|
dud root posted:Do Samsung SSDs juggle data around when idle for wear leveling purposes? Curious if the cells where my large static data is stored have been written to once, or if the controller is smart enough to eventually move 50Gb.italy.holiday.#blessed.mkv to the most used cells Yes we all assume they do wear leveling. I don't think anyone has access to a controller to know for realsies.
|
# ¿ Sep 2, 2016 02:12 |
|
|
# ¿ Apr 28, 2024 17:53 |
|
afkmacro posted:950 pro (512) sustained isn't the only metric that matters, in fact 4k random better represents workstation workloads
|
# ¿ Sep 21, 2016 14:13 |
|
Now we know why the Intel 750s have that gigantic heatsink
|
# ¿ Sep 22, 2016 16:02 |
|
Potato Salad posted:What are y'all using to perform maintenance / imaging on m.2 NVMe drives? You'd have to have some kind of USB 3.1 device to actually use a four-lane NVMe device at full speed, but if all you're doing is maintenance / troubleshooting, 3.0 ought to be fine... Macrium Reflect
|
# ¿ Sep 23, 2016 02:04 |
|
Concurred posted:How soon are we talking? They were just announced late September and I'm finalizing my build next week. I was gonna go with an Intel 750 or 600. Those are not really similar drives other than both being NVMe. The 750 should destroy the 600 of course.
|
# ¿ Oct 11, 2016 00:59 |
|
quote:The Intel 600p only switches into a read-only mode when the spare area is exhausted. Intel also noted users can copy the data from a read-only SSD by installing it as a secondary drive in another computer. Intel provided an official response outlining the recovery procedure, and we include a more detailed explanation in the link. Assuming this is actually what happens, bravo Intel. You simply cannot use a SSD with no spare NAND. I'd happily buy one.
|
# ¿ Oct 12, 2016 19:13 |
|
"The Optane SSD was able to achieve 7.2x times more IOPS at low queue depth and upto 5.21 times the IOPs of conventional SSDs at high queue depths. Optane SSDs also provide 8.11x times lower latency than conventional NAND solutions. An Optane Technology based SSD has 10x times the density of conventional SSD drives. The marketing material also claims it is 1000x faster than the competition available on the market but it isn’t clear to what exactly they are referring to. Optane SSDs will have 1000x the endurance – which, if true, should mean the device has virtually unlimited life span for practical purposes."
|
# ¿ Oct 15, 2016 01:01 |
|
ConanTheLibrarian posted:Ars has some interesting graphs suggesting the 960 Pro's higher specs don't translate to a real world performance boost for PC users: http://arstechnica.com/gadgets/2016/10/samsung-960-pro-review-the-fastest-consumer-ssd-you-can-buy/3/ (last set of charts) I think we are dealing with Windows storage subsystem inefficiencies at this point. Faster SSDs don't do much for workstation workloads.
|
# ¿ Oct 18, 2016 21:10 |
|
goodness posted:Just traded weed for an Intel s3510 1.2tb ssd . Oh sure, its a insane good drive and will probably last as long as you use it.
|
# ¿ Oct 20, 2016 03:53 |
|
The NVMe Drives are so fast you could probably burn one out in a month doing constant writes.
|
# ¿ Oct 23, 2016 14:46 |
|
I swear it doesn't feel like I am waiting for the SSD anymore. It feels more like Windows 10 has inefficiencies that cause delays rather than waiting for a HD to load. Also the task manager says my Intel 750 tends to be around %50 or less usage regardless of what I am doing.
|
# ¿ Nov 15, 2016 16:48 |
|
Yeah.. lame! WTF Intel?
|
# ¿ Nov 23, 2016 23:22 |
|
blowfish posted:Uhhhhhh Well the 750 line is pretty badass. Not cost competitive though.
|
# ¿ Nov 24, 2016 18:23 |
|
Malcolm XML posted:750 line is a data center drive in consumer marketing I know, I own one. Its a rocket of a hard drive. I got the PCIe form factor because the 2.5" one is harder to deal with.
|
# ¿ Nov 25, 2016 19:18 |
|
quote:Thanks for ditching the QA teams, Nadella.
|
# ¿ Dec 7, 2016 16:25 |
|
Combat Pretzel posted:Eh, there's still some dedicated QA over there, but it got apparently drastically reduced and some of the remaining testing responsibilities moved over to the dev teams IIRC. I'm not sure what to think about all of it. In the past, there were huge QA teams and dedicated beta tests. Latter don't necessarily imply being better than the insider preview stuff in terms of reach, but at least there were proper communication channels. Now there's just this lovely nearly useless Feedback app. I've been sending constant feedback with that app. A bunch of my feedbacks have definitely been fixed. Probably a coincidence.
|
# ¿ Dec 7, 2016 17:00 |
|
Totally agree about no return path for communication, nor a bug tracking thing of any kind. Pretty pathetic overall. I feel like I should be getting really familiar with Linux because Windows might be jumping the shark.
|
# ¿ Dec 7, 2016 17:31 |
|
Nice Van My Man posted:So I was starting up my computer, and my 850 EVO was reading SMART status BAD. I tried different cables etc. etc. and eventually just shipped it back to Samsung and they sent me a refurb. Fortunately I had backups, so nothing lost. When it showed BAD status did you try and get any data off the thing? I would have been really interested to know. Based on that answer you can poo poo your pants or not.
|
# ¿ Dec 15, 2016 23:59 |
|
Nice Van My Man posted:Yep, hooked it up to another computer. It connected as "FAT32" with a storage space of 0GB. Fortunately I have a backup, so nothing lost. I was more just scared that something about my computer or usage might fry SSDs, but it seems like I was just being paranoid. Commence pants making GBS threads. Not good at all. (not really) but goddamn SSD controllers!
|
# ¿ Dec 17, 2016 02:53 |
|
Loucks posted:I want an mSATA SSD for my Thinkpad X230. Is the 240GB Intel 525 worth the $50 over the 250GB Sarnsung 850 Evo? My primary concern is reliability. Not worth it at all. Get the 850. Or a Sandisk x400 for a little savings.
|
# ¿ Dec 17, 2016 17:45 |
|
BIG HEADLINE posted:For those of you who have a 960 EVO/Pro or SM951, Samsung put out a new driver today that should net you some extra performance: http://www.guru3d.com/files-details/samsung-nvme-ssd-driver-download.html Also recent firmwares, drivers, and toolbox for Intel 750: https://downloadcenter.intel.com/product/86741/Intel-SSD-750-Series-1-2TB-2-5in-PCIe-3-0-20nm-MLC-
|
# ¿ Dec 23, 2016 16:34 |
|
nielsm posted:Windows still does not handle completely full filesystems well, you should definitely keep 5-10% free on an NTFS volume regardless of overprovisioning on the partitioning or firmware level. BS. What is this based on.
|
# ¿ Dec 28, 2016 00:49 |
|
Is it really the case that the Intel 600p NVMe drive is actually much slower than the 850 EVO SATA? Jebers.
|
# ¿ Jan 15, 2017 18:09 |
|
Eletriarnation posted:No? Gotcha. Thanks, 600p inbound.
|
# ¿ Jan 17, 2017 23:40 |
|
Anime Schoolgirl posted:the 1500mb/sec burst reads are mighty nice anyhow It's a good price for NVMe for sure. Hopefully it wont bite me in the rear end over some odd write-blocking problem.
|
# ¿ Jan 18, 2017 02:04 |
|
Bob Morales posted:I'd happily take an ADATA with a SF2200 You would take a 6 year old SSD controller? k
|
# ¿ Jan 25, 2017 01:47 |
|
Agrajag posted:totally getting myself a 1tb nvme if they start getting cheaper. i am in love with my intel 600p. Yeah the one I got is a goddamn rocket. I am pretty sure it is a bit faster than the 850 evo.
|
# ¿ Jan 25, 2017 03:21 |
|
Dogen posted:Does SMART data mean anything for SSDs? Magician just had an update and says "good" for drive status, which I looked at the data and it has fails on uncorrectable error and ECC error rate, wondering if I need to start worrying about replacing this stupid 850 pro. Post the SMART datas
|
# ¿ Jan 25, 2017 20:18 |
|
Dogen posted:
From googling a bit, you might actually want to RMA that drive. You could also keep backups and keep running the drive. Some say it could be a cable/data issue, one time kind of thing. Others say it could be the beginning of the drive outright failing.
|
# ¿ Jan 25, 2017 21:18 |
|
Ika posted:Does the 960 EVO also have throttling issues? I just installed one in my work PC, and tried backing up a 40gb database file I had copied to it earlier, and by the time the file had finished copying the average speed had dropped to 400MB / s. This is using win7 with the samsung NVM driver but without magician. That is normal. 40GB is enough writes to make the drive perform at its lowest speed since the cache is full.
|
# ¿ Jan 26, 2017 12:51 |
|
Yeah the x400 is a great line of drives. The best cheap option.
|
# ¿ Jan 27, 2017 19:41 |
|
quote:The DC P4800X purportedly offers up to 30 DWPD (Drive Writes Per Day) of endurance, which measures how many times you can fill the drive per day over the warranty period. 30 DWPD is unheard of for NAND-based SSDs; the most endurant modern SSDs top out at 10 DWPD. The DWPD metric can be muddy due to differing capacities, but overall, the DC P4800X can absorb up to 12.3PB of data during its service life. Intel's 450GB DC P3520 SSD, which is NAND-based, can withstand only 590TB, so apparently, 3D XPoint offers almost 21x more endurance than NAND. Hot sex. Still, not 1000x greater endurance than NAND like the hype said way back when.
|
# ¿ Feb 10, 2017 04:11 |
|
WhyteRyce posted:Might have been oversold but this is also the first iteration of this brand new memory type Yeah the devil is in the details and this stuff is rad. I want it.
|
# ¿ Feb 11, 2017 02:53 |
|
Those tiny USB drives just don't match up to even a mechanical HD. Probably your best hope is something like Samsungs USB 3.0 SSD is the smallest while also being fast as hell. https://www.amazon.com/Samsung-T3-P...samsung+usb+ssd
|
# ¿ Feb 16, 2017 21:52 |
|
Maneki Neko posted:So maybe dumb question, I replaced my 5 year old Crucial M4 with a bigger SSD. I know SSD technology has continue to mature over the years, anyone have any sense if there's much life left on that M4? I know I've seen 10+ year endurance lifetimes kicked around for newer gear, any more definitive answers on the older stuff? Run something like Crystal Disk Info and look at the Wear Level Indicator metric. It counts down from 100 to 0 and supposedly around 10 is where the drive should start telling Windows it is bad. http://crystalmark.info/software/CrystalDiskInfo/index-e.html
|
# ¿ Feb 19, 2017 18:56 |
|
lordfrikk posted:Is steady-state something that every drive will reach eventually (I guess like some kind of burn in) or does it require long enough uninterrupted usage after which it will return to "normal" (non-steady-state?)? I'm trying to understand if tests using speeds reached in steady-state are of any consequence for regular consumer workload. Steady state occurs when a drive has burned through its buffers and has to directly write and read to the NAND. Consumer models can sustain high levels for shorter amounts of time and are tuned for bursts of i/o whereas enterprise drives are tuned to provide better performance while in steady-state. Benchmarks tend to hit worst-case scenereos so this means more i/o than the buffers can take and then you get to see steady-state performance. On a consumer machine you will probably never reach a steady-state.
|
# ¿ Feb 23, 2017 17:19 |
|
da anime bulldog posted:So, im building a computer for gaming, and I'm getting back into the swing of learning what all this poo poo means as I select parts. I'm going to go with a Samsung SSD. NVMe is a faster protocol and was created to work with NAND flash storage. It is far more efficient at using solid state memory than the SATA protocol and does really good when hit with a ton of concurrent requests. So, you definitely want NVMe if your stuff has support for it.
|
# ¿ Mar 1, 2017 23:57 |
|
Ataxerxes posted:How do I determine a good controller? Good controller = Intel. If your mobo has extra SATA controllers like Marvell do not use those.
|
# ¿ Mar 9, 2017 16:33 |
|
Bob Morales posted:Somewhat faster than top of the line SSD, not nearly as fast as RAM, what problem are we solving here? quote:3D XPoint has about one thousandth the latency of NAND flash (or about ten times the latency of DRAM), and tens times the density of DRAM. A nice improvement vs NAND to be sure.
|
# ¿ Mar 19, 2017 19:14 |
|
|
# ¿ Apr 28, 2024 17:53 |
|
quote:Running Sysbench 0.5 with a 70-30 read-write split, and with the data stored on a 400GB Intel DC P3700, the rack could handle only 1395 transactions per second. With the 375GB DC P4800X, that number soars to 16480 transactions per second with similar 99th-percentile latency. That's an impressive performance boost from mostly the same hardware. quote:$1,520 for 375GB. redeyes fucked around with this message at 23:04 on Mar 19, 2017 |
# ¿ Mar 19, 2017 23:01 |