Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!ucsd!sdcc6!sdcc10!cs163aeo From: cs163aeo@sdcc10.ucsd.EDU (Professor I.R. Gumby) Newsgroups: comp.sys.atari.st Subject: Re: PC Ditto II -- speed? Summary: Re: Re: PC Ditto II - speed? Message-ID: <105@sdcc10.ucsd.EDU> Date: 6 May 89 19:24:39 GMT References: <8905021217.AA08520@nh.cs.wm.edu> <98@sdcc10.ucsd.EDU> <103@sdcc10.ucsd.EDU> Organization: University of California, San Diego Lines: 85 In article <103@sdcc10.ucsd.EDU>, cs163afu@sdcc10.ucsd.EDU (Some call me...Tim) writes: > The initial release of the IBM PC AT contained a 6Mhz 8086 > processor. It was only later that they upgraded to 8Mhz and still > later the 286 processor. I hate to break this to you, but Intel only manufactures 8086/8088 chips in the following clock speeds; 4.77, 8, and 10. There was never a 6 MHz 8086 or 8088. The Intel 80286 chips are available in these clock speeds; 6, 8, 10, and 12 (from a company that does subcontracting work for Intel, but these are not really Intel chips). Harris makes 16 and 20 MHz 286's. > The PC Ditto II emulator is an 8086 emulator, not a 80286 emulator. > If it were, it would be no problem to include things such as protect > mode, etc. > > Most IBM software runs on all model IBMs. Software that takes > advantage of the 286 (OS/2 etc) is USUALLY expensive (maybe not > universally--that's NOT my point!), and most products are available > in IBM PC XT/AT (non-286!) versions, even if they do support the > extensions. I also hate to burst your bubble, but the very first IBM AT (true blue) was a 6 MHz 286 box that ran for about $20,000 each. When you boot up a 286 box, it is not in protect mode. In fact, all operating systems that use 286 protect mode have to put the instructions and/or BIOS call that puts the processor into protect mode have to do it in the boot program. I know this for a fact because I have worked with original true blue IBM AT's. What makes an AT box is the processor, BIOS, and assorted support chips (bus controllers, slots, DMA controllers, etc.) these differences can be seen and examined closely when you put an XT-286 and an AT back to back and look at them; if you don't believe me (and the other person who responded before me to what is an AT?) then I invite you to take it to comp.sys.ibm.pc. I don't mean this in an insulting for flaming sort of tone. I just want the record straight to what PC Ditto II can't do. With emulators I'm more concerned with what it can't do rather than can do. The more accurate statement for PC Ditto II's performance would be an XT class machine with AT class speed most likely, but the average SI of a typical AT box (which is running at 10 MHz) is about 10, but the SI on an XT box with a 10 MHz 8088 is about 2.1. So remember, 1 MHz on an XT class machine is NOT identical to 1 MHz on an AT class machine. It's just like comparing apples to oranges or more appropriately kilometers to miles. Most IBM benchmarking programs have the bench set relative to two machines, the original IBM XT and the original IBM AT. > BESIDES: I didn't say this--the Avant-Garde representitive did. > And I agree with his appraisal. You may agree and that is your right to do so, but please be sure that you are 100% sure that it's an AT class machine. My peeve is people being misinformed. I admit that I do it too sometimes, but please be careful about it. If this PC Ditto II was really an AT class emulation box, I would jump for joy and use it for cross compilation of my 286 protect mode Minix. I could use my real AT for testing, and the ST with PC Ditto II to build the kernel, but I know that it isn't going to work that way because I know from my experience with IBM XT and AT hardware plus the fact I have read up on PC Ditto II that it will not be an AT emulation box. > What was that reference about Borland entering the ST market? Is > this true? (About two years ago I knew someone who's friend was > "beta testing" Turbo Pascal on the ST, but I haven't heard anything > about it since then.) Turbo C ST is already being marketed in Europe (along with Turbo Pascal I believe). I saw an article on comp.os.minix requesting for information on building the ST Minix from TOS with Turbo C ST (just like we have been doing since Turbo C 1.0). If I can get in touch with some relatives of mine in Europe, I will see what I can do on having it mailed to me. I don't know how far Borland has gone with marketing it, but I do know it exists, but only currently in Europe. Borland could be afraid of the fact that the Atari 8-bit machines were plagued with piracy and are worried about that mentality transforming to the ST. I quite frankly don't blame them. Borland produces a good compiler. I have used Turbo C 1.5 for a year now. Very nice compiler, although it somewhat lags in optimization. If you have the experience, probably GNU C would be better to obtain and make run on your ST from TOS. But if you want the support of a major software company, then going with a company like Borland is a good way to go about having your compiler and support for it too. JCA