Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Saturday December 30 2017, @06:45PM   Printer-friendly
from the perhaps-providing-prompt-prompts-prompts-perceived-performance-primacy dept.

Have you ever had that nagging sensation that your computer was slower than it used to be? Or that your brand new laptop seemed much more sluggish than an old tower PC you once had? Dan Luu, a computer engineer who has previously worked at Google and Microsoft, had the same sensation, so he did what the rest of us would not: He decided to test a whole slew of computational devices ranging from desktops built in 1977 to computers and tablets built this year. And he learned that that nagging sensation was spot on—over the last 30 years, computers have actually gotten slower in one particular way.

Not computationally speaking, of course. Modern computers are capable of complex calculations that would be impossible for the earliest processors of the personal computing age. The Apple IIe, which ended up being the “fastest” desktop/laptop computer Luu tested, is capable of performing just 0.43 million instructions per second (MIPS) with its MOS 6502 processor. The Intel i7-7700k, found in the most powerful computer Luu tested, is capable of over 27,000 MIPS.

But Luu wasn’t testing how fast a computer processes complex data sets. Luu was interested in testing how the responsiveness of computers to human interaction had changed over the last three decades, and in that case, the Apple IIe is significantly faster than any modern computer.

https://gizmodo.com/the-one-way-your-laptop-is-actually-slower-than-a-30-ye-1821608743


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 1, Interesting) by Anonymous Coward on Saturday December 30 2017, @08:39PM (7 children)

    by Anonymous Coward on Saturday December 30 2017, @08:39PM (#615968)

    The most beautiful desktop environment and most responsive computer I have ever seen were the same. I saw Ian Finder (he deserves credit for this) running Mac OS9 (emulated) on a mac book air (~2014 I think). Hand crafted pixel perfect graphics on a very nice display with incredible responsiveness in a tiny form factor. Even running emulated, it was still faster than on the hardware from the OS9 era. He was doing his homework in CodeWarrior.

    Looking at the article, it seems like a lot of the lag is in the keyboard (he measured from when the key started down until the character was displayed in a terminal). I wonder if the results would be even more extreme if he only measured the OS part of the stack (ignoring display and keyboard latency).

    Starting Score:    0  points
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  

    Total Score:   1  
  • (Score: 4, Interesting) by Anonymous Coward on Saturday December 30 2017, @10:39PM (5 children)

    by Anonymous Coward on Saturday December 30 2017, @10:39PM (#615993)

    I didn't see what keyboard the tester was using, but I'm guessing it's some flavor of USB or blue tooth.

    Try plugging in a good ols PS/2 keyboard if your mobo still supports it and watch the latency go away.

    Real gamers use PS/2 keyboards and mice.

    • (Score: 2, Interesting) by Anonymous Coward on Sunday December 31 2017, @03:00AM (2 children)

      by Anonymous Coward on Sunday December 31 2017, @03:00AM (#616049)

      Some people say USB keyboards can't do more than 6 keys at a time... wrong. That is USB in BIOS compat mode. USB HID supports N-key rollover fine. Look for keyboards that show two interface descriptors, one of them "huge" to fit a full key mask (for example "bInterfaceSubClass 1 Boot Interface Subclass", "wMaxPacketSize 0x0008 1x 8 bytes" and "bInterfaceSubClass 0 No Subclass", "wMaxPacketSize 0x0040 1x 64 bytes" as reported by lsusb).

      Some people say USB keyboards are 1.5Mbps limited... wrong again. You can get 12Mbps ones, the above example comes from one of those.

      Some people say USB keyboards can't update fast enough, because polling is slow... wrong for a third time. They are right USB keyboards work via polling, but the speed can be high, for example 1ms interval ("bInterval 1"). That's 1000Hz refresh, 1000 times a second the keyboard gets asked. The controller in the computer should be able to queue all updates, and not disturb the main CPU and the OS for nothing. So far I never had missed keys.

      Old crappy USB mouse uses 1.5Mbps and 10ms. Probably getting a new mouse matching the keyboard will give the "high" speed values. This example mouse is from first waves of opticals, capable of speaking PS2 with dumb adapter.

      PS2 support in main boards is still there, but sometimes with bugs. So it can be tricky, and just hunting for good USB devices the solution.

      PS: some info also avaliable with "mount -t debugfs none /sys/kernel/debug/" then "less /sys/kernel/debug/usb/devices" (look for fields like Spd=12, MxPS=64 and Ivl=1ms).

      • (Score: 0) by Anonymous Coward on Sunday December 31 2017, @11:32AM (1 child)

        by Anonymous Coward on Sunday December 31 2017, @11:32AM (#616115)

        Why would low-speed vs. full-speed possibly matter for a keyboard? Can someone type a million characters per second? Latency isn't different by more than a few microseconds.

        Only difference is probably just the version of the protocol implemented by the interface chip, and has no impact on performance.

        • (Score: 0) by Anonymous Coward on Monday January 01 2018, @02:45AM

          by Anonymous Coward on Monday January 01 2018, @02:45AM (#616315)

          Yes, if the keyboard is polled every 10ms, it means that key that was hit 0.5ms since last report, has to wait 0.5 ms in the 1 ms rate and 9.5 ms in the 10 ms rate (milliseconds, 1/1000 of second, not microseconds) to be noticed. A 60Hz screen refreshes every 16 ms, for comparison.

          So in one case the keyboard could report 16 times per frame, while in the other it reports twice at best, and in some cases only once (think about the out of sync pattern 0, 10, 20, 30, 40 for keyboard but 0, 16, 32, 48 for screen). That is jitter, and will make things worse, as delay varies and your worst case is not very good as they are both very similar so sometimes it's too late, and never fast. The rendering will have less time to react or will have to give up until next frame in a repetitive yet weird pattern.

    • (Score: 0) by Anonymous Coward on Sunday December 31 2017, @11:36AM

      by Anonymous Coward on Sunday December 31 2017, @11:36AM (#616117)

      Best solution to this would be to implement a virtual keyboard in suitable hardware (an Arduino for example). Then you can send a real start signal on a GPIO pin concurrently with the keypress being sent to the host. Of course if you use the same keyboard for all computers, the relative effect will be the same, but it won't hide differences between PS/2 and USB interface. On the other hand, real-world latency does include a keyboard.

    • (Score: 0) by Anonymous Coward on Sunday December 31 2017, @08:21PM

      by Anonymous Coward on Sunday December 31 2017, @08:21PM (#616199)

      Try plugging in a good ols PS/2 keyboard if your mobo still supports it and watch the latency go away.

      Real gamers use PS/2 keyboards and mice.

      Wrong. PS/2 devices can have high latency too. I've compared PS/2 keyboards with other keyboards/mice- use a ruler or similar to press keys/buttons at about the same time and measure the difference in the time. Repeat the tests and even though you don't get the absolute latency you can compare the relative latency between two devices.

      See also:
      http://www.blackboxtoolkit.com/responsedevices.html [blackboxtoolkit.com]
      http://www2.pstnet.com/eprimedevice.cfm [pstnet.com]

      And: https://danluu.com/keyboard-latency/ [danluu.com]

  • (Score: 1, Interesting) by Anonymous Coward on Sunday December 31 2017, @12:52AM

    by Anonymous Coward on Sunday December 31 2017, @12:52AM (#616029)

    Not beautiful, but really fast is my ancient FinalWord II word processing software running on a Win 7 laptop under the DOSbox emulator. Later versions were called Borland Sprint wordprocessor.
        Scroll top to bottom of medium sized docs, 100 pages, 50 lines on a "VGA" screen page, is effectively instant (at the repeat rate of the PgDn key). Jump from home to end is instant.
        Same for search, effectively instant, along with search & replace.
        FWII was written to use a swap file so if DOS crashed, all but the most recent key strokes could be recovered. Now it's so fast that I don't lose much of anything. Used to be the "Swapping..." legend would display for a second every now and then. Now it's rare that I ever see it, it's not up long enough to be displayed.

    While I don't use this emulation very often, it's convenient when I need to look at the source files for a number of long docs that I wrote on CP/M and MS-DOS systems.