CAMM to Replace the Decades-old SO-DIMM Laptop Memory, JEDEC and Dell Argue

[AleksandarK]

Laptop memory has been a controversial topic for many years. Its proprietary standard, SO-DIMM, has shown signs of aging as the decades-old JEDEC standard didn't adapt to other expanding capacities and speed trends. Today, JEDEC and Dell think that the future of laptop memory is in the new CAMM standard that both companies are working on. The introduction of CAMM comes from Dell, whose Senior Distinguished Engineer Tom Schnell is working on it. "We have unanimous approval of the 0.5 spec," said Mr. Schnell for PCWorld. One of the problems that SO-DIMM is facing is the capacity and speed issue, where the current DDR5 SO-DIMM memory stops at around 6400 MT/s. CAMM, on the other hand, starts from that speed and works its way up to offer higher capacities as well.

With Dell introducing CAMM in its laptops, it had no intentions of creating a proprietary solution but rather an expandable and upgradable memory platform with various benefits. With JEDEC's involvement in finalizing this, the CAMM standard is slowly on its way to becoming a viable option for different laptop manufacturers. Dell's Tom Schnell didn't reveal what companies are in the process of creating the final specification; however, we know that 32 of them are present, including Apple. If others join, the standard could take over future laptop designs and offer higher speeds and higher capacities, especially in the mobile workstation space where it matters. Below is an example of a CAMM memory module with a patent showing the SO-DIMM (upper left) versus CAMM (lower right) and CAMM's smaller trace path. With smaller tracing, the latency is also going down, so the new standard will bring additional efficiency. Additionally, devices that are based on LPDDR memory could have an upgrade path with the installment of CAMM.



View at TechPowerUp Main Site | Source

 

[Solaris17]

Dell's Tom Schnell didn't reveal what companies are in the process of creating the final specification

Guess it’s not so “unanimous” then.

Today, JEDEC and Dell think that the future of laptop memory is in the new CAMM standard that both companies are working on.

Tell me more.

The introduction of CAMM comes from Dell, whose Senior Distinguished Engineer Tom Schnell is working on it.

Ah ok, so Dell is trying to strong arm the industry and some dude is trying to complete his next thesis got it.

 

[TheLostSwede]

Guess it’s not so “unanimous” then.

That's normally how these "standards" working groups behave, a lot of arguing in favour of company X's advantages over company Y and why their implementation is better and more important. I've been to some of those meetings and it's not pleasant.

Ah ok, so Dell is trying to strong arm the industry and some dude is trying to complete his next thesis got it.

Yes and no. Someone has to develop what becomes a future standardard at some point or things won't progress. Once they have a solution, they need to argue the advantages of said solution and try to convince others it's a good standard.

I'm not sold on the CAMM form factor, as although it's low profile, it also seems to be a bit over-engineered and some of the larger form factors aren't what I'd call appealing, but if we're going to go past 128 GB of RAM in a laptop, it might be needed. Not sure if Dell is planning on using it in SFF type devices too, but only in laptops doesn't quite make sense.
The upside of CAMM is that we might see laptops with LPDDR memory having an upgrade path.

 

[thewan]

Ah ok, so Dell is trying to strong arm the industry and some dude is trying to complete his next thesis got it.

Good thing the internet didnt exist back when SIMMs and DIMMs were created by some dude from a big name company. If it did we would get this guy complaining about it before and get dejavu all over again.

 

[Solaris17]

Good thing the internet didnt exist back when SIMMs and DIMMs were created by some dude from a big name company. If it did we would get this guy complaining about it before and get dejavu all over again.

Not at all. Density is an issue, but Swede likes to teach just to teach. I know how these technologies come about already. The greater technology “forums” are literally just public facing bodies of these companies in the capacity to create standards.

I just don’t see this case as anything but an attempt for one man to better himself. I would love to read the technical specifications and documentation submitted because I would be curious how many of the density and form factor issues the “other” companies expressed are addressed in this persons initial design, and how many of Dells are addressed in comparison.

If at all of course. For all I know everyone is playing nice and this man is a literal saint with a heart of gold

 

[kapqa]

think we can all agree that soldered ram is not the way forward ... so CAMM welcome!

 

[Solaris17]

think we can all agree that soldered ram is not the way forward ... so CAMM welcome!

I would like todays laptops to have replaceable ram for sure

 

[trsttte]

That small trace path illustration is misleading as hell, yeah on that specific example the traces are longer, but if they turn the socket position 180º like with the CAMM module it would get smaller. They could still make the same argument without being so misleading about it lol

I'm not sold on the CAMM form factor, as although it's low profile, it also seems to be a bit over-engineered and some of the larger form factors aren't what I'd call appealing, but if we're going to go past 128 GB of RAM in a laptop, it might be needed. Not sure if Dell is planning on using it in SFF type devices too, but only in laptops doesn't quite make sense.

Another problem is, at least when they originally showed the CAMM plans, there were way too many different module phisical sizes. 2 could be acceptable (one for higher capacities 128gb and above and one for 64gb and lower for example - this aligns with the sizes they showed, the 64gb fitting in a "half length" module and 128gb on a "full lenght") but 4 or 5 different sizes is just silly.

 

[TheLostSwede]

That small trace path illustration is misleading as hell, yeah on that specific example the traces are longer, but if they turn the socket position 180º like with the CAMM module it would get smaller. They could still make the same argument without being so misleading about it lol

I think you misread the diagram. That's four SO-DIMMs in total, two over and two under, which makes for longer trace paths, but it's also a very rare configuration in a modern laptop, except a few high-end workstation or gaming models. You couldn't turn the socket around in this case, as it wouldn't help, since two of the SO-DIMMs would always be further away from the CPU/SoC.

However, in most "normal" scenarios, it wouldn't make any difference, as long as regular DDR memory is being used. That said, it seems like the SO-DIMM design can't go above 6400 MT/s due to signal integrity issues.

Another problem is, at least when they originally showed the CAMM plans, there were way too many different module phisical sizes. 2 could be acceptable (one for higher capacities 128gb and above and one for 64gb and lower for example - this aligns with the sizes they showed, the 64gb fitting in a "half length" module and 128gb on a "full lenght") but 4 or 5 different sizes is just silly.

Agreed, as Dell presented four different sizes if I remember right, plus the converter for SO-DIMMs. Even though the screw holes are places in the same spaces, the physical size difference is going to be too big depending on how notebooks are designed, that it might prevent upgrades in the future.
However, keep in mind that there will only be one CAMM module in a laptop, as each CAMM module already operates in dual-channel mode. The really small modules might end up being limited to only 32 GB, depending on the DRAM chip density, as they only seem to house eight chips in total, or four per channel. In fact, the orignal small module was only 16 GB as per the picture below.
Admittedly the diagram in this news post shows double sided CAMM modules, it doesn't appear that Dell's design catered for that, or at least they haven't shown any CAMM modules with memory on the bottom.

 

[zlobby]

Jesus, look at the footprint of that thing! It looks like another mobo for crying out loud! And Dell is proud with that?

I would like todays laptops to have replaceable ram for sure

More like sCAMM...

 

[TheLostSwede]

Jesus, look at the footprint of that thing! It looks like another mobo for crying out loud! And Dell is proud with that?

The smallest CAMM module is about the size of two SO-DIMMs next to each other.
Keep in mind that the SO-DIMMs stick out a fair bit from the slots.
And yes, this is a mechanical adapter.

 

[Denver]

When HBM was created I thought that after some iterations and refinements, at some point it would be a replacement for RAM and/or Vram.. now reading this article I am disheartened to see how advances in the hardware sector are stagnating. I only see things that do little or no benefit to the consumer.. Meh.

 

[Vya Domus]

Considering with what proprietary abominations Dell came up with over the years in the computers they sell I wouldn't want to hear their name in the same sentence with the word "standard".

 

[zlobby]

The smallest CAMM module is about the size of two SO-DIMMs next to each other.
Keep in mind that the SO-DIMMs stick out a fair bit from the slots.
And yes, this is a mechanical adapter.

That's more like a point in my case.

 

[TheLostSwede]

That's more like a point in my case.

Well, that's only two SO-DIMMs though, wheras one CAMM should replace four, according to Dell, although maybe not in that size CAMM.
At least that should be the 32 GB one, if we go by the sizes in my previous post, which should be at least 64 GB for DDR5.

 

[zlobby]

Well, that's only two SO-DIMMs though, wheras one CAMM should replace four, according to Dell, although maybe not in that size CAMM.

Still, no matter how I measure, the trace lengths are still the same. Not to mention the mobo-on-mobo thing. But hey, maybe it's just me?

 

[Chrispy_]

There's a lot of negativity here for CAMM but it's better than soldered, which is what we'll get if CAMM isn't adopted. Forget SoDIMM, it's going to be CAMM or nothing soon.

SoDIMM is done, the DIMMs themselves are too bulky for modern laptops, and the slot is even taller. SoDIMM slots make the traces to the IMC longer, which reduces signal integrity and knocks at least a GT/s off the memory speeds which is a big loss just for the sake of flexibility which many buyers will never even take advantage of.

Additionally, laptop BIOSes on most thin & lights are almost exclusively worthless when it comes to XMP or other RAM clock/timing options for SoDIMMs. You're getting the JEDEC defaults and that's it, which makes so many of the faster/tighter SoDIMM kits pointless because your laptop BIOS is just going to run them at super-slow, super-safe, lowest-common-denominator timings. DDR4-2133 and DDR5-4800 etc.

 

[TheLostSwede]

Still, no matter how I measure, the trace lengths are still the same. Not to mention the mobo-on-mobo thing. But hey, maybe it's just me?

Did you take into consideration that the CAMM is dual channel vs SO-DIMM being single channel?

 

[Wirko]

Old people here, does anyone remember a connector that required tightening the screws in order to make contact?

(Many are just secured by screws, think VGA, DVI and probably those on MXM GPU modules, and industrial types too.)

 

[zlobby]

Old people here, does anyone remember a connector that required tightening the screws in order to make contact?

(Many are just secured by screws, think VGA, DVI and probably those on MXM GPU modules, and industrial types too.)

Meh, I remember welding them.

Did you take into consideration that the CAMM is dual channel vs SO-DIMM being single channel?

Not quite, I must admit. But still, there is no way I support this (s)CAMM just because of its aesthetics.

 

[user556]

The article isn't saying what's really changed - The pin count! The single connector must have an enormous grid of contacts.

 

[Wirko]

The article isn't saying what's really changed - The pin count! The single connector must have an enormous grid of contacts.

14 x 44 = 616

 

[TheinsanegamerN]

There's a lot of negativity here for CAMM

That's what happens when inferior proprietary ideas are shoved down consumers throats instead of the perfectly acceptable solution we already have.

but it's better than soldered, which is what we'll get if CAMM isn't adopted. Forget SoDIMM, it's going to be CAMM or nothing soon.

SoDIMM is done, the DIMMs themselves are too bulky for modern laptops, and the slot is even taller.

CAMM is barely any smaller then two SODIMMs. There's barely any savings there. As pointed out by TheLostSwede, all the modules over 16GB are larger then two SoDimms. So if you use dense memory it's actually BIGGER then the current solution.

SoDIMM slots make the traces to the IMC longer, which reduces signal integrity and knocks at least a GT/s off the memory speeds which is a big loss just for the sake of flexibility which many buyers will never even take advantage of.

Funny how desktops with longer trace lengths have no issue pushing far higher speeds then laptops.

Additionally, laptop BIOSes on most thin & lights are almost exclusively worthless when it comes to XMP or other RAM clock/timing options for SoDIMMs. You're getting the JEDEC defaults and that's it, which makes so many of the faster/tighter SoDIMM kits pointless because your laptop BIOS is just going to run them at super-slow, super-safe, lowest-common-denominator timings. DDR4-2133 and DDR5-4800 etc.

This is a red herring argument, as NONE of this changes with the use of CAMM instead of SODIMMs.

 

[Fourstaff]

Making something a standard helps but doesn't guarantee a widespread adoption. If there are no better alternatives than CAMM they are going to win almost by default.

 

[bonehead123]

JEDEC and Dell think that the future of laptop memory is in the new CAMM standard that both companies are working on

Funny, I always thought JEDEC was a governing body instead of a "company", so what this probably means is that duhHell is pushin their "standard" thru JEDEC to get it accepted as such, then they will come up with a way to make it proprietary, so everyone will have to pay them a license fee to use it... go figure

"Dude, your getting duhHell memory" (mainly cause theirs is the only one that works in your machine!)