News Dell's Proprietary DDR5 Module Locks Out User Upgrades

[Admin]

Twitter user iGPU Extremist tweeted photographs of Dell's Compression Attached Memory Module (CAMM) for the upcoming Precision 7670 laptop.

Dell's Proprietary DDR5 Module Locks Out User Upgrades : Read more

 

[hotaru251]

And a new reason to avoid a Dell machine.

Removed

 

[Kamen Rider Blade]

Screw Dell & creating another Proprietary Standard.

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

 

[escksu]

No big deal actually. Most thin/light laptops don't even have RAM slots today. The RAM modules are soldered directly onto the mainboard. Some even solder their SSD to the board instead of putting it on a "gum stick".

Btw, for those who don't know, DEll precision are workstation laptops. They are not intended for home users.

 

[ezst036]

Screw Dell & creating another Proprietary Standard.

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

The last thing that got politicized, it got worse. I don't see how politicizing it will make it better.

Anybody remember the definition of insanity and doing things over and over again?

 

[InvalidError]

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

You won't have to worry about memory standards anymore 6-8 years from now: by the time DDR5 is no longer fast enough to keep up, the first 8-32GB of system memory will be built into the 3D stack and memory expansion may get delegated to 6.0x4/8/16 PCIe RAM now that it has been downgraded to victim cache status.

 

[Giroro]

No big deal actually. Most thin/light laptops don't even have RAM slots today. The RAM modules are soldered directly onto the mainboard. Some even solder their SSD to the board instead of putting it on a "gum stick".

Btw, for those who don't know, DEll precision are workstation laptops. They are not intended for home users.

Be you rest assured, the models with soldered RAM will continue to have soldered RAM.
This is so Dell can have the cost savings of configurable RAM, combined with revenue boost of forcing customers into an overpriced upgrade at time of purchase.
Dell almost certainly will not sell these modules to end users, but if they do they will be many times more expensive than the industry standard.

I'm really getting sick of all the extreme anti-customer monetization that tech companies are allowed to get away with. Duopolies/oligopolies aren't any better than Monopolies, and price-fixing/anti-trust laws are never enforced.

 

[TerryLaze]

You won't have to worry about memory standards anymore 6-8 years from now: by the time DDR5 is no longer fast enough to keep up, the first 8-32GB of system memory will be built into the 3D stack and memory expansion may get delegated to 6.0x4/8/16 PCIe RAM now that it has been downgraded to victim cache status.

LOL..no.
Some 60Mb of cache already causes so many problems and so much increase in price, just imagine that for 8Gb...let alone the space required.
Look at a 8Gb ram stick, all of these chips would have to fit on a CPU die.

 

[teodoreh]

Every time a major company makes something stupid, it's bad for the consumer. This thing reminds me of Apple creating a propientary m.2 slot that couldn't accept usual m.2 drives.
I've seen all sort of <Mod Edit> things - PSUs with propientary size or cables, motherboards with different pins that cannot be placed on standard ATX case, etc. It's all idiotic and destroys the very essence that made PC technology shine over all other 500 standards: Compatibility and expansion potential.

 

[mdd1963]

Best to get one with enough RAM to start with then, as typically , one can anticipate /calculate Dell pricing on any additional item by taking a fair price for an object like a RAM stick or SSD, double that figure, and, for good measure, double it again. Eventually, you will see /reach Dells price for a RAM module.

 

[danger007]

When will companies stop their own preferred standards and get it together to where, especially with laptops/notebooks/tablets, that by having the same standard across all companies they would sell many more laptops as people wouldn't have to worry about proprietary connections. I would be great to see them agree on the "case design even" so that users can have the chance to repair or replace boards, but of course the companies don't want you to do that as they depend on the after market (meaning after you buy their laptop) sales of batteries, various parts like in this case memory and so on.d I wish one day they follow their desktop brethren so we can make choices about the build and features and can assemble the parts in a notebook case because they have standards that are the same across all notebook and laptops

 

[ottonis]

A proprietary RAM module is light-years better than soldered RAM. The latter is not replaceable, let alone upgradeable. The former may even bring some architectural advantages for the laptop, e.g. saving up space and allowing for a slimmer device or for providing more M2 slots.
The best thing Dell could do in the future is to open their specifications to third party vendors so the customer may choose between different offers.
We will see what the future is going to bring.

 

[InvalidError]

LOL..no.
Some 60Mb of cache already causes so many problems and so much increase in price, just imagine that for 8Gb...let alone the space required.
Look at a 8Gb ram stick, all of these chips would have to fit on a CPU die.

While 60MB of SRAM may be mildly problematic, 32GB of DRAM a few years from now should be trivial when HBM3 is aiming for up 64GB at 819GB/s per stack right now. The main challenge is achieving DDR4-like economies of scale to amortize tooling costs. Should be easy enough once we reach a point where every new CPU, GPU, SoC, etc. under the sun needs at least one to meet RAM bandwidth, size and power efficiency requirements. Should IGPs opportunistically scale up with node shrinks, you aren't going to meet 2028 IGPs' memory bandwidth requirements with DDR5 and I'm sure having over 1TB/s of memory bandwidth to a 16+GB local memory pool to load and flush cache lines would enable CPU engineers to pull a few tricks too..

When mainstream DDR5 maxes out at 6.4-8GT/s a few years from now, you will be looking at 10+GT/s before the next generation has a chance of delivering a meaningful performance benefit. That sort of bandwidth across a 128bits wide memory interface that has to cross a 4-5" long total path across the CPU substrate + socket + motherboard + memory slot + DIMM PCB will be extremely difficult to achieve within reasonable power, silicon and PCB materials budgets. It is already proving quite problematic with DDR5, especially when attempting to use two DIMMs per channel. GPUs can manage 18+GT/s across 256+bits wide busses only because the traces are much shorter (~1.5") and the manufacturer has nearly absolute control over everything since it is all soldered to the PCB. Also worth keeping in mind is that GDDR6 needs roughly 10X as much power per chip as standard-volted DDR4/5 does to do its thing. Pretty steep price there and most of it is due to all of the extra signal processing needed to maintain signal integrity at both ends at such crazy speeds. HBM-like interfaces ditch most of that overhead by effectively eliminating traces and most of the power associated with driving or terminating them.

My bet is that DDR5 will be the practical limit for high-speed socketable parallel memory interfaces. Whatever comes next will focus on supplementing on-chip memory, power efficiency, density and cost instead of raw performance.

 

[Kamen Rider Blade]

My bet is that DDR5 will be the practical limit for high-speed socketable parallel memory interfaces. Whatever comes next will focus on supplementing on-chip memory, power efficiency, density and cost instead of raw performance.

My prediction for the future is that IBM's OMI (Open Memory Interface) should be the path foreward, but will still use standard Parallel DIMM's.

The only difference is the Memory Controller will get moved out from the cIOD into it's own chip and be physically next to the DIMM's and will feed back to the cIOD via a Serial Interface with minimal latency penalty like MicroChip's implementation.

This should lower the amount of traces needed from cIOD to Off-Die Memory Controller and allow more bandwidth via more Memory Channels.

It'd also allow the Serial Connection to be "Memory Agnostic" and you can finally use GDDR, DDR, or even HBM if you wanted to.

 

[Aaron44126]

Bad reporting? It does not lock out upgrades in this case. First off, you can buy additional CAMM modules directly from Dell. Second, they are selling a SODIMM interposer so you can use 2× SODIMM modules in place of the CAMM module if you like.

 

[escksu]

Be you rest assured, the models with soldered RAM will continue to have soldered RAM.
This is so Dell can have the cost savings of configurable RAM, combined with revenue boost of forcing customers into an overpriced upgrade at time of purchase.
Dell almost certainly will not sell these modules to end users, but if they do they will be many times more expensive than the industry standard.

I'm really getting sick of all the extreme anti-customer monetization that tech companies are allowed to get away with. Duopolies/oligopolies aren't any better than Monopolies, and price-fixing/anti-trust laws are never enforced.

No big deal at all... As i said, this is dell precision workstation laptops. They are not meant for end-users.. only companies with the need for portable workstations will buy them.

Proprietary stuff is never an issue for such corporate laptops/workstations. Companies dont diy and neither do they upgrade their workstations/laptops. Should the need arise, they buy a new one instead...

Btw, overpriced doesnt exist for computer hardware...the cost is trivial compared to the software the workstation is using. Windows enterprise and office is now basically subscription based... The amount i have to pay for every user a year is more than enough for a brand new laptop....

Maintenance fees for specialised software cost even more. Then even more expensive ones like SAP can cost millions a year... Enough to pay for many workers salary.

So what is hardware cost compared to software?? Nothing....

 

[jasonf2]

If there is a need for a smaller form factor RAM module these companies need to develop it and apply it through the already established standards organizations. All that this does when one vendor rolls out a proprietary module is leave customer machines on a parts island after a couple of years when Dell makes the cost decision to abandon the form factor. RIMM anyone?

 

[ThatMouse]

Thou shall not sell laptops with only 8GB RAM.
Thou shall not sell the 16GB model for $300 more.

 

[jp7189]

Screw Dell & creating another Proprietary Standard.

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

There is... its called the "law" of free market. If people don't like it, they don't buy it, and the model becomes unprofitable.

Realistically people that buy this type of laptop get it for the certified compatibility with software that costs $10's thousands. That crowd isn't going to care about saving $50 on a memory upgrade if it risks the certification.

 

[traxxmy]

Screw Dell & creating another Proprietary Standard.

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

There is no law force you to buy in first. Really biased thinking

 

[USAFRet]

Screw Dell & creating another Proprietary Standard.

There needs to be laws / regulations preventing companies from making "Proprietary Standards".

You are free to vote with your wallet.

I'd wager the majority of Dell customers never change or add to the RAM in their system.

 

[tennis2]

I'd wager the majority of Dell customers never change or add to the RAM in their system.

This is true. And a fair number of corporations buy RAM upgrades from Dell anyway.

Which also makes me wonder if this is really a move to pull in an additional [perceivably] small chunk of revenue/profits from DIYers, or if it really is just another [silly] custom design to shave an extra 10 seconds off assembly time at the cost of paying for custom hardware.
The DIYers that care about adding DRAM are probably going to continue looking at [notebookcheck] reviews before purchase to ensure user-replaceable RAM and avoid buying/recommending these models (resulting in lost revenue). Everyone else will just buy it regardless of what the modules look like because they probably wont do a RAM upgrade anyway.

I don't see a world where this replaces soldered DRAM.

 

[InvalidError]

The only difference is the Memory Controller will get moved out from the cIOD into it's own chip and be physically next to the DIMM's and will feed back to the cIOD via a Serial Interface with minimal latency penalty like MicroChip's implementation.

Even if you do that, you still need a memory-controller-like blob in the CPU/GPU/SoC/etc. to juggle all of the memory IO requests in accordance with application-specific needs and priorities, tag and package them before forwarding to the off-chip controller managing physical access then handle the responses, so this is more like splitting the memory controller into a MAC-PHY pair than completely moving the memory controller off-chip. Even if memory expansion was moved to PCIe, you'd just end up with the CPU's PCIe controller having to arbitrate IO to the memory expansion cards in accordance with the CPU's policies and basically becoming the defacto memory MAC.

While something OMI-like may be the future of off-chip memory, on-package memory provides ~10X as much bandwidth per watt. It will be really difficult for any company wanting to improve their products' power efficiency while increasing performance to pass on it once 2.5/3D-packaging becomes the norm. For something like IBM's Z-series which can have dozens of such modules per drawer, letting modules sleep more often by keeping most of the working set on-package could save hundreds of watts per rack.

 

[bigdragon]

I hope that CAMM does not stay proprietary for long. I remember solutions like MXM and ZIF. The widest possible adoption is necessary for things like CAMM to survive outside of a couple of product generations at a single vendor. Open standards are king. I do appreciate that Dell is trying to condense the footprint of memory and retain modularity.

 

[Eximo]

My bet is that DDR5 will be the practical limit for high-speed socketable parallel memory interfaces. Whatever comes next will focus on supplementing on-chip memory, power efficiency, density and cost instead of raw performance.

Probably why they got together to make a standard for gluing chips together. They know it is coming too.

https://www.tomshardware.com/news/new-ucie-chiplet-standard-supported-by-intel-amd-and-arm