Intel was going to make an SSD that could be overclocked - but plans got canned
Click here to post a comment for Intel was going to make an SSD that could be overclocked - but plans got canned on our message forum
cucaulay malkin
just what we need, K-versions of SSDs for a premium,compatible with Z-series motherboards only.
Devid
Such a good idea to overclock something that is for holding your precious data safe...
Horus-Anhur
Here is the full interview, for anyone curious.
[youtube=f6NnBokT4eo]
Agonist
Good thing this stupid crap didn't come out. Because you would have to have a k model just to overclock it with the Z series board just because it's Intel.
pirlampas
I'd think this is a terrible idea. We can more or less go along with CPUs or GPUs misbehaving or even crashing because of unstable overclocks, but an overclock on a data storage could mean corrupted data being stored, and voilà there goes your system. Of course you can also implement some kind of in-between control system to ensure the data being saved is correct, but that can negate any kind of OC gain easily.
FatBoyNL
Announced in 2013: https://www.guru3d.com/news-story/overclock-your-ssd-intel-makes-a-promise.html
At least the endless waiting ended 🙁
geogan
Crazy idea... the thing would be totally corrupted by bad data with a few days of most users amateur overclocking attempts.
I even had to dial back the speed of my main system RAM recently from 3400MHz back to 3200MHz (Ryzen X570) - as I was getting strange problems with downloads failing with CRC type errors - it was only because these downloads did this kind or error checking that the memory corruption was diagnosed. It was only when I ran the Windows Memory Diagnostic tool during boot-up, that I found this:
8332689
1759
307
0
0
134
0
0
0
0
0
173
0
0
0
0
0
0
0
I couldn't find anywhere online explaining exactly what the T3 and T9 tests are, to narrow down the exact problem (if anyone knows I would be grateful!).
My computer has been running for a very long time with these settings, so who knows what kind of corruption and crashing was going on with me knowing.
And I DID always run Windows based memory tests after the overclocking (using that Ryzen memory overclocking/tuning tool), including OCCT which I thought caught all errors, but none of them did find anything.
kakiharaFRS
I agree totally and I can talk as when I was toying with my Z390 9900k overclock after several BSOD my system drive which was on an NVME got corrupted
impossible to boot->unrepairable OS->cloned an image of it with an USB keys->3/4 of the files were corrupt correct sizes correct names and all but they contained garbage
even pictures and text files I hadn't touched for a year everything got corrupted and you couldn't reinstall windows on top of it as it left scrambled files breaking the OS
I had a clone of the drive but 2 weeks old, would have been catastrophic otherwise
schmidtbag
I thought we were all enthusiasts here? Since when do any of you like the idea of being restricted from overclocking? Intel could've done what they've always done - void the warranty if you do overclock. Only an idiot would OC their storage and put something important on it, but not everyone uses storage for critical/irreplaceable data. Should we not have the right to be stupid about our hardware?
If any of you OC your CPU or RAM on a work PC, you are threatening the accuracy of the data, especially if you're not using ECC RAM. The difference with storage is you can do checksums, or use a filesystem that stores multiple versions of the same file.
In any case, people's obsession over storage performance has always been rather strange to me, considering how only a small handful of real-world applications actually really take advantage of something like a x4 lane NVMe drive.
Undying
No wonder Intel sold their ssd business.
Mufflore
Such a foolish idea in the first place.
I do my best to keep my storage cool to prevent problems (NAND chips have air blown over them where possible).
I wont buy an SSD that gets too hot with a heatsink.
David3k
"Overclocking Lead Engineer" is a job that needs to look for work because it's a solution looking for a problem.
Who in the flying hell thinks that department is a good idea and not just something that should be folded into their Testing and Validation teams?
Reardan
How is that appreciably different from overclocking a CPU, processing your precious data, or memory, holding your precious data?
Mufflore
There are many tools to help establish the overclock is safe without later losing/corrupting data on the way to a drive.
And the chance of corrupting a drive is not high, especially if its not plugged in when testing the overclock 😉
386SX
When RAID6 with Hot Spares may be more destructive than RAID0, noted ... 😀
Pryme
Overclock SSD? Nice, let me buy another Noctua NH-D15 and mount on this bad boy!!
Reddoguk
What i don't get is if the ssd can run faster then why not just release them pre overclocked? Seems they wanted to give you that option and then when it goes tits up there's no chance of any warranty returns.
Also i thought SSD's can only reach 550mb/s and no way to increase that number without M.2 drives.
Mufflore
M.2 drives are SSDs also.
M.2 is a form factor not a drive type. SSD means Solid State Drive, any drive that uses ICs instead of spinning platters is an SSD.
Bitey
back in the day when pci bus speeds were a fixed ratios to main bus, overclocking would also OC the pci and ide/sata controllers and other devices.
It would give an nice boost in interface speed, but corruption and killing devices was not that uncommon side effect.
a bad idea in general
geogan
Yes it is. From what I can tell, most of the speed increases and benchmarks people seem to fall for are the queue depth greater than one types like QD=20 and in real world on a home PC with one user, the queue depth is 99% of the time QD=1 and if you look at any benches there is very little increase in speed between various drives at this depth. But reviewers and manufacturers love pushing the QD=32 type results - we are not running a file server in a data center serving 10,000 people simultaneously!
I notice ZERO difference between my old cheap SATA 2.5" SSD (Sandisk Ultra II) which i used to run as Windows C drive, and my current apparently fastest Sabrent Rocket NVMe4 M2 SSD drive I use now for it.
