13

MS-DOS's debugger DEBUG.EXE did only support the assembly/disassembly of 8086 opcodes. DR-DOS's debugger SID86.EXE, SID.EXE and Novel's debugger DEBUG.EXE did support 80286 opcodes and more, but no AMD, Cyrix, NEC etc. specialties.

So I'm wondering if the CPU manufacturers like AMD, NEC, Cyrix and others provided their own debugger for DOS and if so, if it was offered as Freeware or did it cost money?


Edit: I took the liberty of clarifying the question a bit.

In the DOS days, there were CPU manufacturers like NEC that had additional opcodes in their V20/V30 CPUs which where compatible replacements for the Intel 8080 and 8086. And Cyrix and IIT had extensions in their CPUs and co-processors.

Therefore, it would have made sense for these CPU manufacturers to better support their CPUs and FPUs with their own debugger, allowing developers to take advantage of their additional features and make their CPU and coprocessor products look better in tests.

3
  • 1
    AMD has CodeAnalyst and its successor CodeXL, but that's for Windows, not DOS.
    – dan04
    May 3, 2023 at 18:43
  • 2
    Did they? Yes - Intel has always provided devtools for their CPUs. Were they free? Not at all, at least not in the early days. The first editions of the Intel compiler for Linux cost a lot of money (at least for commercial use)
    – tofro
    May 4, 2023 at 10:38
  • Thank you. But does this also apply to the other x86 CPU and x87 FPU manufacturers? After all, Intel was the market leader with the greatest spread and set the minimum standards for what an x86 CPU and x87 FPU must be able to do. From that point of view, they didn't really need to distribute debuggers cheaply or free of charge. But could the other manufacturers who built extended opcodes into their CPUs/FPUs really do the same? Cheap or free debuggers would have been a useful promotional tool to better sell the CPUs/FPUs and spread the use of the opcodes better. So did they ask for money too?
    – Coder
    May 4, 2023 at 15:14

1 Answer 1

12

No. While DOS competitors did improve on various DOS commands to try and differentiate their products from Microsoft's there was no equivalent motivation for CPU manufacturers whose products were used with lots of different system architectures and operating systems than PC Compatibles running DOS.

An update to a not widely used utility program for a tiny fraction of users of their processors would have been a poor investment. That is before you consider how it would even be distributed in the days before widespread modem use.

6
  • 1
    There was also the third-party debugger SoftICE. As I recall it was fairly expensive but was, and probably still is, the best debugger out there. Unfortunately the last version of DOS/Windows supported was XP.
    – jwh20
    May 3, 2023 at 20:36
  • 6
    “That is before you consider how it would even be distributed in the days before widespread modem use.” — CPU manufacturers didn’t mind shipping manuals, and later, CD-ROMs, to anyone who asked; floppies are cheaper to send than bulky manuals so I don’t think distribution would have been much of a concern. (But that’s minor, +1 from me.) May 3, 2023 at 21:36
  • 1
    Well, there were additional opcodes in the NEC V20/V30 (a replacement for 8080 and 8086) and extensions in Cyrix and IIT CPUs and co-processors. And CPU manufacturers had an interest in making their CPUs and FPUs look better than Intel's. And all that in the days of DOS. So it would have made sense for these CPU manufacturers to give developers better debuggers that support the special features of their CPUs and FPUs.
    – Coder
    May 3, 2023 at 22:24
  • 4
    Serious users would use SoftICE (mentioned above) or a hardware ICE from Intel or another vendor. Everyone else was used to hardship in those days, that's just the way it was. A debugger was usable with just a little extra effort on the user's part as long as there were no new registers. For everything else: you understood octal or hex and that sufficed.
    – davidbak
    May 3, 2023 at 22:52
  • 1
    @Coder: An interesting feature of the NEC V20 is that its documentation used the names IX and IY to refer to the SI and DI registers, probably to avoid any copyright claims by Intel, which made the descriptions of the new instructions a bit confusing until I figured out that the existing instructions using IX and IY were the same as Intel instructions using SI and DI.
    – supercat
    May 4, 2023 at 16:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .