB of density (2-channels or 4 sub-channels), and timings of CL56-70-70 in CAMM2 form factor. That might be interesting to keep an eye on in the future, as it is definitely a new possible world-record contender.



View at TechPowerUp Main Site

 

[Chaitanya]

Real chicken and egg problem with adoption of CAMM form factor of memory.

 

[wNotyarD]

Now this is VERY interesting. I hope CAMM gets mass adopted very soon

 

[Tigerfox]

Placing the CAMM2-module above the socket is a bad idea, because larger modules (tpye B and D) won't fit.

 

[hsew]

Real chicken and egg problem with adoption of CAMM form factor of memory.

It’ll transition all at once. Especially since there is no compatibility between DDR versions. Click Drivers have extended usable life of the desktop DIMM form factor through DDR6 (probably) but there will almost certainly have to be another major innovation for DDR7 DIMMs to exist.

For laptops it’s even more urgent. I doubt there will be DDR6 SODIMMS. Power constraints are much greater.

 

[rv8000]

Meanwhile cu-dimm doing the same and higher speeds, without having to deploy a new standard thats going to cost the end user $$$.

Sick of hearing about camm, belongs in laptops, remains to be absolutely no useful proven benefit.

 

[Konomi]

Meanwhile cu-dimm doing the same and higher speeds, without having to deploy a new standard thats going to cost the end user $$$.

Sick of hearing about camm, belongs in laptops, remains to be absolutely no useful proven benefit.

If you don't like it, don't buy it. Not hard to do. Same applies to even looking at articles that mentions it in the title - you don't need to do that. For most people, high speed memory in general is hardly necessary as you're going to be hindered by the software or OS. Especially so given how developers are releasing poorly optimised and buggy code for fast short-term profits.

 

[TheDeeGee]

Love it, no more issues with large tower coolers =]

 

[rv8000]

If you don't like it, don't buy it. Not hard to do. Same applies to even looking at articles that mentions it in the title - you don't need to do that. For most people, high speed memory in general is hardly necessary as you're going to be hindered by the software or OS. Especially so given how developers are releasing poorly optimised and buggy code for fast short-term profits.

You’re missing the point. Cu-dimm does everything camm does when it comes to speed/performance benefits, WITHOUT requiring new motherboard standards.

Why retool the design of motherboards for no benefit? This is just going to pass increased costs to end users. It doesn’t save any real-estate when it comes to the motherboard layout.

Now if it were to provide 1.5x the frequency or 30-40% better latency there might be a point, but that’s no where to be found. All it’s proven to be so far is wasted development and resources for mb vendors, on boards that will never see the light of day in retail.

Love it, no more issues with large tower coolers =]

My NH-D15 and Trident Z/Corsair Titanium will gladly disagree with that statement.

 

[TheDeeGee]

You’re missing the point. Cu-dimm does everything camm does when it comes to speed/performance benefits, WITHOUT requiring new motherboard standards.

Why retool the design of motherboards for no benefit? This is just going to pass increased costs to end users. It doesn’t save any real-estate when it comes to the motherboard layout.

Now if it were to provide 1.5x the frequency or 30-40% better latency there might be a point, but that’s no where to be found. All it’s proven to be so far is wasted development and resources for mb vendors, on boards that will never see the light of day in retail.



My NH-D15 and Trident Z/Corsair Titanium will gladly disagree with that statement.

If you have low profile memory you can fit the front fan.

I had to go with a 120 on my D15 and case combination. But with CAMM i could fit a 140 no problem.

 

[wNotyarD]

If you have low profile memory you can fit the front fan.

I had to go with a 120 on my D15 and case combination. But with CAMM i could fit a 140 no problem.

Yeah, when I used a DeepCool AK620 Digital, I had to install the second fan in the rear for it not to collide with the ram sticks.

One question about CAMM, though: isn't it projected to have better signal integrity in comparison to DIMM? Even if CUDIMM somewhat circumvents some clock limitations, isn't CAMM supposed to still allow lower timings?

 

[ymdhis]

Make sure the boards also support 12VO and BTF.

 

[Konomi]

You’re missing the point. Cu-dimm does everything camm does when it comes to speed/performance benefits, WITHOUT requiring new motherboard standards.

Why retool the design of motherboards for no benefit? This is just going to pass increased costs to end users. It doesn’t save any real-estate when it comes to the motherboard layout.

Now if it were to provide 1.5x the frequency or 30-40% better latency there might be a point, but that’s no where to be found. All it’s proven to be so far is wasted development and resources for mb vendors, on boards that will never see the light of day in retail.

Because that's the point of innovation. Sure - a competing standard may be able to do those things currently, but as time goes on, CUDIMM may hit a point where it is unable to improve as a result of the design choice. Again - it doesn't even matter for the majority of people, as they're not even going to benefit in most situations, therefore it is pointless complaining about it. Are you planning on buying the motherboard? No. Would you if it wasn't a CAMM2 board? Doubt it because it'd still be expensive most likely and you'd complain about the cost even then. I'm even sure you wouldn't even benefit from more than a gigabit network port on your motherboard as your internet speed most likely isn't even capable of saturating it. You're just complaining over nothing.

 

[rv8000]

Because that's the point of innovation. Sure - a competing standard may be able to do those things currently, but as time goes on, CUDIMM may hit a point where it is unable to improve as a result of the design choice. Again - it doesn't even matter for the majority of people, as they're not even going to benefit in most situations, therefore it is pointless complaining about it. Are you planning on buying the motherboard? No. Would you if it wasn't a CAMM2 board? Doubt it because it'd still be expensive most likely and you'd complain about the cost even then. I'm even sure you wouldn't even benefit from more than a gigabit network port on your motherboard as your internet speed most likely isn't even capable of saturating it. You're just complaining over nothing.

I would absolutely buy a Tachyon. In fact I’ve stated several times on the forums this past year in news articles related to it, but the b650 tachyon never saw a US release (or anywhere for that matter), and gigabyte have otherwise been twiddling their thumbs.

Until there are tangible benefits, camm is a dud for numerous reasons. It’s at the point where we have no consumer boards after 3 years and in terms of performance it’s not even better than cudimm. Innovation for the sake of innovation isn’t always a good thing. Fragmenting standards make the entire ecosystem worse for consumers.

There has yet to be any real, documented evidence of camm being better, just empty promises.

 

[OneMoar]

yay more bendable lga pins and No ram expansion sign me up

 

[wNotyarD]

yay more bendable lga pins and No ram expansion sign me up

How is it not expandable? You take a camm module out, you slap a bigger module in.
While it is not as simple as inserting a new dimm module into an empty slot, dimms aren't without hassle when mixing modules.

 

[PixelTech]

I would absolutely buy a Tachyon. In fact I’ve stated several times on the forums this past year in news articles related to it, but the b650 tachyon never saw a US release (or anywhere for that matter), and gigabyte have otherwise been twiddling their thumbs.

Until there are tangible benefits, camm is a dud for numerous reasons. It’s at the point where we have no consumer boards after 3 years and in terms of performance it’s not even better than cudimm. Innovation for the sake of innovation isn’t always a good thing. Fragmenting standards make the entire ecosystem worse for consumers.

There has yet to be any real, documented evidence of camm being better, just empty promises.

Your aware this article features CAMM2 RAM on the motherboard and not CAMM RAM, right? Leaving off the 2 is not short handing.

 

[rv8000]

Your aware this article features CAMM2 RAM on the motherboard and not CAMM RAM, right? Leaving off the 2 is not short handing.

Yep, any variation (because they ironically can’t agree/settle on a standard) in the desktop space is equally useless. Semantics aren’t going to change that fact.

 

[PixelTech]

Yep, any variation (because they ironically can’t agree/settle on a standard) in the desktop space is equally useless. Semantics aren’t going to change that fact.

Mm, I'm seeing ur argument. There's some disagreement or no collaboration between motherboard manufacturers on the "type" of CAMM2 to use.
Is it fair to say, that because CAMM2 hasn't been finalized or mass produced it's too soon to judge?
This is Computex, some of the tech is prototypes for what companies are working on and developing. While some tech is products that are on the cusp to release to the market.

 

[rv8000]

Mm, I'm seeing ur argument. There's some disagreement or no collaboration between motherboard manufacturers on the "type" of CAMM2 to use.
Is it fair to say, that because CAMM2 hasn't been finalized or mass produced it's too soon to judge?
This is Computex, some of the tech is prototypes for what companies are working on and developing. While some tech is products that are on the cusp to release to the market.

We saw the same exact shenanigans at computex… last year (they demo’d the z890 OCF/some msi boards with camm2 variants, two different socketing types actually). They’ve spun their wheels and gone nowhere.

We just had a new article with a G.skill (might be a diff brand) 10000 mt/s camm2 module at c56, thats worse than 8000 c36/c38 let alone what some cudimm kits are capable of when it comes to latency and frequency.

 

[TheDeeGee]

How is it not expandable? You take a camm module out, you slap a bigger module in.
While it is not as simple as inserting a new dimm module into an empty slot, dimms aren't without hassle when mixing modules.

Indeed, and these days a bunch of motherboards only have two DIMM slots, so they have the exact same "issue" going on.

It's a non-issue really, it's no different than upgrading your GPU.

 

[OneMoar]

Indeed, and these days a bunch of motherboards only have two DIMM slots, so they have the exact same "issue" going on.

It's a non-issue really, it's no different than upgrading your GPU.

And that is exactly the problem want to go from 32 to 64gb ? great throw out the module you have and replace it

dumb standard as far as desktops are concerned especially considering there is zero benefit atm
id be a different story if it was twice as fast or something but IMCS can't operate at the speeds required to care about it atm

 

[Pizderko]

Meanwhile cu-dimm doing the same and higher speeds, without having to deploy a new standard thats going to cost the end user $$$.

Sick of hearing about camm, belongs in laptops, remains to be absolutely no useful proven benefit.

CAMM2 is thinner with more capacity on board if compared with any cu-dimms.
So relax bro, CAMM2/LPCAMM2 are pretty fine.

 

[Hecate91]

CAMM2 is thinner with more capacity on board if compared with any cu-dimms.
So relax bro, CAMM2/LPCAMM2 are pretty fine.
View attachment 400685

Module density makes sense for notebooks, but not for full size desktop motherboards.
If CAMM2 is supposed to replace DIMMs then I would want more than a single CAMM slot on a board.

 

[rv8000]

CAMM2 is thinner with more capacity on board if compared with any cu-dimms.
So relax bro, CAMM2/LPCAMM2 are pretty fine.
View attachment 400685

What does sodimm have to do with any of this? A Camm module that size would take up 2x board area on an atx mb, let alone you could run 4x64gb sticks for twice the capacity. So yes, camm should stay out of desktop and in laptops/mobile where it actually makes sense.