Dell is pushing CAMM, a priorietary RAM module format

[Etern4l]

Dell defends CAMM, its controversial new laptop memory

Just came across this gem.. where Dell can't solder RAM, they will use a proprietary module lol

 

Of course, good old Gordon to the rescue, as always.

 

@Papusan @Mr. Fox  @Aaron44126 @ssj92

[Aaron44126]

Old news? There was a lot of chatter about this in the Precision threads when it first came up. (My own laptop has a CAMM module.) Since then, JEDEC has blessed CAMM as the next standard for laptop memory and other manufacturers are expected to start using it. (SODIMM is going to have signaling issues as RAM speeds continue to ramp up, so something has to be done, and this is better than soldered memory at least.)

https://www.extremetech.com/computing/342342-jedec-says-dells-camm-laptop-memory-tech-will-replace-so-dimm

[ssj92]

Yeah ironically this format is actually better than SODIMM for higher speed memory. It's why laptop SODIMM is stuck under 6Ghz whereas desktop is already at 8Ghz. 

 

If it does become standard at least it should hopefully allow higher speeds and capacity. 

 

I still find it weird DDR5 SODIMM was developed and no one thought about there being a bottleneck at higher speeds. 

[Mr. Fox]

Anything that is a proprietary solution should be treated like a pile of dung and avoided like a plague. Even if it is hypothetically "better" by some measurements. I mean, think about it. Does anyone lack the intelligence to truly believe that Dell, HP, Lenovo, Acer or any other turdbook manufacturer gives a rat's ass about better? The only thing they care about is controlling the use of the products they sell and controlling what parts can be used to repair or upgrade it. Anyone that believes otherwise should look into the ocean beach property I have for sale on the east side of Phoenix. I am offering it at a really good deal.

[Ionising_Radiation]

1 hour ago, Mr. Fox said:

proprietary solution

But... it's not a proprietary solution. @Aaron44126 literally just said Dell offered it to JEDEC as a standard connector/form factor for notebook DRAM, which has been accepted and even expanded by JEDEC.

[Papusan]

1 hour ago, Ionising_Radiation said:

But... it's not a proprietary solution.

Yep, but they intend to milk money out from the change... And Dell don't do this changes for free. We talk about Dell.

 

Some here that still remember the Area-51M lapjoke flagship with memory running below Intel's mem specs? And Dell's new mobile GPU design that should be the future for people/gamers that wanted to upgrade their gaming Jokebooks with new graphics cards? Yep, we talk about the same Tech company.

 

In the future, a person with a CAMM-equipped laptop will be able to buy RAM from any third party and install it in the laptop. Yes, initially, Dell will likely be the only place to get CAMM upgrades, but that should change as the standard scales up and is adopted by other PC makers. The new memory modules are also built using commodity DRAMs just like conventional SO-DIMMs.

 

Dell does indeed hold patents on the CAMM design and there will be royalties, but the company says it’s too early to discuss royalties....

 

Would you trust Putin, Hitler and Stalin? Or the Devil?

 

5 hours ago, Etern4l said:

Dell defends CAMM, its controversial new laptop memory

Just came across this gem.. where Dell can't solder RAM, they will use a proprietary module lol

 

Of course, good old Gordon to the rescue, as always.

 

@Papusan @Mr. Fox  @Aaron44126 @ssj92

 

Beyond the performance advantage, Schnell said CAMM will improve reliability, improve thermals, and save space, which can make it possible to make a thinner laptop. Yep, the goal to reach Apple design continue at full speed.

[Etern4l]

  

5 hours ago, Aaron44126 said:

Old news?

 

Dated April 28. Would be reasonable to assume Gordon got wheeled out for a reason, perhaps they need more of the industry onboard or something? Yays from 20 or so out of 332 JEDEC members doesn't sound like much.  Anyway, apparently nothing has been published yet, so it's not an official open standard (the sources below estimate it might happen in Q3 2023).

 

https://www.makeuseof.com/camm-vs-sodimm-ram/?newsletter_popup=1

https://www.notebookcheck.net/SO-DIMM-laptop-RAM-form-factor-to-soon-be-replaced-with-Dell-developed-CAMM-standard.682337.0.html

 

The way I'm reading the above, Dell developed a new proprietary RAM module "in (some form of) collaboration with JEDEC" and pushed it out to the market prior to any standarization being complete. The actual final standard is likely to differ from Dell's design. Whether the standarisation is actually finalised, the date of that, and whether the JEDEC implementation is compatible with the Dell's design are technically open questions.

[Mr. Fox]

NVIDIA had their totally unnecessary and essentially stupid 12VHPWR cable, now this Dell-backed RAM module morphodite abortion. It's called creating ways to make money and pretending it's better so stupid people will agree/buy it. Necessity is not a component of that, unless you coumt the need of big greedy companies to make more money. They might, but I don't. Dell is the grand master of dreaming up goofball form factors that make them the only source for replacement parts.

[Aaron44126]

55 minutes ago, Etern4l said:

  Dated April 28.

…From last year.  I posted a link to this very article in the Precision 7770 prerelease thread here when it was newly posted.

[Etern4l]

58 minutes ago, Aaron44126 said:

…From last year.  I posted a link to this very article in the Precision 7770 prerelease thread here when it was newly posted.

 

Lol, I'm too tired. The rest of the post looks valid as of today: not a standard yet, CAMM modules are proprietary and not available on the open market. 

[Aaron44126]

2 hours ago, Etern4l said:

Lol, I'm too tired. The rest of the post looks valid as of today: not a standard yet, CAMM modules are proprietary and not available on the open market. 

 

Yes, it is true.  Precision 7X70 and 7X80 systems use CAMM modules before the standardization is done, and as you say, there is a chance that the standard could evolve before it becomes more widely used so future CAMM modules would not work in these systems.  Dell launched these systems with CAMM to demonstrate their commitment to the standard and that the modules work.  You aren't forced to use CAMM in these systems; you can order these systems with a SODIMM interposer and use SODIMM modules instead.  Several users here have done so in order to save some money or install their choice of modules.  I understand the reservations but I can't really fault Dell's approach here.  They're the only manufacturer to get ahead of the problem with SODIMM and put out something that works (other than soldered modules), and they got JEDEC to go along with it.

 

(There's a downside to choosing SODIMM in these systems, though; with only two modules, 64GB is the memory cap... maybe 96GB soon, though, as it does look like 48GB SODIMMs aren't that far off.)

[Etern4l]

1 hour ago, Aaron44126 said:

 

Yes, it is true.  Precision 7X70 and 7X80 systems use CAMM modules before the standardization is done, and as you say, there is a chance that the standard could evolve before it becomes more widely used so future CAMM modules would not work in these systems.  Dell launched these systems with CAMM to demonstrate their commitment to the standard and that the modules work.  You aren't forced to use CAMM in these systems; you can order these systems with a SODIMM interposer and use SODIMM modules instead.  Several users here have done so in order to save some money or install their choice of modules.  I understand the reservations but I can't really fault Dell's approach here.  They're the only manufacturer to get ahead of the problem with SODIMM and put out something that works (other than soldered modules), and they got JEDEC to go along with it.

 

(There's a downside to choosing SODIMM in these systems, though; with only two modules, 64GB is the memory cap... maybe 96GB soon, though, as it does look like 48GB SODIMMs aren't that far off.)

 

To me it looks like they jumped the gun with a proprietary solution. I am not sure about MSI Titan which also supports 128GB, but Lenovo ThinkPad T16 runs 128GB at 4800MHz on SODIMMs, whereas apparently Precision 7770 is advertised to support 128GB at 3600Mhz using "thinner and faster" CAMM... Not great - perhaps it's actually more about the thickness than speed?

[Aaron44126]

We also hashed this over in Precision 7X70 pre-release thread last year.  The 3600 MHz limitation with 128GB is a limitation of 12th gen Alder Lake, not the CAMM module.  (There's an Intel PDF document linked in there confirming the limitation.)  There is tons of misleading support material out there from multiple laptop manufactures (even Dell's own marketing material in some cases indicates that it should work at a higher speed than 3600 MHz).

[Etern4l]

Maybe it's a limitation on the laptop side, but desktop Alder Lake was capable of running 128GB at around 4800MHz, although after a good few bios updates and not with every kit. It was a bit of a mess initially.

 

Edit: On the other hand, DDR4 laptop RAM was always slower, so perhaps we can expect the same from DDR5. I guess is just must have to do with the size of the modules. A DIMM is roughly twice the size of a SODIMM (and there is even a larger discrepancy vs the even smaller CAMM?) specifically, the chip surface is much smaller. Temps, for instance? DDR5 can run hotter. That would potentially make the purported benefits of CAMM moot if not counterproductive, especially for higher capacity modules which run slower anyway. We'll see, although I would personally steer clear of this for now. Not sure if miniaturising and compressing is a step in the right direction from the performance/enthusiast perspective, whatever Dell claims. They didn't actually showcase the performance benefit by offering any RAM over 5200/5600 speeds, did they? In contrast, people are pushing 8200 on the desktop side, and 128/196GB runs at 5200, using the humongous in comparison DIMM modules.

[Mr. Fox]

17 hours ago, Etern4l said:

Maybe it's a limitation on the laptop side, but desktop Alder Lake was capable of running 128GB at around 4800MHz, although after a good few bios updates and not with every kit. It was a bit of a mess initially.

 

Edit: On the other hand, DDR4 laptop RAM was always slower, so perhaps we can expect the same from DDR5. I guess is just must have to do with the size of the modules. A DIMM is roughly twice the size of a SODIMM (and there is even a larger discrepancy vs the even smaller CAMM?) specifically, the chip surface is much smaller. Temps, for instance? DDR5 can run hotter. That would potentially make the purported benefits of CAMM moot if not counterproductive, especially for higher capacity modules which run slower anyway. We'll see, although I would personally steer clear of this for now. Not sure if miniaturising and compressing is a step in the right direction from the performance/enthusiast perspective, whatever Dell claims. They didn't actually showcase the performance benefit by offering any RAM over 5200/5600 speeds, did they? In contrast, people are pushing 8200 on the desktop side, and 128/196GB runs at 5200, using the humongous in comparison DIMM modules.

It would be logical for a number of reasons. Thermal management is already nearly impossible with pre-DDR5 laptops. The need for increased voltage and access to memory settings, a good CPU sample and superior motherboard design are the reality on desktops. The likelihood of laptop manufacturers getting these extra things right and giving purchasers of their products access to associated BIOS features is like slim to none.

[win32asmguy]

On 4/29/2023 at 9:27 PM, Etern4l said:

To me it looks like they jumped the gun with a proprietary solution. I am not sure about MSI Titan which also supports 128GB, but Lenovo ThinkPad T16 runs 128GB at 4800MHz on SODIMMs, whereas apparently Precision 7770 is advertised to support 128GB at 3600Mhz using "thinner and faster" CAMM... Not great - perhaps it's actually more about the thickness than speed?

 

The other 4 SODIMM systems all perform a bit worse if you are only using two SODIMMs. Basically they operate at 4000MT/s regardless of what faster speeds are supported, with the exception of the GT77 where with the advanced bios you can override the behavior (but also greatly increase boot time).

 

With the Precision 7780 and the CAMM to SODIMM interposer it can boot 64GB Kingston Fury at DDR5-5200 CL38. The same motherboard can also take a 128GB CAMM module if capacity is needed over speed. Of course, it is probably a moot point in terms of performance uplift because this system is also limited by its small cooler, lack of undervolting support, low GPU TGP, 240W power brick, and lack of fan speed control.

[Etern4l]

4 hours ago, win32asmguy said:

 

The other 4 SODIMM systems all perform a bit worse if you are only using two SODIMMs. Basically they operate at 4000MT/s regardless of what faster speeds are supported, with the exception of the GT77 where with the advanced bios you can override the behavior (but also greatly increase boot time).

 

With the Precision 7780 and the CAMM to SODIMM interposer it can boot 64GB Kingston Fury at DDR5-5200 CL38. The same motherboard can also take a 128GB CAMM module if capacity is needed over speed. Of course, it is probably a moot point in terms of performance uplift because this system is also limited by its small cooler, lack of undervolting support, low GPU TGP, 240W power brick, and lack of fan speed control.

 

Yikes, that makes no sense. Would expect a performance penalty when using 4 SODIMMs, but not a general penalty on 4-DIMM systems. Are we saying that Lenovo is falsely advertising 4800 regardless of memory capacity? Anyway, even if the P16 performs as advertised, 4800 with 2 SODIMMs is not great. GT77 is 13th gen so perhaps the new IMC is making some difference, although 12th gen would go up to something like 6600 on some desktops.

[saperkus]

14 hours ago, Etern4l said:

 

Yikes, that makes no sense. Would expect a performance penalty when using 4 SODIMMs, but not a general penalty on 4-DIMM systems. Are we saying that Lenovo is falsely advertising 4800 regardless of memory capacity? Anyway, even if the P16 performs as advertised, 4800 with 2 SODIMMs is not great. GT77 is 13th gen so perhaps the new IMC is making some difference, although 12th gen would go up to something like 6600 on some desktops.

I would say, that Lenovo merely stated which modules will be installed in the device - not their real working parameters. Please check the PSREF documentation of P16. With four SO-DIMM slots, you can't go above DDR5-4000 speeds even if only single module is used. With 4x 32GB the speeds will go down to DDR5-3600.

 

Quote

System comes with DDR5-4800 memory and will run at DDR5-4000 and DDR5-3600 due to platform limitation:
4x 32GB configuration runs at DDR5-3600, other configurations run at DDR5-4000.

 

https://psref.lenovo.com/syspool/Sys/PDF/ThinkPad/ThinkPad_P16_Gen_1/ThinkPad_P16_Gen_1_Spec.pdf (page 3)

 

More information about this issue can be found on the HP ZBook Fury G9 QuickSpecs document (page 11). As long as single DIMM is used per channel, the speeds will be 4000 MT/s. In case of two DIMMs per channel, 4000 MT/s speeds will be retained with single rank modules (8GB and 16GB). With three of four dual rank modules (32GB), speeds will be limited to 3600 MT/s. HP also warned, that if single and dual rank modules will be mixed in the same channel, then speed can drop to 2000 MT/s and system may become unstable.

 

https://h20195.www2.hp.com/v2/GetDocument.aspx?docname=c08160125 (page 11)

 

To put it short, we hit a limitation of SO-DIMM standard. Back in 2021 there already were limitations in case of use of DDR4-3200. In case of mixing vendors or capacity, the speeds were often limited to 2933 MT/s.

 

Quote

Transfer rates up to 3200 MT/s for nECC and ECC memory combinations when memory suppliers are consistent. If suppliers
are not consistent, speeds may drop to 2933 MT/s for nECC and 2933 MT/s for ECC memory combinations.

 

https://h20195.www2.hp.com/v2/getpdf.aspx/c07606964.pdf (page 9)

[ssj92]

It's quite clear DDR5 SODIMM has huge limitations. I am running 128GB 2400Mhz DDR4 RAM in my 2019 Area-51m running a 9th gen CPU but these newer 12th gens can't run 128GB unless it's capped to almost previous gen speeds? lol

 

The new 48GB modules will be nice because it should allow 96GB DDR5 and hopefully at full speed. 

 

My 5600Mhz 64GB CL40 kit runs at 5200Mhz CL38 in my m18. Of course if we had full control in bios for timings, speed, voltage, I'm sure we can hit higher. 

[Etern4l]

5 hours ago, ssj92 said:

It's quite clear DDR5 SODIMM has huge limitations. I am running 128GB 2400Mhz DDR4 RAM in my 2019 Area-51m running a 9th gen CPU but these newer 12th gens can't run 128GB unless it's capped to almost previous gen speeds? lol

 

The new 48GB modules will be nice because it should allow 96GB DDR5 and hopefully at full speed. 

 

My 5600Mhz 64GB CL40 kit runs at 5200Mhz CL38 in my m18. Of course if we had full control in bios for timings, speed, voltage, I'm sure we can hit higher. 

 

I think the laptop stuff is consistently lagging behind desktops, where we saw similar issues around 12th gen launch, i.e. 3600-4400 CL38 speeds no matter what modules. Now (several BIOS updates later) running at 5200 CL30 on old Hynix dies and a Z690 board. I believe also on the 12th gen but not 100% sure, but 4800 is def solid.

[seanwee]

On 5/3/2023 at 3:40 PM, ssj92 said:

It's quite clear DDR5 SODIMM has huge limitations. I am running 128GB 2400Mhz DDR4 RAM in my 2019 Area-51m running a 9th gen CPU but these newer 12th gens can't run 128GB unless it's capped to almost previous gen speeds? lol

 

The new 48GB modules will be nice because it should allow 96GB DDR5 and hopefully at full speed. 

 

My 5600Mhz 64GB CL40 kit runs at 5200Mhz CL38 in my m18. Of course if we had full control in bios for timings, speed, voltage, I'm sure we can hit higher. 

I have 2x32gb kits of dual rank dual channel Hynix A die kits on the way for my 7945x + 4090 Strix Scar 17.

 

I will be overclocking them with the smokeless UMAF which allows me to change cpu power limits, vrm settings and overlock ram as well. I am targeting 6000 CL30 @ 1.35v. Expecting up to a 25% fps boost in memory intensive games. The stock 4800 CL40 JEDEC spec is atrocious for amd cpus.

 

On 5/1/2023 at 5:51 AM, Mr. Fox said:

It would be logical for a number of reasons. Thermal management is already nearly impossible with pre-DDR5 laptops. The need for increased voltage and access to memory settings, a good CPU sample and superior motherboard design are the reality on desktops. The likelihood of laptop manufacturers getting these extra things right and giving purchasers of their products access to associated BIOS features is like slim to none.

I will try to take some thermal images pre and post overclock. Thermals shouldn't be overly concerning (desktop kits actually run in comparable conditions nowadays, thanks to nvidia's flowthrough design) but I do wonder how much i will be limited by the suboptimal memory layout (optimised for space rather than speed).

 

As for manufacterers giving access to advanced bios, I know that its present with MSI and Razer albeit its hidden. With AMD systems you are able to access them with Smokeless UMAF tool.

[Mr. Fox]

1 hour ago, seanwee said:

Thermals shouldn't be overly concerning (desktop kits actually run in comparable conditions nowadays, thanks to nvidia's flowthrough design) but I do wonder how much i will be limited by the suboptimal memory layout (optimised for space rather than speed).

DDR5 thermals are also a problem on desktops. At high clock speeds, requiring high voltage to function, so was DDR4, but DDR5 took that to a higher level. When temps breach 45-50°C is where a stable memory overclock becomes unstable. Naked DDR4 and DDR5 are pretty easy to cool with a fan blowing on the chips, but the heating blankets they facetiously refer to as "heatsinks" hurt more than they help.

 

I am water cooling my memory now. That works great. If I detach the water block and have no fan blowing on the memory and start a memory stress/stability test everything is as smooth as silk until it reaches that temperature range, then the errors begin to appear. A few a first, but more as the temperature increases.

 

Hynix A-die generally uses less voltage and runs a few degrees cooler than M-die, but A-die can't handle some of the timings being as tight as M-die allows. What makes A-die good is the higher clock speed makes up for the looser timings and higher latency.

[TheQuentincc]

On 5/3/2023 at 9:40 AM, ssj92 said:

It's quite clear DDR5 SODIMM has huge limitations. I am running 128GB 2400Mhz DDR4 RAM in my 2019 Area-51m running a 9th gen CPU but these newer 12th gens can't run 128GB unless it's capped to almost previous gen speeds? lol

 

The new 48GB modules will be nice because it should allow 96GB DDR5 and hopefully at full speed. 

 

My 5600Mhz 64GB CL40 kit runs at 5200Mhz CL38 in my m18. Of course if we had full control in bios for timings, speed, voltage, I'm sure we can hit higher. 

You probably can by editing nvram value 🙂

[ssj92]

16 hours ago, TheQuentincc said:

You probably can by editing nvram value 🙂

I already enabled xmp and other memory features, but problem is, I don't know what to change for the timings because I can't find the values when using ru.efi