Announcement

Collapse
No announcement yet.

Microsoft plans to completely drop support for x32 in Windows Core 10 OS

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #61
    Originally posted by Anne Wilson View Post

    Rumors maybe not ....
    End of an era: Linux distributions will soon stop supporting 32-bit PCs
    https://www.pcworld.com/article/3089...2-bit-pcs.html
    You are aware of the difference between a 32 vs. 64-bit computer (aka CPU architecture) and a 32 vs. 64-bit Operating System, aren't you? Hint: most of you're 32-bit Windows installtions are running on 64-bit PCs/servers for quite some time now.

    So tell me, 1) when did you last purchase a PC with a 32-bit architecture? and 2) how many of those machines are still active?

    Comment


    • #62
      Originally posted by Knuth Konrad View Post

      You are aware of the difference between a 32 vs. 64-bit computer (aka CPU architecture) and a 32 vs. 64-bit Operating System, aren't you? Hint: most of you're 32-bit Windows installtions are running on 64-bit PCs/servers for quite some time now.

      So tell me, 1) when did you last purchase a PC with a 32-bit architecture? and 2) how many of those machines are still active?
      Yep, basically, they are saying that you won''t be able to install on a computer with an 80486 or Pentium 4 processor. Haven't seen too many of those in new machines for 15 years. If your computer is less than 10-12 years old, it will almost certainly have a 64 bit processor.

      Comment


      • #63
        To be fair, Anne may simply be unaware of the "multiarch support" built into Ubuntu to prevent chucking years of software into the dumpster. Although *nix might not have the "commercial pressure", most commercially-viable distro developers really aren't so dumb as to abandon 2/3 of their software base on a utopian vision of perfection.

        Comment


        • #64
          "To be fair", she should check dates and references before linking to nonsense. Also, we've heard her opinion before, no need to repeat ... and repeat ... and repeat ...
          Dale

          Comment


          • #65
            Originally posted by Raymond Leech View Post
            To be fair, Anne may simply be unaware of the "multiarch support" built into Ubuntu to prevent chucking years of software into the dumpster. Although *nix might not have the "commercial pressure", most commercially-viable distro developers really aren't so dumb as to abandon 2/3 of their software base on a utopian vision of perfection.
            "To be fair", she clearly confused "CPU architecture" with "operating system" and consequently totally misunderstood what her link was actually saying.

            Comment


            • #66
              I have had access to 64 bit for over a year and what I get up to does not tick any of the 64 bit boxes. Having said that when I finish a project I always run in 64 bit to see how it fairs with 32 bit. More often that not there is very little difference between the two from a performance perspective.

              I have just finished an application which does AES, RSA and ECDSA. I have just encrypted a 100MB file and got 892.5ms in 32 bit and 975.7ms in 64 bit; clearing the filecache first on both runs. Now I know that we should not do singleton runs and I normally don't - six to ten is what I normally do. With decent tests there has not been much in it. The 64 bit has a larger binary and it takes much longer to compile. With this project there is no case for 64 bit.

              A lot of folk are marketing 64 bit based upon bragging rights and probably have 32 bit versions which run faster. On the other hand some folk are marketing 64 bit for very good reason where their software has ticked at least one 64 bit box. It is like some folk bragging about their encryption software which has "military strength" becuase AES256 is being used. What load of bull. AES128 is perfectly fine for most people and for sometime to come. Quantum computers for the masses is some way off but even when the security agencies get their hands on them there is some debate whether AES256 has anything to worry about. RSA will be dead in the water, but that is a different story.

              Someone here, in another thread, mentioned that some NHS managers were suggesting that he should up his game with 64 biit. I felt really sorry for the guy. Our NHS has just undergone a security test and everyone of them, I will repeat that, everyone of them failed. Why? Abject ignorance, and probably the same with 64 bit.

              Me: I can let you have a 64 bit version if you want but I must tell you that it runs slower then 32 bit.
              Them: Slower!
              Me: Most of the 64 bit talk is hype. I have a 32 bit copy in the car. I'll go and get it and you can see for youself. While I am out reflect on the fact that I charge more for the 64 bit.
              Them: You charge more!
              Me: Well, yes, it is state of the art, after all. <BIG GRIN>

              Comment


              • #67
                I would like to comment that indeed the lack of a 64 bit compiler is now impacting a small portion of our market place. About 18% of software subscribers are now collecting radar data with multichannel systems. Typical sizes for datasets that use to be about 1gb or less are now 50-500Gb for average sites. We have some old clients and just a few that even are getting ready to collect radar data for monitoring on railways over a 100km or more. You can image our data densities with radar profiles collected at 6cm between 20 channels, with 512 pulse digitization and inline collection at 5cm. We have been temporarily forced to partition large datasets to 100x100 meter blocks but these operations reduce the speed of operations significantly and will lock us out of the market for large sites eventually if not already.

                It would be great to know that PB is working on a 64bit compiler. I will be completely honest, if there were no advancement made in this direction since Drake Software took over i will be forced to abandon alot of the code engines that require us to work with larger datasets beyond 4GB. It would be great if Drake Software can give us an update on developments here and also, an eta whether it is next week, next year or never!!! If it is never we need to know....And if no development in the compiler editing has been made since the takeover, we deserve to know this information as well so we can make timely decisions for our own futures!

                Comment


                • #68
                  Originally posted by dean goodman View Post
                  It would be great if Drake Software can give us an update on developments here and also, an eta whether it is next week, next year or never!!!
                  Adam gave us a roadmap a while back.

                  Some may pick at the roadmap for this reason, but compilers are a slightly different critter. It is inevitable that a good compiler will become self-hosting (multi-platform compilers being the exception).
                  I am legally blind. Please forgive any typos. I do try and catch as many as I can.

                  Comment


                  • #69
                    I have to laugh at some of the comments, both Microsoft and Intel/AMD know that if they ever try the same trick again like Intel did with the Itanium, a vast number of people simply would not buy it. The number of people using XP makes the point that unless there is some advantage to them in upgrading, they just keep using the same box and software. Just as an example, my last XP box is a 9650 Intel Core2 quad and while it handles SSE4.2 and is a 64 bit box, it also runs all of my 32 bit software perfectly on a 32 bit OS.

                    Now with 64 bit code, once you know how to write it and use a toolset that can actually deliver, you can end up with truly eye watering performance and it is for a number of reasons. 16 gp 64 bit registers mean far less stack shuffling preserving the non volatile registers, a faster calling convention with 64 bit FASTCALL, 16 XMM registers, 16 YMM registers, you still have the old MMX/FP registers and you have massive amounts of memory if you actually need it.

                    In 32 bit you have to use /LARGEADDRESSAWARE to get a bit more than 2 gig, in 64 bit you can routinely allocate 50 gig if the box has enough memory and tear through it at rates that most would wet themselves at. Now of course this is great IF you are working on massive quantities of data but a vast amount of tasks performed by software fit into the 32 bit memory addressing range and if written correctly it will perform the intended task(s) easily fast enough without the need to go to 64 bit.

                    As far as our friends who tremble at the prospect of being left behind with their 32 bit software, the solution is to hold their breath waiting for the end of 32 bit Windows with the main gain being that we will never hear from them again.
                    hutch at movsd dot com
                    The MASM Forum

                    www.masm32.com

                    Comment


                    • #70
                      Originally posted by Steve Hutchesson View Post
                      I have to laugh at some of the comments, both Microsoft and Intel/AMD know that if they ever try the same trick again like Intel did with the Itanium...
                      Most who are asking for x64 support in the various communities, probably do not even realize that x64 was invented by AMD and is licensed by Intel. Intel's 64-bit offering ((Itanium) which was co-developed with HP) was a miserable failure. Of course, all of these lovely 64-bit processors are using versions of Windows based on the NT kernel (which was originally created by IBM & DEC). It is nothing short of a miracle that any of this crap works properly as it is literally kludged together with duct tape and bubblegum. PowerBASIC is the most problem-free compiler I have ever used. If PB never saw another release, it would not necessarily be a bad thing, as long as it continued to work on current systems. I will take stability over new features any day.

                      That said, five years ago, I did a contract job for a defense contractor. It required 64-bit, hence I did not use PowerBASIC. If you need 64-bit, there are alternatives that will get the job done. There is no need to wait in hope for 64-bit to come to PowerBASIC.
                      I am legally blind. Please forgive any typos. I do try and catch as many as I can.

                      Comment


                      • #71
                        PB is already very ingenious. It is legitimate to have a dream. I use other tools for 64-bit. But, I hope on PB64. Still, I think PB died. Unfortunately.

                        Comment


                        • #72
                          Originally posted by Bernhard Fomm View Post
                          Still, I think PB died. Unfortunately.
                          Adam gave us a nice roadmap. It will take time, lots of it.

                          At this point, most of us are happy the lights are still on and will not be going out anytime soon.
                          I am legally blind. Please forgive any typos. I do try and catch as many as I can.

                          Comment


                          • #73
                            I have always though one should just program for high dataset and let the computer hardware innovations cover what you did not see coming. And an example would be the growing of data simply by aging of an entity. Keep all the data that you produce and let time take care of handling the storage sizes and speed increase that maybe needed to access that data. Just because it is not the most necessary need of today does not take of tomorrow. Of course you should optimize your software because the data will grow.
                            If you use slow computers for your test environment, you will be happy with the results on a much faster computer. It does take more time and thought but it sure solves a lot of future problems.
                            p purvis

                            Comment


                            • #74
                              Windows 10 Core OS could support Win32 apps in the future
                              https://windowsreport.com/windows-10...os-win32-apps/

                              Comment

                              Working...
                              X