Persistent memory to replace DRAM, but it could take a decade
- Reference: 1708457413
- News link: https://www.theregister.co.uk/2024/02/20/persistent_memory_to_replace_dram/
- Source link:
In a recent [1]webinar , the Compute, Memory, and Storage Initiative (CMSI) of the Storage Networking Industry Association (SNIA) examined emerging memory technologies such as MRAM, FERAM, and ReRAM, with answers to questions supplied in a [2]blog afterward.
These memories, all persistent, are already replacing some established memory technologies, including SRAM, NOR flash, and DRAM, at least in some niche applications. But what about mainstream memory used in PCs and servers?
[3]
To cut to the chase, the experts believe that DRAM will eventually give way to a new memory type, but it is too early to tell which one will triumph, and when this will happen. Their best guess is the early 2030s.
[4]
[5]
One benefit of replacing DRAM with persistent memory is obvious; it keeps its content even without power, meaning there is less danger of losing data. However, DRAM has been around for a long time and is cheap to produce and available in high densities – hurdles that persistent memories will also need to surmount.
The webinar pointed out that Optane, which [6]Intel dropped in 2022 , proved how difficult it is for rival memory technologies to compete with DRAM (and NAND flash) on price.
[7]
One route being taken by many of the makers of these emerging memories is the embedded market. This is seen as "a good path for emerging memory IP," according to the experts, because it allows them to get used in production environments and build up some volume.
"And as the volume and comfort with manufacturing those memories increase, it opens up the possibility down the road of lower costs with higher volume standalone memory as well," the experts wrote.
Those experts on the webinar were Tom Coughlin of Coughlin Associates; Jim Handy, semiconductor analyst at Objective Analysis; and Arthur Sainio, co-chair of the SNIA Persistent Memory Special Interest Group and product marketing director for SMART Modular Technologies.
[8]
Another question that arose from the webinar was whether persistent memories will be able to match the speed of DRAM, to which the answer is yes.
"It appears that that has already happened with the hafnium ferroelectrics that SK hynix and Micron have shown," the experts responded. "Ferroelectric memory is a very fast technology and with very fast write cycles there should be every reason for it to go that way."
MRAM (magnetoresistive RAM) might get up to DRAM speeds with both its writes and reads, and reads will also probably rival DRAM speeds with ReRAM (resistive RAM).
"I think that probably most applications are read intensive and so the read (performance) is the real place where the focus is, but it does look like we are going to get there," one of the experts stated.
[9]China 'readies production' of homegrown high-bandwidth memory
[10]Now is a good time to buy memory because prices rise next year, Gartner predicts
[11]Why the end of Optane is bad news for all IT
[12]Uncle Sam needs novel memory for nuke sims. So why did it choose Intel?
Another issue could be endurance, or the wear issue familiar from NAND flash. Techniques for wear leveling have been developed and incorporated into the software that manages SSDs to handle this, such that the experts contend that "any emerging memory can get by with an equally low endurance as long as it's put behind the right controller."
But a DRAM replacement will need to have higher endurance. Even with the adoption of practices such as mixing types of memory, such that code goes into memory with low endurance and data into higher endurance memory, the lowest acceptable level would be in the tens of thousands of cycles, the experts conclude.
In terms of host interfaces, most emerging technologies have used a NOR flash or SRAM type interface because these were the memory types used in the embedded applications they are targeting.
There is no reason why they could not adopt the same interface as DRAM, although "there are some special twists because you don't have to refresh emerging memory technologies."
But the SNIA experts see CXL, or [13]compute express link , as the future interface of choice going forward as it can present any type of memory to the host processor.
"One of the beauties of CXL is that you put anything you want to with any kind of interface on the other side of CXL and CXL erases what the differences are. It moderates them so although they may have different performances, it's hidden behind the CXL network," one expert explained.
So it looks like the days of DRAM could be numbered, but even the experts don't know what exactly will replace it, or precisely when.
"DRAM has a transition that we don't know when it's going to happen that it goes to another emerging memory technology," the experts wrote. "We've put it into the early 2030s in our chart, but it could be much later than that." ®
Get our [14]Tech Resources
[1] https://www.snia.org/educational-library/emerging-memories-branch-out-2024
[2] https://sniacmsiblog.org/2024/02/emerging-memories-branch-out-a-qa/
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/storage&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZdUvGE-sXZ8HhC9tuKC5MwAAAYQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/storage&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdUvGE-sXZ8HhC9tuKC5MwAAAYQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/storage&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdUvGE-sXZ8HhC9tuKC5MwAAAYQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2022/07/29/intel_optane_memory_dead/
[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/storage&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdUvGE-sXZ8HhC9tuKC5MwAAAYQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/storage&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdUvGE-sXZ8HhC9tuKC5MwAAAYQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[9] https://www.theregister.com/2024/02/01/china_high_bandwidth_memory/
[10] https://www.theregister.com/2023/12/04/memory_prices_gartner/
[11] https://www.theregister.com/2022/08/01/optane_intel_cancellation/
[12] https://www.theregister.com/2022/12/13/intel_doe_nukes_mem/
[13] https://www.theregister.com/2022/08/04/cxl_time_now/
[14] https://whitepapers.theregister.com/
Its gonna be hard to supplant DRAM
If it costs several times more per bit it'll never have a chance, because persistent memory is of little use to most people. DRAM manufacturing has been refined for decades, even if "in theory" some new memory type could compete price-wise it won't have a chance because you can't just spend tens of billions building fabs to make it on the assumption that after 50 years the tech world will be easily convinced to abandon DRAM.
Re: Its gonna be hard to supplant DRAM
If you can execute straight off the "hard drive", surely that could work out being faster than having to read it into RAM first?
Of course that would mean you have to think about programming it in a completely different way, and you would still have to "open" the file if it is on the other end of a network link, because the limitations of the speed of light mean that this is always going to be slower than some sticks of memory sitting about 5cm away from the CPU.
Re: Its gonna be hard to supplant DRAM
The idea of combining storage and execution has been around for a while. Applications would be GPU and AI operations, but not so much general computing.
Re: Its gonna be hard to supplant DRAM
It would be faster, but the same could be accomplished by building a faster SSD with the same stuff. Basically, it's Optane again. You could save a bit of time by changing the executable format you're using from one that loads itself into memory and does a bunch of initialization to one that has already initialized and just stores and loads that state, but the programs will still have to load any configurable state from somewhere persistent which will likely still involve loading files from some kind of storage and processing their contents. I'm not sure loading speed would be different enough for people to care if the price is much higher.
Re: Its gonna be hard to supplant DRAM
"because persistent memory is of little use to most people"
Intel tried to sell the Xpoint memory for awhile in many configurations. The Optane NVMe drives had lowest random access latencies available but most people were already pretty happy with the move from HDD to SSD (and NVMe) that the difference wasn't worth the price difference.
The Optane NVDIMM modules are/were in a [1]different class - the latency being something like 1 / 100th of NVMe drives. The sequential read speeds were only regular NVMe class (even if the memory bus allows way more) but write speeds were much lower.
Too bad those NVDIMM's worked only on Intel Xeon platforms, were expensive (duh!) and initially they were available in small 16/32GB modules, and you could populate maybe three of them per CPU (or per system? getting hazy...) so not that big. I think the later NVDIMM generatiosn allowed much bigger systems.
I'm sure there are applications that will greatly benefit from those super low latencies, but as you say, little use to most people.
[1] https://blocksandfiles.com/2019/07/02/optane-dimm-access-modes/
Yeah, yeah, we've heard this one before.
It was a decade away a decade ago, and a decade from now it probably still will be.
If you dig deep enough, stories like this are usually just bait for research grants.
And security?
Isn’t persistent RAM a security threat?
AFAIK there is at least one attack which takes advantage of the little persistence that normal RAM has, so this would make the attack a lot easier (“shall we have a cup of tea first?”).
Unless they keep some volatile RAM and somehow manage to keep all the juicy stuff there.
One benefit of replacing DRAM with persistent memory is obvious; it keeps its content even without power,
That's going to upset Microsoft support. What will they do when switching it off and on again doesn't work anymore?
Isn't MRAM wear resistant?
The article mentions wear problem, which is, as far as I know, non existent for MRAM. Proven for more than five decades by its predecessors known as "core memory".
And as far as Optane goes: That is Intel management fuck up. Why limit the RAM-Module version only to the most expensive CPUs for no good reason... If they'd opened it up for all Xeons it would have been used a lot, even for "small" only 16 GB variants. Countless usages come to my mind. Those m2 cards were a bit late too.
Just bring back core memory