|
that reminds me, I still need to rehome my Windows boot to the 960GB Optane I bought a few months ago…
|
#
?
Mar 17, 2024 00:30
|
|
|
|
| # ? Apr 27, 2024 18:43 |
|
|
Optane sounds fun, but what's the appeal of wedding your system to dead-end tech?
|
|
#
?
Mar 17, 2024 00:38
|
|
|
How is it wedded? It’s no harder to move from Optane to a new SSD than it is from any conventional SSD. I don’t expect to need ongoing support for any of my drives, really, and when they die/scream-in-SMART they just get replaced.
|
|
#
?
Mar 17, 2024 00:44
|
|
|
If your SSD is in need of a FW patch then things are wrong to start
|
|
#
?
Mar 17, 2024 00:48
|
|
|
Shumagorath posted:Optane sounds fun, but what's the appeal of wedding your system to dead-end tech? It’s a dead end but its write endurance is basically infinite compared to normal ssds.
|
|
#
?
Mar 17, 2024 02:52
|
|
|
We used a ton of those 16GB optane drives as pcie endpoints but I forget what it was that they did wrong, maybe it was having problems with repeated link disables. Anyway shouldn’t be a problem for a legit use when not torture testing pcie links. I still think a disk-on-module is probably a better option for a small industrial drive though.
|
|
#
?
Mar 17, 2024 02:57
|
|
|
Perplx posted:It’s a dead end but its write endurance is basically infinite compared to normal ssds. Why did 3DXpoint lose out to NAND? Cost?
|
|
#
?
Mar 17, 2024 03:04
|
|
|
Shumagorath posted:I guess that’s nice? I’ve never even come close to burning out an SSD, though maybe QLC will start to make me sweat that. Only intel and micron were interested in producing it, and frankly NAND lifespan is still plenty good especially with a lot of redundancy for enterprise products. 3dxpoint’s big feature was supposed to be much much lower latency and transfer speeds so it could be its own memory tier in between dram and NVMe ssds but it never got to that performance level so it was more of just an also ran to PCIe NVMe.
|
|
#
?
Mar 17, 2024 03:07
|
|
|
WhyteRyce posted:If your SSD is in need of a FW patch then things are wrong to start yo dawg I heard you like computers, we put a computer inside your computer so you can have firmware bugs while you have software bugs
|
|
#
?
Mar 17, 2024 03:13
|
|
|
Klyith posted:yo dawg I heard you like computers, we put a computer inside your computer so you can have firmware bugs while you have software bugs
|
|
#
?
Mar 17, 2024 03:50
|
|
|
Having done both I’m feeling very triggered right now
|
|
#
?
Mar 17, 2024 04:28
|
|
|
A certain SSD vendor will never admit you found a bug and will instead insist it’s the specs fault for not allowing fault tolerance and special messages to mitigate said bug. It’s exhausting
|
|
#
?
Mar 17, 2024 04:32
|
|
|
Klyith posted:Is there any tweaking or tuning for how much separation between the voltage levels is used? Or is that a thing that's not really possible to do? From the very basics I know of how the cells work, that seems like it would trade better retention (more gap between levels) for endurance (higher voltage putting more stress on the cells). Please note I am not a real flash expert, I just used to chat with an engineer friend who worked on enterprise SSD controllers so I learned some flash facts from him. No idea re: levels but if I had to guess, I'd guess it isn't possible. Just on first principles, the more electrons you shove in the trap, the more they want to get out. Doesn't seem like that would mix well with leaky insulation at end of life. On pSLC - multi-level write modes have worse write performance than SLC because the controller inside the flash chip has to titrate charge into the cell. They inject some, read to see where the cell's at, if it's too low loop until done. It does seem plausible that repeated cycles of applying high voltage to force a little more charge in causes more wear than a single big injection.
|
|
#
?
Mar 17, 2024 05:17
|
|
|
How are Inland brand drives? MicroCenter is usually pretty solid on other stuff
|
|
#
?
Mar 17, 2024 16:20
|
|
|
Nissin Cup Nudist posted:How are Inland brand drives? MicroCenter is usually pretty solid on other stuff At least before they were solid, just standard drives with Phison controller.
|
|
#
?
Mar 17, 2024 16:30
|
|
|
Nissin Cup Nudist posted:How are Inland brand drives? MicroCenter is usually pretty solid on other stuff Middle of the pack. Easier to return and get a replacement if you are close to MC and get a bum one.
|
|
#
?
Mar 17, 2024 16:31
|
|
|
there are a lot of reasons why Optane died. the tl;dr though is basically cost/capcity. you were basically paying more for a drive that was 1/10th the capacity. optane gets a lot of poo poo for failing but I think that's a tad unfair. it's the same poo poo computational storage, K-V SSDs, storage accelerators, ZNS, and probably FDP will struggle with. yes, tremendous benefit can be squeezed out under optimal, well written code. but at the end of the day the customers all just seem to want something that's cheap as poo poo, works like everything else, and doesn't require a rewrite of whatever poo poo software stack they are having to deploy. having to explain "literally none of your old code will run on this ZNS drive unless maybe you switch over to this other experimental file system also your performance may be poo poo" to a customer probably instantly sours them. WhyteRyce fucked around with this message at 17:25 on Mar 17, 2024 |
|
#
?
Mar 17, 2024 17:22
|
|
|
makere posted:At least before they were solid, just standard drives with Phison controller. Before what?
|
|
#
?
Mar 17, 2024 17:23
|
|
|
Subjunctive posted:Before what? Couple years ago, I haven't been keeping up-to-date.
|
|
#
?
Mar 17, 2024 17:43
|
|
|
makere posted:Couple years ago, I haven't been keeping up-to-date. ah, OK. I thought there might have been An Event like a PE takeover or change of suppliers or whatever, whew Microcenter is one of the last good things in computer retail, and I cling to it even though I have never been to one
|
|
#
?
Mar 17, 2024 17:46
|
|
|
Subjunctive posted:Before what? The thing. You know, that Thing.
|
|
#
?
Mar 17, 2024 17:57
|
|
|
I was reading a review of a SK drive and the gist was that it was perfect in every way except for not having a heatsink. What does an SSD heatsink even look like since I wasn't aware SSDs even needed a heatsink? Can you just place a drive in-line with a case fan or such?
|
|
#
?
Mar 17, 2024 20:25
|
|
|
Nissin Cup Nudist posted:I was reading a review of a SK drive and the gist was that it was perfect in every way except for not having a heatsink. What does an SSD heatsink even look like since I wasn't aware SSDs even needed a heatsink? Can you just place a drive in-line with a case fan or such? Just a block of metal, with some minor amount of fins. Ex the WD Black 850X only has "fins" from some internal holes in the metal bar. They're more heat spreaders than sinks. However, you can generally ignore a review talking about heatsinks on SSDs being necessary. Number one, most current mobos come with heat spreader plates where the NVMe drives get installed. And laptops can't fit drives with heatsinks at all. Two, it's just not a huge deal even if you don't have one. The controller chip on modern SSDs can get hot enough to throttle itself down, just like a normal CPU. (In fact these controller chips are specialized ARM CPUs, thus my joke earlier about putting a computer in your computer.) When that happens they naturally will have lower performance. But the thing is, pretty much the only thing that makes them throttle hard are artificial benchmarks testing maximum sustained sequential performance. Real world normal use will rarely do that. If your case isn't a hotbox, a SSD heatsink is not a requirement. (There are places they are needed, like the PS5 storage expansion slot which is a tiny box with zero airflow.)
|
|
#
?
Mar 17, 2024 20:46
|
|
|
priznat posted:Only intel and micron were interested in producing it, and frankly NAND lifespan is still plenty good especially with a lot of redundancy for enterprise products. I think performance was better for the DRAM modules, but RAM prices fell just as it entered the market, so it was squeezed out off the memory mode. And the persistent memory mode needed stack adaptation, so it was too slow to get going.
|
|
#
?
Mar 18, 2024 08:41
|
|
|
Slotting 3DXpoint PMEM meant losing a slot for RAM, and that's extremely silly. Ram is ten times faster, and rewarming a cache after boot is not nearly as hard as having a bunch of developers rework your app to capitalize on not-as-fast-as-ram, not-appreciably-faster-than-ssds-in-a-way-that-mattets storage. Optane--as in the storage side of it--was cool but prices never worked out for it.
|
|
#
?
Mar 18, 2024 15:29
|
|
|
Optane DIMMs were regular dram backed by Optane so it was like the memory hierarchy within a memory hierarchy if you used the right mode It’s not just persistent, it’s huge amounts of memory. I’ve seen enough demos and presentations to know that there is a potentially valuable performance improvement to be had that isn’t just no need to reload from disk after reboot, but yeah rewriting anything is a hard sell even if it’s for potentially good reasons. ZNS seems pretty dead in favor of FDP because of the hard requirement of a complete rewrite even though the former gives you a real cost advantage. And I think many companies weren’t keen on jumping all in to a single sourced tech from a historically finicky vendor There still is interest around the concept. It’s why CXL is forging ahead. WhyteRyce fucked around with this message at 15:47 on Mar 18, 2024 |
|
#
?
Mar 18, 2024 15:32
|
|
|
WhyteRyce posted:ZNS seems pretty dead in favor of FDP because of the hard requirement of a complete rewrite even though the former gives you a real cost advantage.
|
|
#
?
Mar 18, 2024 16:47
|
|
|
I don’t even know FDP is that much easier (the spec is kind of junky tbh) but it lets you use it as a traditional SSD in parallel to or as backup to your application teams figuring out just what the hell they are supposed to be doing so maybe that’s enough
|
|
#
?
Mar 18, 2024 16:56
|
|
|
It even kind of seems like CXL is a bit DOA with no real great use cases coming up, additional memory tiers that require a poo poo ton of software support and system planning probably not where it’s at.
|
|
#
?
Mar 18, 2024 18:07
|
|
|
I used to think nothing of stuff like dual sockets being an issue for anyone because I’m a brain dead HW guy but then I talked to application teams that absolutely hate numa and lol CXL is numa on steroids get in losers we are going over an interconnect
|
|
#
?
Mar 18, 2024 18:20
|
|
|
WhyteRyce posted:I used to think nothing of stuff like dual sockets being an issue for anyone because I’m a brain dead HW guy but then I talked to application teams that absolutely hate numa and lol CXL is numa on steroids get in losers we are going over an interconnect Yeah Software folks are gonna dictate this for a while at least and I think they hate it. Even adding memory expansion is a big headache and that is the simplest of CXL applications. Doing cache coherency and whatnot, they’re gonna go bananas.
|
|
#
?
Mar 18, 2024 18:27
|
|
|
Zorak of Michigan posted:The form factor is called disk on module. BobHoward posted:Retention starts out high when the flash is at 0 write cycles, because that's when the insulation around charge traps or floating gates is in excellent condition. As the flash accumulates write cycles, insulation gets damaged, making the cells leakier, which means they fade faster. Saukkis posted:I suggest Intel Optane 16GB M.2 NVMe sticks, they are available plentiful on eBay for $5. Thanks, this is exactly the info I'm looking for. Those optanes might be perfect but either way, I'm happy with the retention of a low-write SSD. Most importantly I don't need to spec the rest of the hardware around the storage since there's options for anything I choose.
|
|
#
?
Mar 19, 2024 19:45
|
|
|
Are the various secure erase ops any worse on SSD endurance than a normal write cycle? I understand there’s deleting an AES key and forcing all cells to 1 or max_voltage…?
|
|
#
?
Mar 19, 2024 19:50
|
|
|
Shumagorath posted:Are the various secure erase ops any worse on SSD endurance than a normal write cycle? I understand there’s deleting an AES key and forcing all cells to 1 or max_voltage…? AFAIK most SSDs just do the bit where they delete the encryption key and the translation table. Effectively zero endurance use. For the ones that also do an extra erase of the flash, it's an erase that hits everything at once. I'd assume this counts as half of a P/E cycle, and so is basically 50% of a full drive worth of endurance? But I dunno if erase and program have the same or different wear. When talking about NVMe drives, you can send distinct commands for Sanitize Crypto Erase or Sanitize Block Erase, if you use a tool like nvme-cli with the discernment.
|
|
#
?
Mar 19, 2024 23:57
|
|
|
Shumagorath posted:Are the various secure erase ops any worse on SSD endurance than a normal write cycle? I understand there’s deleting an AES key and forcing all cells to 1 or max_voltage…? When I looked at it on SATA drives, the ATA Secure Erase command permitted both of these implementation choices. Some manufacturers publish technical documentation which discloses how they do it. Others, you can infer based on observing how the drive behaves during a secure erase (mostly, how long does it take - 10 or more seconds hints at bulk erase, a few seconds suggests key destruction), or other factors like SMART stat counters before and after the wipe. Don't count on one vendor doing the same thing every time. IIRC, I found that Samsung 840 appeared to be a bulk erase family, while 850 used key destruction.
|
|
#
?
Mar 20, 2024 00:06
|
|
|
Klyith posted:When talking about NVMe drives, you can send distinct commands for Sanitize Crypto Erase or Sanitize Block Erase, if you use a tool like nvme-cli with the discernment. BobHoward posted:When I looked at it on SATA drives, the ATA Secure Erase command permitted both of these implementation choices. Some manufacturers publish technical documentation which discloses how they do it. Others, you can infer based on observing how the drive behaves during a secure erase (mostly, how long does it take - 10 or more seconds hints at bulk erase, a few seconds suggests key destruction), or other factors like SMART stat counters before and after the wipe.
|
|
#
?
Mar 20, 2024 00:31
|
|
|
nvme drives are not required to support the sanitize operation, or any the specific sanitize operations. Always need to check the identify controller output for support which is probably not easily doable in windows really surprised someone would not crypto erase though you should still be able to kick off secure erase and specify crypto vs. user data erase via nvme format though WhyteRyce fucked around with this message at 00:47 on Mar 20, 2024 |
|
#
?
Mar 20, 2024 00:39
|
|
|
I'd be interested to know what's in this thing for secure erase: https://www.startech.com/en-us/hdd/sm2dupe11
|
|
#
?
Mar 20, 2024 01:16
|
|
|
Shumagorath posted:I noticed this, but I've also found drives that would only support SBE and fail SCE. WhyteRyce posted:really surprised someone would not crypto erase though The simplest reason would be because they're not an Opal 2.0 self-encrypting drive. For example, my SN770: code:Even on a QLC drive I can't see it being a problem unless you had reason to repeatedly do secure erases. That would be a pretty weird scenario.
|
|
#
?
Mar 20, 2024 01:24
|
|
|
|
| # ? Apr 27, 2024 18:43 |
|
|
Makes sense since iirc Opal and IEEE 1667 are dead after so many manufacturers hosed up a good thing and everyone went back to implementing poo poo in software. Speaking of which... is write amplification a thing / of significance with software full-disk crypto? I remember back in the Sandforce days it would ruin their deduplication / compression special sauce.
|
|
#
?
Mar 20, 2024 01:31
|
|