From: james.harris.1@gmail.com   
      
   On 23/10/2021 10:21, Rod Pemberton wrote:   
   > On Thu, 21 Oct 2021 21:02:45 +0100   
   > James Harris  wrote:   
   >   
   >> You guys might be interested in this. The guy claims it as the   
   >> world's fastest-booting PC. That may be something of an overclaim but   
   >> this is the fastest I've ever seen a PC boot - about 2.5 seconds to   
   >> MS-DOS 3.3. And that's on an 8086!   
   >>   
   >>     https://youtu.be/DaSkda4XW3k   
   >>   
   >   
   > Didn't you bring up this issue previously?  Or, was that someone else?   
      
   No, I don't recall seeing that video before.   
      
   >   
   > Well, that is for an early small BIOS, with a very small text-based OS,   
   > probably without any device drivers, with a relatively fast for the era   
   > 8086.  Yes?  And, it really didn't seem like it was loaded from a   
   > floppy (real slow) or MFM hard disk (slow).  So, I suspect it would've   
   > been much slower if it was.  Didn't he say something about a ROM drive?   
      
   IIRC the PC booted from ROM.   
      
   ...   
      
   > So, your definition of "boot" doesn't only include executing the BIOS   
   > and transferring execution to the boot loader?  I.e., you define "boot"   
   > as including executing the BIOS/UEFI, executing the boot loader (etc)   
   > too, and executing a small text OS too, but not executing a large GUI   
   > OS.   
      
   I didn't try to define 'boot'. You could give it different legitimate   
   meanings, e.g. to include firmware time or not and there could be a   
   debate about when a machine has finished booting.   
      
   For the latter, I'd say a machine has booted when it can be interacted   
   with normally by a user or by a remote program. By 'normally' I means as   
   long as no one experiences particular delays waiting for a component   
   part to start working.   
      
      
   > I agree that booting/loading could be faster, but then you'd need to do   
   > some sort of save-state of the OS, like certain OSes do when the   
   > computer "sleeps" or "hibernates" or does a "quick shutdown".  You'd   
   > also need some way to quickly reset the hardware to the correct states   
   > to match that of the OS save-state.  And, if the OS boots fast, because   
   > the machine is fast, what's the point?   
      
   That seems unlikely. Given that Wolfgang's OS boots in about 2 seconds   
   from when it gets control I doubt hardware discovery needs to be slow,   
   albeit that Wolfgang tends to use known hardware.   
      
      
   --   
   James Harris   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)