Talk:Solid-state drive/Archive 4
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 |
Persistence
Article states:
SLC to MLC NAND to NOR
10× more persistent 10× more persistent
so both are more persistent? seems strange to me...
- The problem is that there are two different tables made to look like one. SLC is more "persistent" than MLC. Nand is more "persistent" than Nor. It is only a "marketing breakthrough" to attempt to combine these into one imaginary concept. I suggest the two tables be separated. --A D Monroe III (talk) 22:18, 9 July 2014 (UTC)
sequel to SSD
have heard news that there will be a sequel to SSD drives called "RRAM drives":
https://wiki.riteme.site/wiki/Resistive_random-access_memory — Preceding unsigned comment added by Traversable47w0rmh0le (talk • contribs) 10:27, 27 December 2013 (UTC)
- RRAM is one of a few technologies that may well replace NAND flash currently used in most SSDs, but an SSD-compatible product that uses RRAM will likely still be called an SSD. If the external host interface changes it might get re-termed Solid State Storage (SSS) or something, but not due to the internal memory technology. --A D Monroe III (talk) 23:04, 9 July 2014 (UTC)
- Well, there's already a newer external host interface, SATA Express (and associated NVM Express), and SSDs haven't changed their name. But, who knows. :) — Dsimic (talk | contribs) 05:35, 10 July 2014 (UTC)
Understanding the Robustness of SSDs under Power Fault
I see mention of reliability issues - but no mention of the largest study on the matter! See: https://www.usenix.org/conference/fast13/understanding-robustness-ssds-under-power-fault Test35965 (talk) 22:21, 8 July 2013 (UTC)
- Well, the PDF of this study is within the Solid-state drive § External links section, it was added there on July 8, 2014. — Dsimic (talk | contribs) 05:45, 10 July 2014 (UTC)
Dubious
From the comparison table: "HDDs generally have slightly lower write speeds than their read speeds." Well, the bits pass the R/W head in exact same speed during reading and writing, which makes me suggest that the read and write speed must be always the same. Objections? 80.223.182.224 (talk) 16:42, 22 May 2014 (UTC)
- As there was no objections, I modified the text. 80.223.182.224 (talk) 17:17, 30 May 2014 (UTC)
- It isn't dubious. If you look at the detailed specs for hard drives (for example here) you will often find a slightly longer (slower) average seek time listed for writes vs. reads. It's not the actual data transfer, it's the time the drive has to spend to get ready for the transfer. Jeh (talk) 21:07, 30 May 2014 (UTC)
- HDD writes have stricter tracking requirements (seek-settle) than reads; reading a little off-track is harmless, but writing off-track can corrupt adjacent data. This means writes must sometimes wait an extra rev to get fully on track, thus overall writes are slightly slower than reads. --A D Monroe III (talk) 22:11, 9 July 2014 (UTC)
- A D Monroe III, any chances for a reference, please? — Dsimic (talk | contribs) 05:31, 10 July 2014 (UTC)
- HDD companies don't like to publish such inner-working details as to why, but the seek time specs they do publish (such as Jeh's link above) show write seek times about 1ms longer than read seek times. Since the original comparison text didn't give a "why", these data sheets specs should be fine as a reference. --A D Monroe III (talk) 14:54, 10 July 2014 (UTC)
- You're right, but in fact I was hoping for a "why" reference. :) — Dsimic (talk | contribs) 20:35, 10 July 2014 (UTC)
Proposal to merge "Disk on module" with "Solid-state drive"
- The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- The result of this discussion was to merge Ultimatemythbuster (talk) 01:03, 21 July 2014 (UTC)
I would like to propose that the article Disk on module be merged with this article. The former is a stub article and might benefit from being merged into a new section on the SSD page.Ultimatemythbuster (talk) 12:37, 11 July 2014 (UTC)
- Support: Makes sense to me. — Dsimic (talk | contribs) 21:23, 14 July 2014 (UTC)
Is SSD = Solid State Disk a mistake?
There's been a minor edit skirmish (not up to war levels) on the "correct" definition of the acronym for SSD – Solid State Drive or Solid State Disk. I think we agree most sources say Drive, but can we state that saying Disk is "mistaken"? SSDs have neither disks nor drives; I think both are equally "wrong". We'd need more than our own opinions to declare one as a mistake; we'd need a source, and I don't think any respected source could claim either is right or wrong; at most one might be "preferred". So, I've removed any right/wrong indication, and added this per WP:BRD. --A D Monroe III (talk) 20:20, 18 August 2014 (UTC)
- Anon's addition is grammatically incorrect and redundant, since the same parenthetical text already points out that they contain no actual disk. The introductory sentence should keep things simple. Save the nitpicking for later once the basics have been established. --Imroy (talk) 20:38, 18 August 2014 (UTC)
Could you please explain why SSDs are not drives?! Its the same as in HDD = hard disk drive. You wouldn't say hard disk disk, would you? Look at: http://en.wiktionary.org/wiki/drive#Noun "(computing) A mass storage device in which the mechanism for reading and writing data is integrated with the mechanism for storing data, as a hard drive, a flash drive.". So as you can see Drive is correct. And Disk is wrong, obviously.
- First, Anon IP, just to answer your question, HDDs have a "drive" -- the spindle motor; SSDs do not. So, yeah, SSDs aren't "disks", but they also aren't "drives", technically.
- But that really doesn't matter. Some people call SSDs "disks"; that's just a fact. Maybe they "shouldn't", but stating that here won't change that. Even if it could change people's minds, we Wikipedia editors aren't allowed to even attempt to do that. Wikipedia collects, not corrects information. See Wikipedia's core sourcing policy WP:Verifiability, not truth. If you can find a reliable source that states "disk" is "wrong", then we can state that; in fact, must state that. Without such a source, we must just inform readers of the existing situation, however much we might like to wish it were otherwise. --A D Monroe III (talk) 15:12, 20 August 2014 (UTC)
Why do you say "they also aren't "drives""??? Have you read what I wrote?! Look at: http://en.wiktionary.org/wiki/drive#Noun "(computing) A mass storage device in which the mechanism for reading and writing data is integrated with the mechanism for storing data, as a hard drive, a flash drive." So there you have the prove that Drive is correct and no one says Disk is correct. Why don't you change it? — Preceding unsigned comment added by 134.93.55.218 (talk) 16:29, 20 August 2014 (UTC)
- Anon, I did indeed read your source. It definitely implies that an SSD can be called a drive. It does not say it cannot be called anything else. (And it's a tertiary source; Wikipedia relies on WP:secondary sources.)
- Wikipedia policy forbids the change you suggest. Wikipedia editors cannot pass judgment on the correctness of commonly-held information, only report it. (For example, the Holy Roman Empire was not holy, was not Roman, and was not an empire, but that's what everyone calls it, so we call it that too.)
- There is no source given that actually states "Solid State Disk" is "wrong", so we can't say it is.
- Note that even though we state SSD is sometimes called "Solid State Disk", as we must, we currently quickly follow this with "though it contains no actual disk". That's as far as we can go. Please let us preserve this existing compromise. --A D Monroe III (talk) 15:49, 21 August 2014 (UTC)
...and lower power
I suggest adding something about SSD being lower power than spinning drives, in the excellent phrase:
Compared with electromechanical disks, SSDs are typically more resistant to physical shock, run silently, have lower access time, and less latency.[7]
Suggested change: add "require less power"
Compared with electromechanical disks, SSDs are typically more resistant to physical shock, run silently, require less power, have lower access time, and less latency.[7]
Also, since there is an SLC vs MLC table, and now TLC is out, TLC should at least be introduced, perhaps with a reference to: Multi-level_cell
which has been updated to TLC.
WardXmodem (talk) 23:15, 18 September 2014 (UTC)
- Hello! Hm, when compared to HDDs, not all SSDs require less power to operate. Just have a look at those power-hungry SSDs in form of PCI Express expansion cards – some older models needed additional power on top of what PCI Express provides, and even required active cooling. — Dsimic (talk | contribs) 10:47, 22 September 2014 (UTC)
eMMC vs SSDs
This HowToGeek article says that eMMC is different from SSD, I am a bit confused and was wondering if both this and the other article could explain the distinction.
The SSD article is in the Category:Solid-state_computer_storage while the MCC article is in Category:Solid-state_computer_storage_media, so my best guess is that SSD is the hardware while MCCs (cards) are software? But "card" sounds more like hardware to me... why would a card be called a media? --64.228.88.135 (talk) 23:18, 30 November 2014 (UTC)
- I'm unable to find the Wikipedia article for eMCC as your wikilink was a disambiguation page. Can you provide a link please? --Chamith (talk) 00:00, 1 December 2014 (UTC)
- Sorry, wrong consonant-doubling, I screwed up and wrote EMCC by mistake, fixed link to eMMC, embedded-multimedia, C is either card or controller, or possibly both. Right now solid-state storage redirects to this article but perhaps that is wrong and I am misled by that? Is there any kind of SSS besides drives? For example would a MultiMediaCard be described as a "solid state card" with cards being smaller/slower than drives? --64.228.88.135 (talk) 00:22, 1 December 2014 (UTC)
- No, solid-state storage doesn't redirect to eMMC. Those are two different articles about two different storage medias. And you asked whether why eMMC falls into Category:Solid-state_computer_storage_media. Actually the word media doesn't mean software, it stands for a particular form of storage material for computer files, for example magnetic tape or discs. In other words it's a physical media. eMMCs are memory cards and it is also known/used as non-traditional SSDs. You can describe it as a "solid state card" if you like because it is a solid state storage media. Like hard drives SSD technology has improved a lot. You can consider eMMC as a basic step/version of SSDs available now.--Chamith (talk) 00:36, 1 December 2014 (UTC)
- Sorry, wrong consonant-doubling, I screwed up and wrote EMCC by mistake, fixed link to eMMC, embedded-multimedia, C is either card or controller, or possibly both. Right now solid-state storage redirects to this article but perhaps that is wrong and I am misled by that? Is there any kind of SSS besides drives? For example would a MultiMediaCard be described as a "solid state card" with cards being smaller/slower than drives? --64.228.88.135 (talk) 00:22, 1 December 2014 (UTC)
Thank you for the answer, beginning to clear up a bit. Based on MMCs being solid-stage storage but NOT a solid-state drive, I am going to change the solid-state storage redirect to this article into a disambig and use this SSC term. I also need to amend that edit based on misunderstanding to the MMC intro. --64.228.88.135 (talk) 01:13, 1 December 2014 (UTC)
Sales section (More information suggested)
Hi, can u please give us any information about the percentage of how many notebooks and desktop systems have those SSDs integrated? And how many computers will have them in future? --Martin — Preceding unsigned comment added by 47w0rmh0le (talk • contribs) 17:26, 22 December 2014 (UTC)
SSD automatically fragments everything to avoid hot spots?
Pink love 1998 (talk · contribs) added this to the SSD/HD comparison table, in the row about fragmentation:
The critical purpose of the SSD algorithm is to distribute the data to various locations to prevent heat build up on a particular spot, so SSD's are typically always fragmented.
with edit comment:
Added info, please don't delete I don't know how to insert references and advise you to read about SSD's
Inserting references is only barely more complicated than just typing the reference info in prose. Just add <ref> (your reference) </ref> following your text. If you want to be precise about location, the ref goes before any immediately following space, but after any immediately following punctuation.(like here) Others will improve the formatting by using a cite template. Eventually you'll pick up on that.
Or, you could add a section to the talk page here, giving your refs. Someone else will pick them up and put them in the article.
However, even if referenced, this claim raises an issue, and this is the real reason I deleted it. When most people speak of "fragmentation" on a hard drive, we're talking about the fact that large (and sometimes small) files do not occupy contiguous ranges of LBAs. This fragmentation is visible to a file system from the outside of the drive, can be corrected by defragmenters, etc. This has well-known impact on HD performance. It even impacts SSDs, although minimally: to access a part of the file that crosses an extent boundary, two different I/O requests must be issued by the FSD and implemented by the disk driver and other drivers in the stack.
The... call it distribution of content from sequential LBAs to different areas of the chips in an SSD is different. It is not visible to e.g. file system drivers, and a defragmentation run would not "fix" it. To the host speaking to the drive through its SATA connector, a file that occupies sequential LBAs still appears to be contiguous even though those LBAs might not be physically contiguous on the chips.
There is no analog to this in a normal hard drive, other than the occasional "sparing" of bad sectors.
If this point can be referenced to a WP:RS, it certainly belongs in the article. But I don't think it belongs in the table row that discusses file system fragmentation, certainly not as the first sentence in the entry for SSDs. That table row is just not talking about this sort of thing. Wherever it goes, it needs to be described so as to distinguish it from the non-contiguous-LBA sort of fragmentation. Jeh (talk) 21:04, 13 January 2015 (UTC)
- Yes, isn't heat the reasonable cause for the fragmentation. The SSD keep track of the next memory location using a link pointer. With the argument of latency (very low latency consider it as RAM in RAM the access speed of all locations is the same) it doesn't matter where the data is stored and because of the higher latency compared to that of HDD, defragmentation is needed because it takes longer to access data from random location(it's funny do you know that ssd uses DDR2 ram technology) . — Preceding unsigned comment added by Pink love 1998 (talk • contribs) 22:28, 13 January 2015 (UTC)
- Hello! As Jeh already explained it very well, fragmentation pretty much doesn't exist in the context of SSDs. Pink love 1998, your post seems highly confused, as I simply don't understand what heat, link pointers, and DDR2-related technology you refer to? Again, defragmenting an SSD can't do anything but shorten its life by wearing out underlying flash memory. — Dsimic (talk | contribs) 00:12, 14 January 2015 (UTC)
- Oh, fragmentation still exists. It just doesn't matter (except only just barely due to increased numbers of I/O requests to the drive).
- No, SSDs don't use DDR2 technology. DDR means "dual data rate", which describes the bus the DIMMs plug into, not the memory cell structure. Jeh (talk) 02:00, 14 January 2015 (UTC)
- Right, thanks for the correction. I didn't express myself clearly enough, it should've been something like "fragmentation pretty much doesn't matter in the context of SSDs, compared to the way it affects the operation of HDDs". At the same time, SSDs might use DDR2 memory for their buffers or storage of in-memory metadata structures, but that has nothing to do with the fragmentation of stored data. — Dsimic (talk | contribs) 02:33, 14 January 2015 (UTC)
Over provisioning
A section on over provisioning would be useful imo. Its a popular issue is discussed alot. Chendy (talk) 14:30, 12 February 2015 (UTC)
Maximize SSD Lifetime and Performance With Over-Provisioning
- Hello! Over-provisioning is already mentioned a few times in the article, linking the term to the Flash over-provisioning article (better said, a redirect), which provides a rather good description. Repeating that in greater detail might be pretty much redundant, if you agree. — Dsimic (talk | contribs) 03:12, 17 February 2015 (UTC)
Amdahl's law
"In applications where hard disk seeks are the limiting factor, this results in faster boot and application launch times (see Amdahl's law).[105]" I know the law and that SSDs are better under parallel I/O load, but is the law applicable here? Is it immediately obvious to people why or is this WP:OR? Note, I didn't find the law in the ref (the first page, there are 17..). comp.arch (talk) 13:09, 27 February 2015 (UTC)
- Hello! Well, Amdahl's law is "used to find the maximum expected improvement to an overall system when only part of the system is improved". SSDs are obviously only one part of a computer system, but I'd remove the "(see Amdahl's law)" wording anyway. — Dsimic (talk | contribs) 12:19, 2 March 2015 (UTC)
conventional hard drives
Congratulations, now most web pages and government documentation refer to Conventional Hard Drives (instead of Magnetic Hard Drives, etc ). Is it possible to state that Conventional means the majority of the mechanical/magnetic hard drives at the time of writing?
10 years time, conventional may mean a completely new type of Hard Drive. — Preceding unsigned comment added by 164.39.36.212 (talk) 14:56, 29 October 2015 (UTC)
External links modified
Hello fellow Wikipedians,
I have just added archive links to one external link on Solid-state drive. Please take a moment to review my edit. If necessary, add {{cbignore}}
after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}}
to keep me off the page altogether. I made the following changes:
- Added archive https://web.archive.org/20120203173241/http://blog.2ndquadrant.com:80/en/2011/04/intel-ssd-now-off-the-sherr-sh.html to http://blog.2ndquadrant.com/en/2011/04/intel-ssd-now-off-the-sherr-sh.html
When you have finished reviewing my changes, please set the checked parameter below to true to let others know.
An editor has reviewed this edit and fixed any errors that were found.
- If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
- If you found an error with any archives or the URLs themselves, you can fix them with this tool.
Cheers.—cyberbot IITalk to my owner:Online 22:38, 17 January 2016 (UTC)
Done, looks good. — Dsimic (talk | contribs) 08:02, 18 January 2016 (UTC)
Contentious
This page is very contentious. Could it be made to read less as a defense of HDD and more informative. That'd be great.2602:304:B0D8:6220:C5B4:6BAA:AC2F:43B7 (talk) 19:59, 28 May 2016 (UTC)
- Contentious? The article states SSDs are way faster, but way more expensive than HDDs. Can you give examples of what needs improving? --A D Monroe III (talk) 21:02, 15 June 2016 (UTC)
- Concur. Please give specific examples of where you think it reads "as a defense of HDD". Jeh (talk) 15:54, 17 June 2016 (UTC)
Hard drives and altitude
This portion needs a substantial adjustment. As the cited source notes, many hard drives have an altitude limit of 12000 meters above sea level, but drives designed for high-altitude operation are readily available from manufacturers. This is also becoming less true, helium-filled hard drives are taking over for large capacity hard drives and if they're impervious to helium, high-altitudes are unlikely to be an issue. The mention of the breather hole should be merged with this. 207.172.210.101 (talk) 01:30, 24 June 2016 (UTC)
- As always, we'd need references that helium-filled drives can also work properly on high altitudes. — Dsimic (talk | contribs) 13:57, 24 June 2016 (UTC)
SLC, MLC and TLC NAND
Should this not be included in the article? http://www.speedguide.net/faq/slc-mlc-or-tlc-nand-for-solid-state-drives-406 BP OMowe (talk) 14:41, 25 August 2016 (UTC)
TBW
I have seen the term TBW used in SSD specifications and came here to try to find out what it meant. I was disappointed. I eventually discovered that it stood for Terabytes Written in reference to an SSD's expected lifetime. Please can one of the article's main editors add this. Viewfinder (talk) 13:03, 16 November 2016 (UTC)
Graphics
The most intuitive dysfunction refers to the moving parts in an HDD. Some people are unclear on that through influence only. As a matter of fact, it IS known that there are no moving parts in an SDD, like the disk found in the HDD, which IS slower than an SSD. For future reference try checking out how an SSD IS like a disc. — Preceding unsigned comment added by 2602:301:7751:160:140f:74da:270a:b1df (talk • contribs) 12:17, 21 August 2017 (UTC)
Alignment in Wikitables
I think the data looks better in the section I edited (30/12/2017), but would look better still if the colon's were aligned instead of simply centering the text.
In MathJax this is `\begin{align} ... &: ... &: ... \end{align}`, with the ampersand aligning the next character, in this case the colons, but is it possible with Wiki markup?
Darcourse (talk) 07:57, 30 December 2017 (UTC)
Susceptibility to magnetic fields
This section is bad. The cited reference includes some talk on the subject, but lacks any expert statements. At best it includes a few people who claim hard drives were damaged by magnets, but I doubt many of the individuals are experts. The magnetic fields required for writing to a modern hard drive are very intense. My understanding is you could put a rare earth magnet directly onto a modern disk platter and the magnetic field of the magnet would fail to damage any data (instead dust and tiny scratches from the contact might well damage the platter). 207.172.210.101 (talk) 01:40, 21 June 2016 (UTC)
- Hello! You're right, thank you for pointing it out! The reference was a low-quality one, so I went ahead and made the changes that provided accurate information and much better references. — Dsimic (talk | contribs) 08:42, 21 June 2016 (UTC)
- Those are decent, but this these mean that portion of the article needs fixing. Ideally I'd link to sections 4 and the epilogues of the first link (Peter Gutmann paper), which effectively say modern disks (>1GB) are essentially immune to external magnetic fields. The kjmagnetics reads like an amateur experiment (not necessarily bad, but be careful of conclusions!) and says the same thing, their report of mechanical scrapping could well have been due to distorting the case of the drive rather than anything having to do with properties of the magnetic field. "Very old hard drives (less than a gigabyte) may have been at some risk from external magnetic fields, but any drive larger than a gigabyte is essentially immune to external magnetic fields"? 207.172.210.101 (talk) 20:09, 21 June 2016 (UTC)
- Hm, I'm not sure that 1 GB is specified in references as a clear capacity-based division between susceptible and resistant drives... Am I missing something? — Dsimic (talk | contribs) 13:52, 24 June 2017 (UTC)
- Indeed, 'tis not. The reference was stating post-1990 hard drives were pretty well immune. Bit more recollection, I think 100MB drives were coming out around then, so that may be a better rough guide. The real issue is larger drives have to be less susceptible otherwise the write process would corrupt nearby bits (therefore storage size is a better guide than manufacture date). I don't have any references other than my memory. 207.172.210.101 (talk) 22:23, 25 June 2016 (UTC)
Soppprt versi.9.0 Samsungalaxy (talk) 14:52, 23 February 2019 (UTC)
64GB SATA SSD from 1978; 41 years ago?!?
SATA was created in the year 2000 ( https://wiki.riteme.site/wiki/Serial_ATA ). And in 1978 the best hard disks were in the megabyte range... So please consider the following image caption in the article as maybe not fully correct:
A Super Talent Technology 2.5" Serial ATA solid-state drive Date invented 1978; 41 years ago Invented by Storage Technology Corporation — Preceding unsigned comment added by 193.170.243.131 (talk) 12:22, 27 March 2019 (UTC)
- It was correct but misleading, better now. Tom94022 (talk) 19:48, 27 March 2019 (UTC)
A Commons file used on this page has been nominated for deletion
The following Wikimedia Commons file used on this page has been nominated for deletion:
Participate in the deletion discussion at the nomination page. —Community Tech bot (talk) 14:51, 5 April 2019 (UTC)