Why iOS feels faster than Android, and why computers don’t feel as fast as they used to

I wanted to share a short but interesting post– with data– about input lag, latency and complexity. There’s something here for both young and old users but the tl;dr is that if you seem to recall computers of yore feeling snappier, you’re not crazy– they probably were! It’s the same story for iOS vs, Android.

Excerpted from the original post: “Fancy gaming rigs with unusually high refresh-rate displays are almost competitive with machines from the late 70s and early 80s, but “normal” modern computers can’t compete with thirty to forty year old machines.”

I encourage folks to read the original article and take a look at the latency tables for various devices. https://danluu.com/input-lag/

Conversation 26 comments

  • ErichK

    Premium Member
    26 February, 2018 - 7:30 pm

    <p>I guess computers today are "faster" but we make them do <strong><em>more</em></strong>, so it's a never ending cycle.</p><p><br></p><p>Who among us hasn't had the experience of sitting in front of an under-powered PC, loaded Word, and waited a few seconds before the first characters we typed appeared on the screen. I would imagine there's probably thousands of lines of code those characters get passed through before they ever even see the light of day.</p><p><br></p><p>But I do remember when I first typed "dir" at the DOS command line on my new 486 in 1993 … boy did that sucker fly! ?</p>

  • wright_is

    Premium Member
    01 March, 2018 - 12:06 am

    <p>Having used PCs from the early 80s on, I would say that they are definitely a lot faster now, even the feeling.</p><p>Go back and boot an original IBM PC from floppy disk… It takes several minutes to boot! Loading WordStar took longer than loading Word today.</p><p>Typing on an original IBM PC in WordPerfect, it was possible to overfill the buffer, waiting for it to catch up at times, so that would be a couple of seconds delay between keypress and the text appearing on the screen.</p>

  • jules_wombat

    26 December, 2018 - 9:17 am

    <p>Yep the frustrations of modern computing. Simply creating a Word document, as it attempts to connect to services and profiles, before opening up. Even Linux Ubuntu is pretty flaky compared to TRS-80 machines. But a lot more capable I guess. </p>

  • madthinus

    Premium Member
    26 December, 2018 - 11:09 am

    <p>With added power comes more complexity. As we have added layers on layers of abstraction, the lag has increased. I am sure if we start from scratch today we could do better at the cost of losing all the software that works. . Once a decade Microsoft mentions Input lag and that they are trying to reduce it. Nvidia, Oculus and the fine VR folk is the latest of input lag warrior trying to reduce it. We will see if they get it done. </p>

  • jchampeau

    Premium Member
    26 December, 2018 - 12:36 pm

    <p>I was thinking about this in a slightly different context just a few days ago. The integrated climate control/infotainment system in my 2016 Jeep Grand Cherokee is the laggiest, buggiest, most fickle hunk of crap I've ever seen in a vehicle. The touchscreen is so slow to respond sometimes it's distracting, which is exactly what you <em>don't</em> want in a vehicle. The built-in navigation is terrible, the layout is awful (it takes too many screens/taps to get anything done), the integrated voice control is difficult to use, the app-based remote start that uses Sprint's network fails half the time, cover art only displays half the time, etc., etc., etc. Even if I ignore the touchscreen and just hit the "warmer/cooler" temperature control rocker switch, it's slow to show the new setting on the screen. If I hit it five times rapidly, all five pushes might register or maybe just one or two will. So frustrating. My last vehicle–a 2010 Silverado pickup–had the Bose speaker system but a simple green LED-based head unit as opposed to the color touchscreen option, and the climate control was all separate. Oh, how I miss the simpler interface that was responsive and just worked.</p>

    • Daekar

      01 January, 2019 - 11:25 am

      <blockquote><em><a href="#387268">In reply to jchampeau:</a></em></blockquote><p>Amen, brother. This is one of the reasons I'm keeping my old car – I do NOT want all this complicated nonsense in my car when it adds zero utility.</p>

    • AnOldAmigaUser

      Premium Member
      01 January, 2019 - 2:58 pm

      <blockquote><em><a href="#387268">In reply to jchampeau:</a></em></blockquote><p>I do not understand why car makers insist on putting these systems in the car, rather than providing software that can be used with a phone or tablet. The average age of a car on the road today, I believe, is over 10 years, which is an eon in technology. The systems are bound to be obsolete in 3 years, if they worked in the first place. If they provide a simple interface and a docking location for a phone/tablet, the software can be kept up to date.</p><p>My pet peeve is cylinder deactivation and stop/start systems. The first typically causes problems with the engine mounts because of the uneven stresses applied, and the second I just do not trust to not leave me screwed in the middle of an intersection.</p><p><br></p>

    • IanYates82

      Premium Member
      01 January, 2019 - 9:44 pm

      <blockquote><em><a href="#387268">In reply to jchampeau:</a></em></blockquote><p>Agreed.</p><p><br></p><p>The Bluetooth connection screen in my Hyundai is so irritating, it can't possibly have ever been used by the person responsible for the software. Lots of modal dialogs with "please wait!" in them. </p><p>Also really silly things like asking "Am I sure" in a popup after choosing a Bluetooth device, but the popup is so big it obscures which device I tapped on, and the popup itself doesn't include the device name. So the "Am I sure" prompt is really useless since it does nothing more than require me to always say yes. It does this even when I don't have an active Bluetooth connection and have explicitly opened up the Bluetooth devices list to activate such a connection. There's clear user intent here, so of course I am sure!</p><p><br></p><p>Grrr :)</p>

    • minke

      18 March, 2019 - 8:00 am

      <blockquote><em><a href="#387268">In reply to jchampeau:</a></em></blockquote><p>Give me some nice buttons, switches, and dials I can feel and adjust without looking while driving and I am much happier. When you have to get out the car manual, and read it for 20 minutes, in order to drive a borrowed or rental car, you know we are taking a step back.</p>

  • jimchamplin

    Premium Member
    26 December, 2018 - 11:30 pm

    <p>Every now and then I like to install and run Debian Linux or FreeBSD on real hardware either with no GUI, or I compile CDE from the 90s. It's absolutely amazing to run fully command-line from SSD on a machine that's got 12GB of RAM and a quad Xeon CPU.</p><p><br></p><p>FreeBSD usin' like 90MB of RAM.</p><p><br></p><p>Shit is literally instant.</p>

  • longhorn

    01 January, 2019 - 12:10 pm

    <p>Spectre, Meltdown and Foreshadow software fixes are not making computers faster. Kernel performance (doesn't matter which OS) goes down 30 % in some scenarios. Does anyone know when Intel will release a fully fixed architecture?</p><p>Maybe AMD will get there first as AMD was less affected.</p>

  • Winner

    01 January, 2019 - 4:07 pm

    <p>My work PC has latest Win10, plenty of RAM, SSD, plus latest MS Office.</p><p>It is as slow as molasses to open an existing Powerpoint file, I swear much slower than it was on Win 7 with a spinning drive.</p>

    • IanYates82

      Premium Member
      01 January, 2019 - 9:41 pm

      <blockquote><em><a href="#389652">In reply to Winner:</a></em></blockquote><p>Anti-virus… They keep adding more bells and whistles to AV products – things like "smart heuristics" and "machine learning anti ransomware", etc. Slows things down a LOT.</p>

      • Daekar

        02 January, 2019 - 11:56 am

        <blockquote><em><a href="#389703">In reply to IanYates82:</a></em></blockquote><p>Another excellent reason not to run anything except the built-in Windows solution.</p>

    • wright_is

      Premium Member
      02 January, 2019 - 12:49 am

      <blockquote><em><a href="#389652">In reply to Winner:</a></em></blockquote><p>On my PC, a 77MB, 38 slide presentation opens in around 1.5 seconds.</p><p>An 80 slide presentation with lots of linked video opens in around 1 second.</p><p>What sort of presentation are you trying to open?</p>

  • wright_is

    Premium Member
    02 January, 2019 - 12:45 am

    <p>They feel faster, to me. </p><p>Having worked on computers since the early 1980s, I have to say that a modern PC feels a lot faster than the old stuff I used to use and program.</p><p>For example, I worked on a financial reporting system that packed up the local accounts and sent them to HQ back at the end of the 80s. That took around 4 hours to work through the accounts and exporting them to a .csv file for transer when I started, I made a couple of minor changes, one being only displaying every 100th account, instead of every account being processed and the processing time dropped to 30 minutes! (HP 125 CP/M and HP 150 MS-DOS machines)</p><p>Roll forward to the mid 90s, a similar financial budgeting program was packing up the data and sending it back to the HQ in around 2 minutes. (P90 based desktop)</p><p>These days the packing would take a couple of seconds. (Core i5 laptop)</p><p>So, it depends on what you mean. Latency to display a character on the screen is a meaningless statistic, the PC is there to actually do something and it can do a lot more somethings in the same amount of time, compared to the past.</p>

    • Greg Green

      09 February, 2019 - 9:47 am

      <blockquote><em><a href="#389716">In reply to wright_is:</a></em></blockquote><p>I think for large data work there clearly improvements. But regular day to day tasks of opening small files, webpages, programs, all seem to take longer.</p>

  • Daekar

    02 January, 2019 - 11:56 am

    <p>I appreciate the drastic increase in computing power, but I hate that the rate at which we have piled additional stuff on the system meets or exceeds the rate of increased computing power. Sometimes it's called for, and sometimes I just want to smack the designer that decided, "yeah, we need layers of animation, AI, and a mandatory network connection for that functionality that doesn't use remote resources because reasons." No, you don't. You <em>so don't</em>.</p>

    • Greg Green

      09 February, 2019 - 9:46 am

      <blockquote><em><a href="#389879">In reply to Daekar:</a></em></blockquote><p>I remember someone a long time ago saying about performance Intel giveth and Microsoft taketh away. Though that’s not quite so accurate anymore since Intel is making only incremental improvements.</p>

  • Thom77

    02 January, 2019 - 3:32 pm

    <p>I sometimes believe there is a conspiracy to bloat software to make hardware obsolete to force you to buy new hardware. I also sometimes suspect that there is outright nerfing of operating systems speed capabilities after a certain amount of time. My Ipad Air slowly started becoming unusable after each update. I think 10.0 was the big one where i was like WTF. I've had a mid range Samsung phone become sluggish which it wasnt when I first bought it, and no resets fixed it. In the case of the Ipad, I know about the aging battery issue causing performance decreases to save battery … but what NOBODY mentioned, is that the performance was the same WHEN PLUGGED IN, which is the way I use my Ipad 75% of the time. I know the system updates have new features … but I can't imagine why it would bog down your system when your not using any of the "new" features … unless … they were designed to bog down you system regardless.</p><p><br></p><p>I recently bought a low end Samsung J3 Luna Pro to carry me over for a while when my Iphone SE broke. There is no reason why even a $60 phone should be this sluggish in 2018. Then I realized there IS a reason why it should be this sluggish….. so I will buy the more expensive phones.</p><p><br></p><p>When I was younger, I remember the 486 upgrade we got was noticeably faster then the 386 we had previously. Truth be told, most casual users of computers will barely notice any difference now between each variation of (insert name) Lake chipsets. </p><p><br></p><p>I sincerely believe that chip manufactures and computer manufactures know that there is very little difference for most consumers between the new version tech vs last years. And considering the power in computing today and the small increases in power year to year, there is little reason to upgrade …. unless …. artificial manipulation occurs which promotes the illusion that the "new" one is far superior to the one you already have.</p><p><br></p><p>On a side note, since I have my tin foil cap on, I also have wondered if Intel develops a chip with Power D from the beginning, and then produces 4 variations of the same chip in the span of 4 years with ever decreasing nerfs (Chip A, B, C .. then D which always existed from the beginning).Its not too outlandish … I remember when the 720kb "floppy" disk and the 1.44mb variation was LITERALLY the same product, with only the notch in the side directing the system to recognize it as one or the other. </p>

    • Xatom

      04 January, 2019 - 12:06 am

      <blockquote><em><a href="#389986">In reply to Thom77:</a></em></blockquote><p>Some keen observations. I have 2 identical ipads one personal one business. I never update any os ( save Win 10 over which i have no control). My business ipad is required to be updated. To your point the difference in speed between the one with the out of the box os vs the 'upgraded' business one is vast. </p><p><br></p><p>In the early days feature additions and hardware speed created a compelling need to upgrade. Now there is literally no reason to upgrade. There hasn't been serious desktop software innovation in a decades other than pure feature blot that no one wants or needs. All this is an attempt at planned obsolescence. Apple battery-gate tells us all we need to know. </p><p><br></p><p>Apple's phone sales problem is the result of this. Fine products-old os- no reason to upgrade until the phone breaks. And they are being watched carefully for any attempt to take advantage of the installed base by the cult. Both Apple and other mobile phone makers wont recover from products that are at the end of their innovation cycle. Just have to wait for foldables and real 5g to drive the next wave.</p>

      • Bill Strong

        04 April, 2019 - 1:41 pm

        <blockquote><em><a href="#390688">In reply to Xatom:</a></em></blockquote><p>It is not exactly true that there has been no innovations in desktop software. It might be considered true in the case of consumer consumption software, but professional and gaming innovations continue on. Especially in the 3D space, Blender is now using Game engine style tech to update its view-port in real-time for a more WYSIWYG experience. Allegoritmic's, now Adobe's, Substance Designer and Substance Painter are highly innovative ways to view and create artistic and realistic 3D objects. TV programs are being made in Game Engines such as Unreal, rather than using huge render farms. </p><p><br></p><p>Even Databases are changing, since crypto currencies are essentially a distributed database tech. We have a group changing to reinvent the internet aver at IPFS. We have VR demos that show the power VR could have to create 3D objects, landscapes and real life structures. I am surprised I haven't seen any cyberpunk HUDs yet, but I am sure I will.</p><p><br></p><p> Programming languages also innovate on a regular basis, such as Rust and Red-Lang, as well as academic ventures such as a mathematically proven safe kernel, seL4, and a mathematically proven https stack now used in Windows and Firefox, EverCrypt.</p><p><br></p><p>Its only the consumption side of the Desktop experience that hasn't innovated, which is why its lunch was so easily eaten by iOS and Android.</p>

  • siv

    02 January, 2019 - 6:49 pm

    <p>You have to remember that back in the days of Apple 2 the machine was very simple and I think not doing multi-tasking, so clearly it would seem snappier, I could never afford Apple kit, but had Sinclair ZX81, Spectrum, QL, then switched to Atari kit 520 and eventually a Mega ST4 and even in that period the machines were still pretty simple. The ST like the Apple Macs of the time didn't really do multi-tasking you could run TSR's but they weren't really multi-tasking they just froze the main app whilst you accessed the accessory and then resumed when you finished with the accessory. The Amiga was probably one of the first PC like machines to start multi-tasking. </p><p><br></p><p>Also in the early days particularly in the PC space the MHz rises had a direct impact on performance and each new 386 to 486 to Pentium and beyond pushed the clock speed up dramatically. This you noticed as everything did run faster. </p><p><br></p><p>Eventually we hit the limit of silicon at around 4 GHz and then the only way to extract extra performance was by multi-tasking operating systems and increasing the cores in the CPU which is where we are now. So you won't seem to feel the snappiness any more because for each individual task they are not getting any faster most PC CPUs are now in the 2.5 GHz to 4.0 GHz range and most reasonably priced laptops and PCs are more towards the 2.5 GHz range. </p><p><br></p><p>Where you do feel it as others mentioned below is in the ability to achieve more in the same time as you can theoretically hand off parts of a calculation or some other processing to separate core in the CPU so the work is carrying on in parallel rather than in series as it was back in the early days.</p><p><br></p><p>Regarding the latency between a key press and it appearing on screen, again the original article does partially explain it, but I think with Windows the way multi-tasking is working can affect this, particularly if you have an anti-virus program scanning in the background or Windows is checking for updates. Also a lot of the patches we have applied to mitigate some of these processor look ahead malware has made our machines seem a lot slower, my ASRock Z97 motherboard definitely slowed when I applied the patches for Meltdown and Spectre to its firmware.</p><p><br></p><p>It's all a compromise.</p><p><br></p><p><br></p>

  • minke

    18 March, 2019 - 7:57 am

    <p>This is so true. At work I just recently got the latest, greatest iMac with the upgraded processor, etc., and it is so laggy compared to my 5-year old PC at home running Ubuntu. Just opening and closing basic programs like Word or Excel is slow and triggers unnecessary dialog boxes in many cases. I really like online office suites too, but then you become dependent on the speed and quality of the Internet connection, which can make everything drag. But, compare this to the old dial-up modem days and the speed is rather amazing.</p>

    • StoneJack

      18 March, 2019 - 8:49 am

      <blockquote><em><a href="#413352">In reply to Minke:</a></em></blockquote><p>It is because of MS office, not the Mac.</p>

      • minke

        04 April, 2019 - 7:14 am

        <blockquote><em><a href="#413357">In reply to StoneJack:</a></em></blockquote><p>OK, then it's the hardware and software together that make everything a drag. What really ticks me off though are the obvious problems that just go on and on and on and seem to never get solved, like OneDrive lack of syncing reliably in a timely fashion, or the continued use of username and password login like we were doing it 20 years ago except now we must use a password that is both impossible to remember and type so we have to first login to the password manager, which for some reason won't do autofill so you have to find the right entry then copy and paste using control c + control v. Some things just seem to always get worse rather than better.</p>

Windows Intelligence In Your Inbox

Sign up for our new free newsletter to get three time-saving tips each Friday

"*" indicates required fields

This field is for validation purposes and should be left unchanged.

Thurrott © 2024 Thurrott LLC