Linked by Thom Holwerda on Tue 21st Nov 2017 16:03 UTC
Windows

I wiped off my Windows 10 installation today. It wasn't because of the intrusive telemetry or the ads in the start menu but desktop composition. It adds some slight but noticeable latency that makes typing feel uncomfortable. In Windows 7 you can turn it off.

If you're fine with unresponsive UI operations and graphical tearing, then, sure, go back to Windows 7 or earlier and turn off compositing to get a few ms back when typing.

Order by: Score:
Linux + xOrg
by Bill Shooter of Bul on Tue 21st Nov 2017 16:40 UTC
Bill Shooter of Bul
Member since:
2006-07-14

That setup allows turning off vsync, if you really want to do that like a crazy person.

Just search vsync, and you'll find tons of people trying to enable that in xorg, rather than disable. Its better with it on.

Reply Score: 6

RE: Linux + xOrg
by Megol on Thu 23rd Nov 2017 12:27 UTC in reply to "Linux + xOrg"
Megol Member since:
2011-04-11

That setup allows turning off vsync, if you really want to do that like a crazy person.


Crazy... Says more of you than anyone else.


Just search vsync, and you'll find tons of people trying to enable that in xorg, rather than disable. Its better with it on.


Why is it better? And you better have a technical and accurate argument.

Reply Score: 0

RE[2]: Linux + xOrg
by Bill Shooter of Bul on Sun 26th Nov 2017 05:31 UTC in reply to "RE: Linux + xOrg"
Bill Shooter of Bul Member since:
2006-07-14

Crazy... Says more of you than anyone else.

Basically its a tradeoff. Either you have screen tearing or increased latency. Most people prefer not having screen tearing. This input lag thing is kind of nuts. Maybe you can notice it, but we're starting to get into pseudo audiophile territory here. If you start telling me it can't be rigorously tested, then I'll have all the proof needed...

Seriously would be interested in a set up that had vsync on and off with a xorg windows system and run users through various tasks to see which set up they preferred, and if they could actually detect the latency. I'd completely accept the results of a proper study regardless of my initial assumptions.

Reply Score: 2

RE[3]: Linux + xOrg
by ssokolow on Sun 26th Nov 2017 13:57 UTC in reply to "RE[2]: Linux + xOrg"
ssokolow Member since:
2010-01-21

Thankfully, adaptive sync solutions like FreeSync 2 and G-Sync should eventually reduce the trade-off to irrelevancy.

(I wouldn't consider G-Sync a valid contender in the long run compared to an open standard like the optional adaptive sync portion of the DisplayPort 1.2a spec that FreeSync relies on, but G-Sync apparently doesn't require the proprietary scaler module in laptops since the display is connected more directly to the GPU.)

Edited 2017-11-26 13:58 UTC

Reply Score: 2

RE[3]: Linux + xOrg
by Megol on Sun 26th Nov 2017 19:04 UTC in reply to "RE[2]: Linux + xOrg"
Megol Member since:
2011-04-11

"Crazy... Says more of you than anyone else.

Basically its a tradeoff. Either you have screen tearing or increased latency. Most people prefer not having screen tearing. This input lag thing is kind of nuts. Maybe you can notice it, but we're starting to get into pseudo audiophile territory here. If you start telling me it can't be rigorously tested, then I'll have all the proof needed...
"

Ah, you actually know what you are talking about! Have to apologize as I thought you were yet another complainer without any clue. Well except for the "crazy" thing, still applies.

Today we have double buffered and synchronized drawing. It doesn't show tearing but it have a worst case latency approaching 2x the update rate of the screen.

IOW we have in a normal system a latency of 1/60 to 1/(2*60) seconds for a buffered, synchronized update. That is ~16.7 to ~33.4 milliseconds.

But that is only the raw latency ignoring any input -> processing -> update overheads.

This is obviously detectable by a normal human being and irritating for those that have better response time like FPS gamers.

Tearing while distracting isn't normally detectable. The most visible cases are those of moving chunks of screen data like when scrolling or moving windows.

But there are techniques for reducing those cases. Or was. It's probably harder to avoid it when targeting decoupled streaming optimized processors (=GPU).

One easy technique is optimizing for what the user is most likely looking at - text being entered or the mouse pointer. Prioritizing updates close to where the user is looking makes tearing artifacts less visible.

My main point is this: no vsync can result in visible tearing but generally lowers the perceived input->output latency. Why would it be crazy if someone prefers the possibility of tearing over a detectable additional latency? It's just a matter of preference.


Seriously would be interested in a set up that had vsync on and off with a xorg windows system and run users through various tasks to see which set up they preferred, and if they could actually detect the latency. I'd completely accept the results of a proper study regardless of my initial assumptions.


That would be interesting.

Reply Score: 2

RE[4]: Linux + xOrg
by zima on Mon 27th Nov 2017 13:55 UTC in reply to "RE[3]: Linux + xOrg"
zima Member since:
2005-07-06

This is obviously detectable by a normal human being and irritating for those that have better response time like FPS gamers.

Or so they claim... but one has to wonder how many are placebophiles (the author of the discussed news article certainly seems to be one) - for example, they might be noticing, when v-sync is disabled, not the decreased latency but tearing, and "convince" themselves it's better that way because it "should be" (so for a proper ABX test, you'd have to introduce "artificial" tearing when v-sync is enabled, to get around such easily noticed visual cue)

Reply Score: 2

Office 2016 Lag
by daedalus on Tue 21st Nov 2017 16:59 UTC
daedalus
Member since:
2011-01-14

Not sure if it's related to compositing in Windows itself or just in Office, but ever since we updated Office in my office (on PCs running Windows 7), I find using Word, Excel etc. is a horrible experience. Drag-selecting, typing, scrolling, zooming, all suddenly feel laggy and disconnected.

But that's ok, because now the text cursor fades in and out instead of instantaneously turning on and off, presumably for that more luxurious feel... Or something.

Reply Score: 4

RE: Office 2016 Lag
by Kochise on Tue 21st Nov 2017 17:03 UTC in reply to "Office 2016 Lag"
Kochise Member since:
2006-03-03

"Improving users' experience"

Reply Score: 2

RE: Office 2016 Lag
by silviucc on Tue 21st Nov 2017 17:12 UTC in reply to "Office 2016 Lag"
silviucc Member since:
2009-12-05

Are you by any chance using a high polling rate mouse (like a gaming model)?

I have the same issue.

More on this thread:
https://www.reddit.com/r/Windows10/comments/5fkyxt/windows_10_lag_wi...

Reply Score: 4

RE[2]: Office 2016 Lag
by daedalus on Wed 22nd Nov 2017 08:48 UTC in reply to "RE: Office 2016 Lag"
daedalus Member since:
2011-01-14

Nope, it's a bog standard HP mouse, and the Windows operations themselves (Explorer, other applications, dragging windows around) are all fine.

Reply Score: 2

RE: Office 2016 Lag
by Morgan on Wed 22nd Nov 2017 12:43 UTC in reply to "Office 2016 Lag"
Morgan Member since:
2005-06-29

I would venture to say it's Office. I'm keeping our systems at work on Windows 7 for as long as it's safe to do so. I'm currently testing Office 2016 on one of our workstations and it has introduced severe lag when typing. It's a weird kind of lag too; you can type a character in any Office program and after about half a second it will generate the character on the screen, and it does so by wiping from left to right at what appears to be 60Hz. In other words, it's synced to the screen refresh but it's delayed by around 500ms. It's simultaneously laggy and smooth, and it makes for some seriously frustrating typing sessions.

Going back to the previous Office version (2010) eliminates the bug.

Reply Score: 2

RE[2]: Office 2016 Lag
by Odwalla on Wed 22nd Nov 2017 13:51 UTC in reply to "RE: Office 2016 Lag"
Odwalla Member since:
2006-02-01

That's the Office typing animation. You can disable that.

Reply Score: 2

RE[3]: Office 2016 Lag
by Morgan on Wed 22nd Nov 2017 16:15 UTC in reply to "RE[2]: Office 2016 Lag"
Morgan Member since:
2005-06-29

Dear god, having to go into the registry to disable something that should never have been created in the first place...that's maddening.

Thanks for the tip!

Reply Score: 2

RE: Office 2016 Lag
by Greatquux on Wed 22nd Nov 2017 13:33 UTC in reply to "Office 2016 Lag"
Greatquux Member since:
2017-02-17

I have found disabling animations helps - this is in Windows system properties. Excel and Word feel a lot faster then. But yeah, Office has been getting worse since about 2010 or so.

Reply Score: 1

RE[2]: Office 2016 Lag
by daedalus on Wed 22nd Nov 2017 15:38 UTC in reply to "RE: Office 2016 Lag"
daedalus Member since:
2011-01-14

Hmmm, will have to look at disabling them so... I wonder how long they'll keep that option around for ;)

Reply Score: 2

RE: Office 2016 Lag
by Parry on Sat 25th Nov 2017 17:11 UTC in reply to "Office 2016 Lag"
Parry Member since:
2014-06-03

Office 2016 had a bug (was fixed just a few weeks ago) on Surface Pros that caused a typing lag of a about 1 second when working in large documents. It was affecting ALL our Surface Pro users and we can to roll them back to Office 2010. Turning off Office '16 animations didn't help but thankfully a recent update seems to have fixed the problem.

Reply Score: 1

It's not really funny, but true.
by Vaisanen on Tue 21st Nov 2017 17:06 UTC
Vaisanen
Member since:
2017-11-21

"Don’t you find it a bit funny that Windows 95 is actually snappier than Windows 10? It’s really a shame that response times in modern computers are visibly worse than those in twenty years ago."

Not only that, it's also much less bloated, being that it was developed in times, when hardware constraints were actually taken into account. Didn't serve ads, either, a huge boon in itself. Unresponsive, sure, sometimes, especially if one was careless. Graphical tearing, maybe, but an OS's main function is not to be looked at. And if it was still supported, I'm certain I wouldn't be the only person to go back to it.

Reply Score: 2

Sauron Member since:
2005-08-02

Too true, my Windows 98 computers agree with you. And they are blown away by my 25 year old Amiga! ;)

Reply Score: 6

Veto Member since:
2010-11-13

Yeah. The Amiga was very much a fully synchronous vsynced machine making the graphics and desktop super fluid - as long as it could keep up.
Even the mouse was synchronous, as it was sending the movement pulses directly and the counters read at every vsync, instead of using a latency-inducing serial bus.
But it was also a simpler time with 2D-only graphics and few or no layers between the application and hardware...

Reply Score: 5

daedalus Member since:
2011-01-14

Sounds like you didn't get to grips with the pretty simple concepts involved. And not sure why you're talking about OS 1.0 - a lot of stuff improved even in the early days - 1.2 or 1.3 were standard in 1987.

Anyway, the poster was talking about a 25-year-old Amiga, so likely an OS 3 machine. Icons stayed wherever you put them if you told them to, lassoing icons was indeed possible (including later adding and removing items from the selection, and selection across multiple windows), and very little in the way of basic operations required the Shell - in fact I can't think of anything that could be described as a "basic operation" that Workbench didn't cover. Sounds like you were a jealous ST owner back in the day.

Reply Score: 3

Sauron Member since:
2005-08-02

Correct, I was referring to my A1200. Although I do have a A500, a A500+ and a A2000 in my arsenal, my 1200 gets the most use.
Like you I haven't got a clue what the poster is on about, I get non of that even on my Kickstart 1.2 A500.
Anyway, all off topic now but to add to the topic, I run Windows 7 and don't experience any lag of any sort that I've noticed.
I won't run Windows 10, not now, not ever. It takes Operating Systems to a level I don't agree with and won't enter!

Reply Score: 1

Bill Shooter of Bul Member since:
2006-07-14

You're insane. It was also developed in times where security was an after thought as were buffer overflows. Blue screens were a daily occurrence.

Also, back in the day, we considered it a resource hog. The upgrade from win 3.1 required more ram, better graphics cards, and cpu cycles.

Edited 2017-11-21 17:44 UTC

Reply Score: 4

leech Member since:
2006-01-10

At least Win9x didn't randomly allow some applications to stream, while others can't even load a web page. Not sure what's causing it yet, as I've disabled the network throttling in the Windows 10 registry...

I still remember when I upgraded to 32mb of ram and thought it was ridiculous for an operating system to require so much, when you could run TOS in 1mb with room to spare. Still loved that I had a 4mb Mega STe and created a 2mb RAMDISK, copied over all of Ultima 6 into it and played the game out of RAM.

Reply Score: 2

Bill Shooter of Bul Member since:
2006-07-14

Not familiar with the issue in win 10 you are referencing. Works fine for me ( for various definitions of "fine").

Also not sure I have any programs that load web pages that aren't themselves web browsers ( no web view apps) . Could be an internal issue in the app.

Reply Score: 3

darknexus Member since:
2008-07-15

You're insane. It was also developed in times where security was an after thought as were buffer overflows.

That's not the point, unless you're seriously trying to say that a secure operating system needs to be slow as mud in order to be that way. I think you're a lot smarter than that.

Reply Score: 1

Bill Shooter of Bul Member since:
2006-07-14

No, the point is you shouldn't use win 9x. It was a terrible code base riddled with bugs. Don't pretend it was an awesome experience. It was not.

Reply Score: 2

RE: It's not really funny, but true.
by rtfa on Wed 22nd Nov 2017 09:33 UTC in reply to "It's not really funny, but true."
rtfa Member since:
2006-02-27

Well, yeah. Windows 95 was sh*t and very bare. You could press Escape in the login dialog and go straight to the desktop. Plus, what ever Intel gives, Microsoft takes away

Reply Score: 0

canary in the coalmine
by garyd on Tue 21st Nov 2017 17:49 UTC
garyd
Member since:
2008-10-22

I appreciate the benchmarking and scientific method used to prove the theory but if this were really an issue with the OS I would expect that gamers would have started howling about this a couple of years ago when Win10 was still in beta test.

Edited 2017-11-21 17:51 UTC

Reply Score: 3

v RE: canary in the coalmine
by TemporalBeing on Tue 21st Nov 2017 21:05 UTC in reply to "canary in the coalmine"
RE[2]: canary in the coalmine
by feamatar on Tue 21st Nov 2017 21:53 UTC in reply to "RE: canary in the coalmine"
feamatar Member since:
2014-02-25

what is that special sku?

Reply Score: 1

RE[3]: canary in the coalmine
by TemporalBeing on Tue 21st Nov 2017 22:17 UTC in reply to "RE[2]: canary in the coalmine"
TemporalBeing Member since:
2007-08-22

what is that special sku?


Looks like it became "Game Mode" - https://www.pcworld.com/article/3194510/windows/how-to-activate-game...

Something you can activate/deactivate just to get higher performance while gaming.

Reply Score: 4

RE[4]: canary in the coalmine
by Nelson on Wed 22nd Nov 2017 18:09 UTC in reply to "RE[3]: canary in the coalmine"
Nelson Member since:
2005-11-29

This is not a special SKU.

Reply Score: 3

RE[5]: canary in the coalmine
by TemporalBeing on Wed 22nd Nov 2017 18:12 UTC in reply to "RE[4]: canary in the coalmine"
TemporalBeing Member since:
2007-08-22

This is not a special SKU.


Initial stories around it was it was going to be. Looks like they just rolled it into an update; though don't be surprised if you need certain SKUs to use it.

Reply Score: 0

RE: canary in the coalmine
by Drumhellar on Wed 22nd Nov 2017 07:32 UTC in reply to "canary in the coalmine"
Drumhellar Member since:
2005-07-12

IIRC in Windows, when a game is fullscreen, compositing is bypassed, so it doesn't actually affect games.

Reply Score: 4

RE: canary in the coalmine
by Odwalla on Wed 22nd Nov 2017 14:00 UTC in reply to "canary in the coalmine"
Odwalla Member since:
2006-02-01

There was no scientific method involved. The blogger admits his data gathering is suspect, admits his posted data is probably skewed to be worse than it really is, and only tested a single computer. That's about as anti-scientific method as you can get.

Reply Score: 3

RE[2]: canary in the coalmine
by grat on Wed 22nd Nov 2017 19:29 UTC in reply to "RE: canary in the coalmine"
grat Member since:
2006-02-02

Don't forget the blogger completely left out the specifications of his monitor-- Many modern LCD monitors are slow by default, and latency of 20+ ms isn't unusual.

I can understand the frustration with latency, but not for typing-- I learned to develop over a 1200 baud modem, and can type 60+ WPM, so talking about "latency" makes me giggle.

Latency in gaming is far more annoying.

Reply Score: 3

RE[3]: canary in the coalmine
by ssokolow on Wed 22nd Nov 2017 22:02 UTC in reply to "RE[2]: canary in the coalmine"
ssokolow Member since:
2010-01-21

Likewise (it's one of the reasons I have a mechanical keyboard) but, in my case, it's because, once Firefox 3.0 fixed the dropped characters problem, my Firefox has been incredibly prone to GC pauses due to the massive number of tabs and extensions I use.

(I'm hoping Quantum will fix that once I've finished migrating my extension needs to stuff that'll outlive 52ESR, but I worry because my mother's copy of Firefox Quantum seems to have reintroduced the "characters get dropped if you out-type the browser too badly" problem on her Facebook.)

Reply Score: 2

RE[4]: canary in the coalmine
by Alfman on Wed 22nd Nov 2017 23:07 UTC in reply to "RE[3]: canary in the coalmine"
Alfman Member since:
2011-01-28

ssokolow,

Likewise (it's one of the reasons I have a mechanical keyboard) but, in my case, it's because, once Firefox 3.0 fixed the dropped characters problem, my Firefox has been incredibly prone to GC pauses due to the massive number of tabs and extensions I use.


I never knew what caused that, I always assumed it was FF being slow because it was loading something, but it seems plausible that the delays may actually be caused by a garbage collection process.

(I'm hoping Quantum will fix that once I've finished migrating my extension needs to stuff that'll outlive 52ESR, but I worry because my mother's copy of Firefox Quantum seems to have reintroduced the "characters get dropped if you out-type the browser too badly" problem on her Facebook.)


My problem is with rampant memory usage and the bookmarks being painfully slow, these problems still remain. FF crashed once on the day it was updated and I had to kill it, but otherwise it's been working for me since then.

Edited 2017-11-22 23:08 UTC

Reply Score: 2

So....
by brostenen on Tue 21st Nov 2017 18:15 UTC
brostenen
Member since:
2007-01-16

You got problems with Windows. Tell me all about it, while I type this message, using Linux.

Reply Score: 0

RE: So....
by Bobthearch on Tue 21st Nov 2017 19:59 UTC in reply to "So...."
Bobthearch Member since:
2006-01-27

And then you'll get right back to real work using GIMP, OpenOffice, and QCad. LMAO!

Reply Score: 1

RE[2]: So....
by leech on Tue 21st Nov 2017 20:07 UTC in reply to "RE: So...."
leech Member since:
2006-01-10

I don't know... I mostly do all my work from Tilix...

Not everyone needs to do photo manipulation and document creation.

Besides, I honestly don't know why you couldn't do all document creation / photo manipulation with OpenOffice / Gimp.

Reply Score: 1

RE[3]: So....
by daedalus on Wed 22nd Nov 2017 09:03 UTC in reply to "RE[2]: So...."
daedalus Member since:
2011-01-14

You can of course create and modify documents with Open/Libre/whatever. What is less easy is sharing those documents with others, and reading documents from others. Sure, if they use standard fonts and don't go too crazy with the formatting you should be fine, but having to explain to a client who's using MS Office why they get blank pages in the procedure you sent them is surely not going to look great for your company.

Reply Score: 4

RE[4]: So....
by Lennie on Wed 22nd Nov 2017 17:05 UTC in reply to "RE[3]: So...."
Lennie Member since:
2007-09-22

I just send people a HTML email if I need more style than text email. I basically never use an office suite.

And really don't understand what people are doing with it.

Edited 2017-11-22 17:06 UTC

Reply Score: 1

RE[2]: So....
by Savior on Wed 22nd Nov 2017 10:05 UTC in reply to "RE: So...."
Savior Member since:
2006-09-02

The usual anti-Linux bullshit. Your only valid point might be OpenOffice (btw. it's called LibreOffice nowadays; that goes to show how well-informed the criticism is). As for the other two, very few people need heavy graphics editors, such as Photoshop, and even less CAD software.

I know that every "power user" (wannabe) out there pirates Photoshop onto his (usually pirated) Windows, but most of them would perfectly be fine with something like Paint.NET or Pinta. And GIMP is horrible, but Photoshop is not the epitome of usability, either.

Reply Score: 2

RE[3]: So....
by BluenoseJake on Wed 22nd Nov 2017 15:41 UTC in reply to "RE[2]: So...."
BluenoseJake Member since:
2005-08-11

OpenOffice still exists, LibreOffice is a fork of OpenOffice. LibreOffice is developed by the Document Foundation, and OpenOffice by Apache.

It hurts your arguments (some of which i agree with, btw) if you start off with an obvious (and easily research-able) mistake

Edited 2017-11-22 15:42 UTC

Reply Score: 3

RE[4]: So....
by ssokolow on Wed 22nd Nov 2017 21:58 UTC in reply to "RE[3]: So...."
ssokolow Member since:
2010-01-21

Look at the commit logs.

Forking LibreOffice away from the Sun/Oracle bureaucracy caused a massive spike in the commit rate and it's still got a healthy development pace.

OpenOffice is languishing so much that they have trouble releasing security patches in a prompt manner. (And don't take my word for it. The project's refusal to admit that it's practically unmaintained has been commented on multiple times in places like LWN.net)

Reply Score: 3

RE[5]: So....
by BluenoseJake on Thu 23rd Nov 2017 12:54 UTC in reply to "RE[4]: So...."
BluenoseJake Member since:
2005-08-11

That doesn't really matter. OpenOffice still exists, whatever mental gymnastics you want to undergo doesn't change that.

Reply Score: 1

RE[6]: So....
by ssokolow on Fri 24th Nov 2017 08:58 UTC in reply to "RE[5]: So...."
ssokolow Member since:
2010-01-21

For a not-very-useful definition of "still exists".

LibreOffice came about when more or less everyone involved in OpenOffice.org had either been reassigned or left the project because of how Oracle was mismanaging it.

LibreOffice is OpenOffice.org in everything but the name and the Sun CLA.

Apache OpenOffice is the name, a copy of the pre-LGPL code and little else.

Heck, the whole reason Oracle was convinced to give it to the Apache Foundation was because IBM had built Lotus Symphony as a closed-source product based on OpenOffice.org via a CLA-enabled licensing agreement with Sun that Oracle inherited and IBM wanted Oracle to try to stem the bleeding of developer effort to an LGPLed fork.

Edited 2017-11-24 09:02 UTC

Reply Score: 4

RE[3]: So....
by daedalus on Wed 22nd Nov 2017 15:42 UTC in reply to "RE[2]: So...."
daedalus Member since:
2011-01-14

Ummm... Nope. OpenOffice and LibreOffice are two separate entities these days. But regardless, the criticism stands for both forks.

Reply Score: 3

RE[2]: So....
by grat on Wed 22nd Nov 2017 19:41 UTC in reply to "RE: So...."
grat Member since:
2006-02-02

And then you'll get right back to real work using GIMP, OpenOffice, and QCad. LMAO!


I do occasionally use Gimp. I'm not a fan, but fortunately, there are a number of other editors out there, and it's not so bad for basic operations like converting, cropping, or adjusting levels.

For more sophisticated operations like photography there's programs like Darktable.

OpenOffice (well, LibreOffice) works just fine in allowing me to see and edit documents written by coworkers in Office, and I can send those edited documents back, and they can open them just fine, so yeah-- LibreOffice is a good replacement for MS Office.

I also run a local copy of davmail, which allows me to play nice with our Exchange server using standard IMAP, CalDav and LDAP clients. Akonadi (KDE) is a pain in the posterior, but KMail and KOrganizer work well for email and calendar.

I don't do CAD.

I did have to run a Windows VM under VirtualBox for VMWare Remote Console and the VMWare client, but the web version of the VMWare vSphere interface is pretty good now, and there's a linux Remote Console app, so it's pretty rare that I open the Windows VM anymore.

Reply Score: 3

RE: So....
by feamatar on Tue 21st Nov 2017 21:49 UTC in reply to "So...."
feamatar Member since:
2014-02-25

Is wayland out of beta by now?

Reply Score: 1

RE[2]: So....
by Bill Shooter of Bul on Wed 22nd Nov 2017 12:21 UTC in reply to "RE: So...."
Bill Shooter of Bul Member since:
2006-07-14

Not sure Beta was ever applied to Wayland, but its on by default in Fedora Gnome.

Not that it matters for composting lag. It should be the same kind of issue for compositors implementing Wayland, unless there is a way to turn of vsync.

Reply Score: 3

RE: So....
by The123king on Wed 22nd Nov 2017 08:17 UTC in reply to "So...."
The123king Member since:
2009-05-28

Windows might be an archaic, buggy, unintuitive sack of hopelessness, but i've never had it break an install during updates. With linux, it was a regular occurrence.

Reply Score: 0

xristos
Member since:
2014-04-25

Keyboard and mouse latency bug me a lot on Macs.

All the way back to the first release of OSX on my PowerBook Ti.

Even the mouse is slow (actually the mouse has felt slow even on MacOS 7-9, but the keyboard latency started with OSX)

I think if you're a long time Mac user it might not be noticeable. But for somebody like me who grew up using Windows and then switched to a Mac, it really is a big problem.

Windows10 may have introduced latency, but it still feels better than OSX.

Reply Score: 3

rener Member since:
2006-02-27

yeah, same here, especially notable some years ago when my first 4k display arrived and Apple had those 15" rMBP intensionally driver crippled not to do 60Hz MST on it. Running on 30Hz was painful and really laggy (Windows DID do 60Hz MST on that Nvidia 650-something). Much more than I think it should have been, Using it with 30Hz on Linux felt like twice as fast. Guess the mouse cursor is double or triple buffer on macOS. Maybe they should use the dedicated cursor overlay and update it directly and not part of the whole multi-buffered screen composition: https://rene.rebe.de/2014-01-27/dell-ultrasharp-24-4k-up2414q-on-mac...

Edited 2017-11-22 18:20 UTC

Reply Score: 2

Comment by ssokolow
by ssokolow on Tue 21st Nov 2017 21:21 UTC
ssokolow
Member since:
2010-01-21

During the first few releases where the KDE devs were saying "OK, NOW the problems are fixed", compositor latency was a big part of why KDE 4.x felt unacceptably heavier and more sluggish than KDE 3.5.x, but I didn't realize it before I gave up on trying every release to see if I could come back from LXDE.

Maybe a year ago, the lack of a monitor mount and LXPanel's non-support for non-rectangular desktops forced me to try KDE again and the latency is now no longer noticeable... though KWin still does require me to toggle compositing after it bugs out after a week or two of being logged in.

Ironically, I'm now preparing to migrate in the opposite direction to either KDE (if 5.x is acceptable by the time *buntu 14.04 LTS is no longer supported) or LXQt to flee the creeping influence of the GNOME 3.x HIG on my Lubuntu apps.

Reply Score: 3

RE: Comment by ssokolow
by grat on Wed 22nd Nov 2017 19:44 UTC in reply to "Comment by ssokolow"
grat Member since:
2006-02-02

Honestly, good drivers and config settings are more important than the desktop environment under Linux.

Reply Score: 2

Windows 7
by Alfman on Tue 21st Nov 2017 22:42 UTC
Alfman
Member since:
2011-01-28

Thom Holwerda,

If you're fine with unresponsive UI operations and graphical tearing, then, sure, go back to Windows 7 or earlier and turn off compositing to get a few ms back when typing.



I'm using windows 7 right now and have no idea what you are talking about. Are you referring to a specific application? Anyways, I've always disabled aero and set the win2k theme, I find it both more appealing and lighter on system resources!

Edited 2017-11-21 22:43 UTC

Reply Score: 5

Odwalla
Member since:
2006-02-01

The blogger admits his testing method is flawed and says the Windows 10 numbers he posted are likely worse than real-world. His testing machine utilizes a mobile CPU but he doesn't inform us as to the power-state of the CPU at the time of testing. The laptop used for testing also does not have a discrete graphics chip, instead relying on an embedded Intel HD Graphics 4400. Which isn't a powerhouse of a GPU at all.

With a sample size of one and demonstrable flaws in the data gathering and results there's nothing to be learned by this blogger's post.

Reply Score: 4

dpJudas Member since:
2009-12-10

With a sample size of one and demonstrable flaws in the data gathering and results there's nothing to be learned by this blogger's post.

It gets worse than this.

The guy is complaining about the DWM presentation delay. The average time is 8 ms in his test because on average there's about 8 ms to the next vsync (most LCD monitors updates at 60 hz, which is 16 ms).

The DWM is not new. It has been there since Windows Vista, and his "solution" for Windows 7 is to force the DWM to be off. In other words, he might as well have written a story about Windows Vista introducing a keyboard typing delay.

If someone truly is annoyed about this delay, they could solve the problem by buying a gaming monitor, aka. one of those running at 120 hz of 144 hz. That would bring his average delay down to 3-4 ms. If he claims he can still feel it after that he's full of BS.

The DWM has its share of problems - especially when apps miss the composition deadline. However, putting your Windows 7 into forced legacy mode is in my opinion not a proper solution to that. It will also leave you royally screwed when Windows 7 goes out of support.

Reply Score: 5

dpJudas Member since:
2009-12-10

I just realized his solution doesn't even work! When you turn the DWM off the video buffer is indeed updated immediately, but the monitor doesn't refresh until the next vsync anyway.

That means the average delay before the change meets your eyes is still 8 ms. All he gained was to get screen tearing. It also makes me somewhat skeptical that he can really feel the delay.

Reply Score: 4

Alfman Member since:
2011-01-28

dpJudas,

That means the average delay before the change meets your eyes is still 8 ms. All he gained was to get screen tearing. It also makes me somewhat skeptical that he can really feel the delay.


I think the point wasn't necessarily that the user would perceive small delays, but that when many such delays are aggregated they become perceptible. Therefor finding a way to reduce any of them could be helpful in reducing the overall lag.

Here's an interesting talk covering the perception of lag using different modes of input:

How Much Faster is Fast Enough? User Perception of Latency & Latency Improvements
https://www.youtube.com/watch?v=d3MaC_1zpMs

(see 11:56 for delay perception chart)

Applied Sciences Group: High Performance Touch
https://www.youtube.com/watch?v=vOvQCPLkPt4

Touch screen dragging lag is easily perceptible down to around 10ms because our visual senses can directly see the effects of latency, but our ability to perceive differences in temporal lag with keyboards and touchpads is different.

Edited 2017-11-22 03:10 UTC

Reply Score: 3

dpJudas Member since:
2009-12-10

I think the point wasn't necessarily that the user would perceive small delays, but that when many such delays are aggregated they become perceptible. Therefor finding a way to reduce any of them could be helpful in reducing the overall lag.

You might be able to convince me you'd notice the DWM delay with mouse and touch operations, but there's just no way I think anyone notices it with the keyboard.

In any case, if its bad with keyboard it should be ten times worse with the mouse. I haven't heard of anyone complaining loudly about mouse lag since Windows Vista.

Reply Score: 3

dnebdal Member since:
2008-08-27

You might be able to convince me you'd notice the DWM delay with mouse and touch operations, but there's just no way I think anyone notices it with the keyboard.

In any case, if its bad with keyboard it should be ten times worse with the mouse. I haven't heard of anyone complaining loudly about mouse lag since Windows Vista.

He does explicitly address that: The mouse cursor is handled entirely outside compositing; it's got its own little rendering path. (It's just a bitmap put on top of everything else, and all windows needs to do to move it is to send new coordinates to the drivers/card.)

Edited 2017-11-22 11:31 UTC

Reply Score: 3

dpJudas Member since:
2009-12-10

He does explicitly address that: The mouse cursor is handled entirely outside compositing; it's got its own little rendering path. (It's just a bitmap put on top of everything else, and all windows needs to do to move it is to send new coordinates to the drivers/card.)

Games running in Windowed mode is affected. Drawing program are also affected. Dragging a window is affected. None of those things use such a rendering path.

Maybe I'm just blind, but I can't notice any mouse lag when I grab and drag the window on this Windows 10 machine.

Reply Score: 3

Alfman Member since:
2011-01-28

dpJudas,

You might be able to convince me you'd notice the DWM delay with mouse and touch operations, but there's just no way I think anyone notices it with the keyboard.


I didn't say I did notice the DWM delay, but if several subsystems are adding delay, then the point is that all the lags add up and one could potentially notice the aggregate delay. I think the empirical numbers from the video I linked to in my last post give a good idea of the lowest total latency we can perceive. Once you begin to notice the lag, it doesn't matter where it comes from, optimizing any of the sources of lag can help even if it's a small amount on it's own.



In any case, if its bad with keyboard it should be ten times worse with the mouse. I haven't heard of anyone complaining loudly about mouse lag since Windows Vista.


Windows has a fast code path for cursor updates, I do not perceive lag. I suspect it may be entirely interrupt driven so it's fast even under heavy system load.

With keyboard input, the lag is usually below my perception threshold except in relatively rare circumstances. For example, typing a URL in firefox will sometimes lag enough for keys to buffer up while firefox is looking something up. Interactive office applications can sometimes lag under system load. This could be addressed by prioritizing interactive event loops and making sure all blocking code paths are taken out of the UI event loops. But most developers are lazy and don't bother with any of this, it's much easier to blame the background processes.

I use notepad2 as a system wide improvement over ms-notepad. The one thing I find frustrating at times is when I edit remote files, the entire UI will lock up if there's a network issue because the file IO happens in the GUI event loop.


With some low fps games where mouse events are only scheduled to run in a relatively slow event loop, I've found that not only can I perceive mouse lag, but it can become mildly annoying. I'm not an avid gamer, but I don't suppose its a problem for modern 60+fps games.


Ultimately, like you I don't think interaction lag is generally a problem for most PC users, however it does seem to be a bigger problem for touchscreens.

Edited 2017-11-22 15:21 UTC

Reply Score: 3

dpJudas Member since:
2009-12-10

I didn't say I did notice the DWM delay, but if several subsystems are adding delay, then the point is that all the lags add up and one could potentially notice the aggregate delay.

There are not several subsystems adding delay to what he is complaining about. There's only the DWM composition to presentation delay. That delay is a constant, which is some fixed value between 0 to 16 ms depending on exactly when the DWM does its composition for the next presentation.

I bring the mouse into the discussion because there really is no need to point at studies here. Open your eyes, grab the titlebar of any given window, and start dragging it around. If you can't even notice that lag then there is just no way you'd notice it with the keyboard.

I think the empirical numbers from the video I linked to in my last post give a good idea of the lowest total latency we can perceive. Once you begin to notice the lag, it doesn't matter where it comes from, optimizing any of the sources of lag can help even if it's a small amount on it's own.

If the lag is increased by other sources, then I'd say you should probably consider fixing those than jump back to an earlier version of Windows. But let's look at the facts here - his test program doesn't measure anything else than the average delay to the next DWM presentation. My comments are based on his test program and nothing else.

He didn't say "my text editor is so shitty at updating that I chose to go back to an earlier version of Windows as a solution". He more or less said Windows itself forced him to go back.

With keyboard input, the lag is usually below my perception threshold except in relatively rare circumstances. For example, typing a URL in firefox will sometimes lag enough for keys to buffer up while firefox is looking something up. Interactive office applications can sometimes lag under system load.

He was directly blaming the DWM. Thom also mentioned the OS, so he apparently read it roughly in the same way. If we are discussing "slowmo app updates screen really slow and I can notice that sometimes", then yes, I agree, I can notice if an app is slow too.

Ultimately, like you I don't think interaction lag is generally a problem for most PC users, however it does seem to be a bigger problem for touchscreens.

I don't know enough about touchscreens to comment on that. But he wasn't talking about touch input (not the same as keyboard input).

Reply Score: 4

Alfman Member since:
2011-01-28

dpJudas,


I bring the mouse into the discussion because there really is no need to point at studies here. Open your eyes, grab the titlebar of any given window, and start dragging it around. If you can't even notice that lag then there is just no way you'd notice it with the keyboard.


Ah, but I cited the data specifically so that we could speak about latency perception more objectively.

I don't experience latency on this win7 system, but never the less it would be a fallacy to declare there's no problem just because we don't experience it in our tiny sample.

If you look carefully some of the windows 10 points end up being more than one frame behind the windows 7 ones, so win 10 may be doing something that adds more overhead than desktop composition should require. Does it matter to user perception? Obviously that depends on how much latency there is in the whole system. It may make a difference in high latency applications.



I don't know enough about touchscreens to comment on that. But he wasn't talking about touch input (not the same as keyboard input).



So... you didn't even look at the links then? The microsoft link especially clearly shows the effects of touch latency down to around 10ms.

And I'm not saying it's the same thing as keyboard latency (although the other researcher rigorously measured this as well), just thought other data points would be interesting for the discussion ;)

Edited 2017-11-23 08:34 UTC

Reply Score: 3

To be fair...
by Poseidon on Wed 22nd Nov 2017 06:41 UTC
Poseidon
Member since:
2009-10-31

To be fair, the only OS that doesn't have that issue is perhaps macOS, but apple made sure to invest on the graphical backend when going into it. On the downside of it though, it requires quite a bit of power, so older machines get graphically taxed to keep up sometimes.

KDE 3 was really good on this as well, but 4 changed that a bit, and I haven't used it recently to know if that's the case anymore. OpenGL accelerated rendering help quite a bit though, if you had the power.

Reply Score: 1

RE: To be fair...
by Carewolf on Wed 22nd Nov 2017 16:19 UTC in reply to "To be fair..."
Carewolf Member since:
2005-09-08

Macs have it worse than any desktop. It is not caused by lack of performance, the lag is an intentional design decision. It makes things feel more smooth by making transitions everywhere. This is something Macs pioneered and why they feel so slow and sluggish to Windows or Linux users regardless of hardware.

Reply Score: 3

Not sure it is really decompositing
by ThomasFuhringer on Wed 22nd Nov 2017 09:00 UTC
ThomasFuhringer
Member since:
2007-01-25

All I know is that GUIs were already fluent and snappy ten years ago on that day's standard hardware.
Now we have a multiple of computing resources on our desktop and have issues with rendering the mouse cursor. What a mess!
It must be something else. I suspect layers of 'managed code' somewhere underneath. - Or simply BAD code by underqualified, sloppy programmers.

Reply Score: 0

fix...
by sarreq on Wed 22nd Nov 2017 14:04 UTC
sarreq
Member since:
2010-03-14

more ram, and disable the page file

Reply Score: 0

Nope
by birdie on Wed 22nd Nov 2017 16:36 UTC
birdie
Member since:
2014-07-15

> unresponsive UI operations

Never noticed that about Windows 7. Must be a problem with your setup, Thom.

Reply Score: 3

RE: Nope
by Megol on Thu 23rd Nov 2017 12:24 UTC in reply to "Nope"
Megol Member since:
2011-04-11

Or him talking out of his ass - again.

It's not even like he try to generate traffic by intentionally trolling like some other sites, just writing things that doesn't make sense and aren't true.

Reply Score: 0