Talk:SATA Express

Compatibility and SATA ports

Here's a good question from Matthiaspaul, provided as an inline comment within the article — so I'll quote it and move here for a better chance of sparking a discussion:

What about the (IMO more important) other way around: Is the embedded SATA port a mandantory feature or optional for the vendors? Will it be possible (perhaps by accepting a speed penalty) to access SATA Express drives with legacy operating systems (such as DOS), which to not have disk drivers and do not access the hardware directly, but rely solely on the BIOS INT 13h to access media (or on EFI machines, an emulation of the BIOS), and where this BIOS may only use legacy PATA commands?

I'll try to provide an answer in a few minutes. -- Dsimic (talk) 23:06, 2 October 2013 (UTC)

• SATA 6 Gbit/s port is a part of the SATA Express specification, as far as I can see — so it's mandatory for manufacturers to include it in the pinout, so to speak. On the other hand, AHCI compatibility is there for both SATA SSDs and for PCI Express SSDs, so it doesn't depend on whether SATA port (better said, SATA device) is used or not — PCI Express SSDs can be accessed through the AHCI driver as well as through the NVMe driver, as it's listed within the article.
  Everything so far was clear, now we go into the muddy waters. :) Having INT13h support implemented is up to BIOS for the controllers running in AHCI mode, and not in the ATA compatibility mode. Personally, I've never used that, so I can't speak from my experience... Having a look into people's experiences, it seems that some BIOSes are doing proper INT13h emulation for AHCI controllers, while other BIOSes unfortunately don't.
  Hope it helps. -- Dsimic (talk) 23:22, 2 October 2013 (UTC)
  • In other words, that's more into the territory of BIOSes and their INT13h emulations. -- Dsimic (talk) 23:41, 2 October 2013 (UTC)
    • Here's a good paper on the Compatibility Support Module (CSM). --- Dsimic (talk) 21:14, 6 October 2013 (UTC)

Stimulus for SATAe

See also: User talk:Dsimic § eSATA

(This revert by Dsimic)

The previous intro (which should summarize the article) contains a bunch of non-sequiteurs that a reader could not contextualize. For example, it abruptly began its main paragraph this way:

"Instead of the usual approach of doubling the speed of SATA interface, PCI Express was selected for achieving data transfer speeds greater than 6 Gbit/s. It was concluded by the designers of SATA interface that doubling the native SATA speed would take too much time, too many changes would be required to the SATA standard, and would result in much greater power consumption..."

What "usual approach"? (Usual approach for what?) Selected for speeds "greater than 6Gb/s" (What is special about "selecting" a greater speed than the existing speed for any next generation successor to anything?). Doubling the native speed would "take too much time" (compared to what? Why was there time pressure? Nothing hints at a context at all). It would take "too many changes" (uncontexted) and "much greater power consumption" (than what? By what standard of "too much"?)

The paragraph simply didn't make sense and a reader who read the intro should be able to get the "gist" of the key points, which they could not. The reason for all these issues was as follows:

• SATA 1, 2 and 3 were designed for HDDs
• SSDs became a rapid growth/adoption area shortly after SATA 3 came out, and within 2 years could swamp SATA 3 and needed more bandwidth.
• Developing a replacement on such short notice wasn't practical (explains the issue with "time"), and would have required disadvantageous power and cost as well (explains the issues with "changes" and "power").
• An interface with the required bandwidth and power handling was already well established, commonplace and mature, so they adopted that for the next generation of SATA instead (explains that SATAe uses PCIe and why this is so).

@Dsimic: Can you find a way to merge these into the existing wording, if you don't like the way I did it? The issue needs resolving. FT2 ^{(Talk | email)} 00:50, 11 July 2014 (UTC)

Hello there! You're totally right, the lead section's paragraph quoted above isn't that great for readers not already familiar with the history of SATA, its usage over time etc. Thank you for pointing that out, and I'm already working on expanding that paragraph according to your suggestions. — Dsimic (talk | contribs) 00:55, 11 July 2014 (UTC)

Done, please check out changes to the article. In a few words, SATA Express § History section was expanded first according to your suggestions, and then the lead section was edited to sum it up properly. Though, lead section is still quite brief (when compared to your edit), but providing brief summaries is what the lead sections are supposed to do – hope you'll find it good and clear enough.

Of course, I'm more than open to any further discussion. — Dsimic (talk | contribs) 02:45, 11 July 2014 (UTC)

Hey FT2, your latest additions to the article aren't bad per se, but they're simply way too wordy and make the article look like a forum post. In other words, we don't need to explain every single bit of the history events which led to the invention of SATA Express, and the style of language should be a bit more formal. I've reverted your edits for now, while I'll go through them in detail later today and integrate them into the article.

Hope you're fine with that. — Dsimic (talk | contribs) 14:33, 17 July 2014 (UTC)

I'm fine with tone improvements - that's always good and rarely an issue.

But I am a little concerned that your idea of what an article should contain, means effectively, stripping out how the article's topic came about, which is core historical information, and is worth the extra space and lines. I'm reassured you will go through them but I'm concerned that you might ignore that the historical background is key here and not in any way "off topic". Please consider that the actual information you reverted should probably be reinstated in full, with tone improvements or shortening if you can find good ways to do that. I'm out for a few hours, please think on this and I look forward to seeing whats up when you are happy with your updates! FT2 ^{(Talk | email)} 14:44, 17 July 2014 (UTC)

Don't worry, I agree that historical background is important, it's just that we don't need to cover every single step, and to explain every single bit. :) I'll go through everything in detail later today or early tomorrow latest, and I think you'll be satisfied with the end result. :) — Dsimic (talk | contribs) 15:01, 17 July 2014 (UTC)

I'd say that I have a good solution for how to include a plenty of historical details without clogging up the SATA Express § History section, just please give me some time. :) — Dsimic (talk | contribs) 15:13, 17 July 2014 (UTC)

Well, I've gone over and over your edits on the SATA Express § History and lead sections, therefore I've spent a lot of time and put a lot of effort into that, and after defluffing and compaction ended up with these edits to the article.

Please don't get me wrong – I'm not disrespectful by any means, but we simply don't need to provide a "fool proof" explanation of everything that happened under the sun since SSDs even became available. Also, for example, it's even better not to mention specific manufacturers while describing such a timeline, as that eliminates any chances for bordering the timeline with an advertisement while not losing anything of the overall value.

I know, you'll feel that a lot of detail was lost during the compaction I've performed, but you'll also see that pretty much eveything that's important is still there, incorporated in a formal tone into the rest of the SATA Express § History and lead sections. With all that additional content, if a reader is unable to grasp the whole thing, I'd say that reading a different kind of articles should be suggested. :)

Of course, I'm more than open to further discussion. However, please keep in mind that I've spent a lot of time doing this, and while I respect all editors, I'd expect the same level of treatment for myself. — Dsimic (talk | contribs) 06:36, 18 July 2014 (UTC)

Connector mating matrix: table labels upside down?

Is it just me, or are the column labels at the tops of the columns in the "Connector mating matrix" table (within the "Connectors" section) upside down? — Preceding unsigned comment added by Rfg-ronbaby (talk • contribs) 19:15, 29 December 2017 (UTC)

"up to 1969 MB/s"?

Seeing the calculation of the figure in the text of the article leads me to believe that it only includes the PCIe lines. But the description of the standard postulates that they are an addition to the standard SATA 3.0 lines, not a replacement. If so, then the speed should be 1969 MB/s + 750 MB/s = 2719 MB/s, no? DenisRS (talk) 06:53, 31 May 2019 (UTC)

A SATA device uses either native SATA (up to 600 MB/s) or PCIe (up to 1969 MB/s) with PCIe 4.03.0. SATA doesn't use multipathing, so you can't aggregate. --Zac67 (talk) 07:21, 31 May 2019 (UTC)

Furthermore, PCIe lane 0 uses the same pins as the SATA data pairs, so even if multipath were supported at higher layers, there simply aren't enough pins to accomodate both. I've edited the one sentence in the article I could find that might give you the wrong idea. Is it clearer like this? @Zac67:, I think you meant PCIe 3.0 rather than 4.0? Digital Brains (talk) 08:55, 31 May 2019 (UTC)

Yes, thx @Digital Brains: --Zac67 (talk) 09:29, 31 May 2019 (UTC)

Somehow I had forgotten about the second SATA data port :). PCIe lane 0 shares pins with the SATA data pairs of one SATA port and PCIe lane 1 shares pins with the SATA data pairs of the other port. Digital Brains (talk) 09:43, 31 May 2019 (UTC)

Can we merge this page with the main SATA page?

It makes sense to not waste an entire article for something that could just be a subheadding. Alexceltare2 (talk) 14:54, 25 January 2023 (UTC)