Instant PC Boot - Why not, really?

Sigh,Where should I begin.

How about with this: Flash memory and Bubble Memory are totally different things. Bubble memory is (was?) a serial magnetic device in which small magnetized regions are shifted along a magnetic substrate. Flash memory uses charge stored in the gate region of MOSFET transistors - much faster, and random access.

Next:

"all bootable systems were ready for user input in a matter of milliseconds, right, because the OS was written on Mask-ROM"

Not true at all. If you ever worked on a PDP-11 or (heaven forbid) a PDP-8, you know that the OS had to be loaded before any program could be run. In fact, I used to have to toggle the boot loader into a PDP-8 by hand, which then drove the paper tape reader that read in the rest of the system. In terms of computing power and complexity, the desktop computer of today is most comparable to the mainframes of 30 years ago (e.g., IBM-360 or, from my fond memory, CDC-6400). These machines all had to be booted from disk or tape, just like todays PCs.

As far as multitasking goes, I'm not sure what you mean by "The Never really achieved "Multitasking" magical-goal.."

My desktop machine at home is running a web server and a database server (MySQL), as well as checking my email and doing various network tasks such as keeping my dynamic IP updated in a DNS server somewhere, all while I'm playing a real-time web based shoot-em-up game. I think we can safely say multitasking is alive and well :-)

With regard to software

Software and the functions it performs have gotten enormously more complex. I grew up on Fortran and yes, I have programmed in Forth (YUK !). These types of language are totally unsuited to graphical user interfaces (I built one in Fortran once. Again, YUK!) and the code was complex and often not reusable. The more recent object-oriented paradigm attempts to hide some of the complexity and encapsulate functions and data for simplification and reusability.

The "good old days" weren't all that good by the standards of what you can do today. Things were simpler because they did much, much less. Let's not go back there!

I agree with Steve that bubble memory lasted little only for the name which is left over. Most of the stored information is sequential on storage medium and that makes it slow. Even those Flash disks used for information give sequential information from random access system but somewhat faster. The reason is that format is not changed and only media is altered. Another problem is that you use information supplied by others and they also have some way to claim that it is their information and CDROM etc are information media supplied by the software companies. You use what you like. Only when you use limited software, it can reside anywhere for faster usage. You can also notice the way parallel processing machines are made. You gain on speed if you wish. Speed to few 100s of Terra Flops should be possible.

Steve,

Good answer to a post from left field somewhere.

Maybe he had a small computer with the operating system in ROM and CMOS memory (like my old HP-75C) otherwise I have no idea what he could possibly be talking about, except that he apparently has a slow machine right now.

Every so often the issue of using non-volatile memory comes up to allow a machine to come on "instantly" in its last state, but so far at least nothing much has come of it.

I am certainly no fan of MS, but how does he think you can have more than one "free running stand alone apps" without crashing instantly over one conflict or another? A "big brother" is necessary in computers to create a virtual machine platform that the apps make all their calls to. Without the virtual machine interface, each app requires its own set of specific hardware drivers, and we know what pain that was.

Von Neumann type CPUs don't really multitask in the real time sense to any extent, no matter what you do, but when they are fast enough, it seems as if they do, and that's what counts. The computer as a whole multitasks to a large extent on a low level because of its dedicated hardware, buffering and so forth, but that gets right back to the need for "something" to manage it all at the intermediate level.

As to PCs getting slower and slower, in reality the human interface hasn't changed all that much, as far as getting a response from a command if no disk reads or writes are involved, but the response we get now is so much "richer" than the 80 x 25 ASCII character based screen of the past, and WYSIWYG (What You See Is What You Get: meaning the printed output matches what you see on the screen) is the norm. But the high resolution, 32 bit color GUI and bitmapped fonts require a lot of computing power and high bus speeds across the board. When it comes to applications, and what they are now able to accomplish, the present PCs and software are astounding.

I do agree with him on the proliferation of "bloat", in both Windows and apps in general, but that trend has tended to track increases in processor power from the beginning. Old programs, running on old computers were "lean and mean" because they had to be. Right now the biggest single software issue slowing machines down is the unfortunate need for so much protection/detection, and/or prevention of malware.

I too worked (briefly) with a PDP-8, with the paper tape reader, tape punch, and teletypewriter. That company also had an electronic/electromechanical analog computer that was in constant use, mostly involving wave simulations for the Navy.

I enjoyed programming in Forth, but then I had become well accustomed to heavy programming on my HP-41CX calculator, which for those who don't know it was actually a small ½ bit computer masquerading as a calculator. With all its plug-in application modules, card reader, non-volatile memory, real time clock, HP-IL (a network implementation) peripherals (catalogued digital cassette tape drive, printers, RS-232, parallel and HP-IB interfaces, disk drive, monitor/TV interface etc), it was probably the apex of calculator development. I have a good portion of the peripherals. The astronauts used it on the early shuttle flights to calculate re-entry. The HP stack oriented RPN programming language had much in common with Forth.

I still use both my trusty HP-34C and HP-41CX, and the 34C is used daily and sits right on my desk.

For those interested in learning more about this amazing calculator:

http://www.hpmuseum.org/hp41.htm

There are also a number of HP calculator simulations (that work) for Windows at that site.

Enough of my off-post reminiscences!

Regards, Greg

By "free running stand alone apps" I referred to any app which does not need a registry check-up, or installation routine, but instead, once compiled and encapsulated, can be copied to any medium accessible and run As-Is (segmented to RAM) from there and on.

This however, does not prevent Time-Sliced Multitasking, as still roughly defined today, only without the "hussle" of registry checkup endless iterations, which is frankly meant to "hunt" piracy, rather than orchestrate processes or RAM Allocation

It's quite possible to write stand-alone apps. I have a number of them on my machine. There's no install process, you just drop them in a folder and run them.

Yvalmate has a good point. My routine is to start up my computer, sit around with nothing to do for up to a minute, until the lan sign-on comes up, then go get a cup of coffee, which I can usually finish before the computer is up and running, unless software updates are scheduled, then it can be another 15 minutes or so before I can actually start work.

The PC manufacturer that comes up with a product that can come on instantly would really have a leg up.

I agree that instant-on would be a good thing. One option you can try is to use "Stand by" or "Hibernate" instead of powering off.

However, the point in my reply to Yvalmate's post was that software and systems are more complex because we're asking them to do a lot more, and that his memory is faulty in recollecting that machines used to come on "instantly."

Just imagine if, instead of going to get your coffee after flipping the power switch, you had to spend your 15 minutes toggling in the boot loader

Loved your comment and appreciated it, stand corrected on Flash and Bubble issues, but these were an example of some kind of EEPROM to enable maintenance of variables and updatable segments required for the fast boot (The original idea submitted to MS R&D here in Israel was to sell the OS as a Plug-In Card...).

Just before you all bite my head off, keep in mind I really meant to speak of older-days PC's versus current ones, running some ten-thousand faster than today, in which context I compared for instance between, say Sinclair QL and today's PC's.

I don't even mind being corrected on the multitasking issue, as long as we all keep in mind that even NT's multitasking using what's called "virtual Multi Processors" are nothing more than a crude division of RAM space for what is eventually a threaded multi-process down the shoot of a single interrupt column. All this is not much more than an elaborate attempt to impress people, with a much smaller success in mnemonic terms of efficiency.

My issue with programming languages is that it all got into an "Object Library" culture, which significantly reduced runtimes efficiency, by filling it with irrelevant compilation garbage, while reducing the algorithmic accuracy of the handling mechanisms since the modern "scripts" were pre written for generalised purposes, not specific enough for each task in mind.

My main argument is not really technical, but cultural, only demonstrated with (half-cooked?) technical terms. Please, forgive my know-how and get to the point.

I can't claim to be old enough to remember those bubble-whatevers <*wink*> but I do think that to some degree, in spite of the increase in speeds/power/flops (or whatever unit you prefer) that our ability to create truly optimised software is very limited. I've been toying with programming microcontrollers and it's stunning to see how fast stuff can really run when there's only a few lines of code running at a only a few megahertz.

I also think anti-virus software is evil...after moving my ol' dependable Dell Latitude Cp 400 (<--yep, that's the speed in MHz) to xubuntu linux it was shocking to see how fast it did boot without the overhead of AV software and windoze. Let the experiments continue...

I have read in a "MaximumPC" Magazine (I cannot remember which month, might have been last November) that RAM manufacturers are currently developing RAM that will have the ability to store and hold information without a Charge (Like a USB Key) but have the speeds of traditional RAM. This would enable all your standard system Info to be keep saved in RAM even while the Machine is off, so when you boot up there is no need to load it because its already there. Close to instant boot-ups might just be around the corner, but with today's large and complex programs and Interfaces.

Instant boot? Why not just never turn it off...works for me.

Here is another suggestion. Load Windows on a 10 GB hard drive and use it for your boot drive, use another larger drive for installation of your apps and data, then WHEN not IF Windows crashes or gets infected, reformat the 10 GB drive, and reinstall Windows. or..........buy a Mac, I am thinking about it.

Nice... I think you just got my gisted blues

I am surprised no one has mentioned PDA's as an example of instant-on computers. I recently graduated from a laptop to a PDA out of necessity, and find that I can do quite a bit of my normal work outside the office with this device, including downloading test equipment data, even live logging. With MS's Excel (the only thing they have ever done well, in my opinion), I can have a good preview of the information, as well. They are programable (I use PocketC, and am looking at Forth for the same purpose- both being quite reasonable in price- evem free. in some cases). The PDA's can talk to the Internet (although I have a personal rule against ever putting one of my own computers on line), and even make phone calls. With Bluetooth, IR, WiFi, and who knows what else, the potential is significant. Expansion cards give me more storage that I need on a daily basis (although I still need my desktops for long-term storage). The built-in apps demonstrate very clearly that one does not need bloated software to maintain a calendar, track contacts, or keep daily notes.

And it turns on in less than a second of hitting the button.

Charlie

A good point for my taste: Maybe PDA's are not popular enough to "justify" the triple scrutiny for maleware pirated copies etc, a point arising once again in cases such as MAC OS, as well as myriad of firmware based aparats and gadgets from multi-codec video players to controller-based household fixtures...

"...it turns on in less than a second of hitting the button..."

I would bet that its operating system is loaded to RAM from ROM, as was the case with 1980s PCs such as Atari, Texas, Sinclair, Commodore, Apple - the whole operating system was up and running in less than a second.

And that's on hardware thousand times slower than today's.

PCs load their operating system from some kind of EEPROM (be it HDD, Flash, SSD) which is about a thousand times slower than Mask-ROM.

Were you around in the early 80s? I was self taught with an Atari assembler. And a digital electronics correspondence course. It was what was available at the time. Where I lived. I did take some formal classes later when they became available.

I did just the opposite of what you suggest. And Atari even used some of my routines. I was quite ignorant of the situation and received no money for it. I got quite addicted for awhile and was only concerned with precisely controlling and timing hardware registers........if you have ever gotten into something like that. It was a new kind of magic for me. I only had analog experience.

I was the first one in town to acquire 64K of memory..........what would one do with so much memory? 16K was the usual amount. Well, i wrote a boot routine where the OS from ROM.....was loaded into RAM......and then modified to my liking. I had discovered a lot of no-op instructions in the ANTIC video routines. And during sync and blanking....I could set up routines to display and maintain sweeping radar screens and such.....completely in the background. I could customize all kinds of things to the OS.

Years later, when windows came out, I asked my boss, a windows 2000 programmer..........why don't we do what you just suggested......because of that long boot time. Put OS on ROM. win 2000 was the best at the time. But win was never stable until 7. BSOD was very common.

And you are the first person since, to mention it.

Of course the size of the Atari OS was peanuts, and even with putting OS into RAM, it seemed instant.

That's what I recall. I just put together a new desktop with win 11 pro. 10 second boot.