I wish it had made it to thumb drives for its superior data retention. It doesn't work by holding a charge; it's phase-change, so it retains data longer.
What makes Intel Optane stand out (2023)
131–140 of 164 posts
Re: What makes Intel Optane stand out (2023)
#132I use two Intel 905 SSDs as mirrored cache devices for ZFS.
Why would you mirror a cache device?
Re: What makes Intel Optane stand out (2023)
#133Earlier quoted context omitted.
Optane practical TBW endurance is way higher than that of even TLC flash, never mind QLC or PLC which is the current standard for consumer NAND hardware. It even seems to go way beyond what's stated on the spec sheet. However, while Optane excels for write-heavy workloads (not read-heavy, where NAND actually performs very well) these are also power-hungry which is a limitation for modern AI workflow.
You're conflating two things. Yes, Optane would survive more writes. But it wouldn't survive more TBW/$, because much larger flash drives were available cheaper. Double the size of the drive using identical technology, and you double TBW ratings.
This was very clearly not true at the time for the actual implied TBW figures of even a tiny Optane drive, and is not even true today when you account for the much lower DWPD of TLC/QLC media. Do the math, $1/GB vs $0.1/GB where the actual difference in DWPD per spec sheets is more like two or three orders of magnitude, with the real-world practical one being quite possibly larger. (Have people even seen Optane actually fail in the wild due to media wear out? This happens all the time with NAND.)
Re: What makes Intel Optane stand out (2023)
#134Earlier quoted context omitted.
> were those weird hybrid SSDs even implemented by actual hardware, or were they part of the giant series of massive kludges They were definitely part of the series of massive kludges. But aside from the Intel platforms they were marketed for, I never found a PCIe host that could see both of the NVMe devices on the drive. Some hosts would bring up the x2 link to the Optane half of the drive, some hosts would bring up…
Why on Earth did Intel implement this as a 2x2 device? They could have implemented multiple functions or they could have used a PCIe switch or they could have exposed their device as an NVMe device with multiple namespaces, etc. (I won’t swear that all of these would have worked nicely. But all of them would have performed better than arbitrarily splitting the link in half.) Maybe they didn’t own any of the IP for th…
Or perhaps they just made a number of incredibly poor decisions. They seem to have been doing that for the better part of a couple decades now.
Re: What makes Intel Optane stand out (2023)
#135Earlier quoted context omitted.
> I'm not seeing a lot of regrets from folks who moved to TLC and QLC NAND, and those products are more popular than ever. That's interesting. Even TLC has huge limitations, but QLC is basically useless unless you use it as write-once-read-many memory. I wish I have bought a lot of SSDs when you could still buy MLC ones.
> QLC is basically useless unless you use it as write-once-read-many memory The market thoroughly disagrees with your stupid exaggeration. QLC is a high-volume mainstream product. It's popular in low-end consumer SSDs, where the main problem is not endurance but sustained performance (especially writing to a mostly-full drive). A Windows PC is hardly a WORM workload.
Seems like it is though? Most consumer usage does not have much churn. For things like the browser cache that do churn the total volume isn't that high.
The comparison here is database and caching workloads in the datacenter that experience high churn at an extremely high sustained volume. Many such workloads exist.
Re: What makes Intel Optane stand out (2023)
#136Earlier quoted context omitted.
Optane practical TBW endurance is way higher than that of even TLC flash, never mind QLC or PLC which is the current standard for consumer NAND hardware. It even seems to go way beyond what's stated on the spec sheet. However, while Optane excels for write-heavy workloads (not read-heavy, where NAND actually performs very well) these are also power-hungry which is a limitation for modern AI workflow.
You're conflating two things. Yes, Optane would survive more writes. But it wouldn't survive more TBW/$, because much larger flash drives were available cheaper. Double the size of the drive using identical technology, and you double TBW ratings.
> it wouldn't survive more TBW/$
Yes it would, by an almost arbitrarily large margin. You can test this out for yourself. Overwrite one of each in an endless loop. Whenever the flash based drive fails, replace it and continue. See how long it takes for the optane to fail.
You should be able to kill a typical consumer flash drive in well under a week. Even high end enterprise gear will be dead within a couple of months.
Re: What makes Intel Optane stand out (2023)
#137Did it ever have real input costs which justified its pricing or was it totally a synthetic price to cover an imagined value above base, and some IPR?
I may be wrong, but I tend to think the fab cost has very little to do with the complexity, but not nothing, and the outcome cost is very unrelated to price at a viable yield.
If optane failed, it was probably because Intel didn't like the economics of pricing it to succeed.
Re: What makes Intel Optane stand out (2023)
#138Earlier quoted context omitted.
Same for the Larabee / Knights architecture. Would sure be fun to play around with a 500 core Knights CPU with a couple TB of optane for LLM inference. Intel's got an amazing record of axing projects as soon as they've done the hard work of building an ecosystem.
> 500 core The newest fully E-core based Xeon CPUs have reached that figure by now, at least in dual-socket configs.
But I'm surprised PCIe based CPU+RAM modules aren't a "thing" since that's basically what a GPU is if you ignore all the rather fundamental differences. Seems like it would be convenient to cheaply attach additional compute without worrying about all the other stuff.
I suppose I'm just reinventing SXM at this point. The BC-250 comes close but despite the formfactor it isn't actually a PCIe card. Although if it integrated a 100 Gbit SFP slot it might actually be superior to a solution that resided in a host system. But the BC-250 is very much an anomaly as opposed to the norm.
Re: What makes Intel Optane stand out (2023)
#139Earlier quoted context omitted.
> 500 core The newest fully E-core based Xeon CPUs have reached that figure by now, at least in dual-socket configs.
But I'm surprised PCIe based CPU+RAM modules aren't a "thing" since that's basically what a GPU is if you ignore all the rather fundamental differences. Seems like it would be convenient to cheaply attach additional compute without worrying about all the other stuff. I suppose I'm just reinventing SXM at this point. The BC-250 comes close but despite the formfactor it isn't actually a PCIe card. Although if it integr…
You need CXL to extend the cache coherency properties of actual RAM over a remote link. That's costly tech. Otherwise, you're relying on the OS (and even the compiler/basic libraries, since you need to make fences, etc. OS-visible) to paper over the differences by doing its own implementation of distributed shared memory (this is known as a 'SSI' or single-system-image approach) which has significant challenges and is closer to the spirit of setting up swap.
Re: What makes Intel Optane stand out (2023)
#140Earlier quoted context omitted.
But I'm surprised PCIe based CPU+RAM modules aren't a "thing" since that's basically what a GPU is if you ignore all the rather fundamental differences. Seems like it would be convenient to cheaply attach additional compute without worrying about all the other stuff. I suppose I'm just reinventing SXM at this point. The BC-250 comes close but despite the formfactor it isn't actually a PCIe card. Although if it integr…
You need CXL to extend the cache coherency properties of actual RAM over a remote link. That's costly tech. Otherwise, you're relying on the OS (and even the compiler/basic libraries, since you need to make fences, etc. OS-visible) to paper over the differences by doing its own implementation of distributed shared memory (this is known as a 'SSI' or single-system-image approach) which has significant challenges and i…
I didn't mean anything like that. Just the equivalent of a GPU with the ability to run arbitrary CPU oriented programs.
Of course GPUs do many tasks very well but there are also plenty of problems that aren't well suited to them. Well I suppose I've answered my own question at this point. There probably just aren't enough real world problems that aren't amenable to running on a GPU while also being either compute or memory bandwidth bound.
Still the near-monoculture does strike me as odd. I guess GPUs have bifurcated into enterprise versus consumer at this point but otherwise all we've got is a single CPU example from over a decade ago and a single alternative take on the concept from Fujitsu. Is it just due to the obscene cost of masks for modern process nodes?