test please delete

[voodooflux]

Hmmm, neither the PS3 or the 360 will recognise my external hard drive, so I'll have to copy the video onto a memory stick instead.

[voodooflux]

Hmmm, neither the PS3 or the 360 will recognise my external hard drive, so I'll have to copy the video onto a memory stick instead.

Or I could stream it from the NAS...hmmm, choices, choices...

[BrilloPad]

The Amstrad PC1512 (and PC1640) preserved registers across BIOS calls. This fact was clearly documented in the Amstrad Technical Reference Manual.

What wasn't clearly documented was that, in doing this, it differed from every other IBM PC (and compatible) BIOS except Compaq's.

As a result, when I was first writing games, we shipped a game to the publisher which, as part of initialising the display when the game started, would make a BIOS call with registers that had been modified by the previous BIOS call, with the result that it switched the CGA card from graphics to text mode, thereby totally buggering things up

I could only debug it by walking up to the Tandy shop and asking the salesman to let me check it on one of their display models. Furthermore, I had to make sure that the fix occupied the same number of bytes as the original code, for easy patching of the disks that had already been duplicated

It didn't take long to sort it out - I just replaced the BIOS-calling code with code that hit the hardware directly, went back to Tandy to check it, and shipped the patch.

On the bright side, the offending code had been written by another chap before I started at the company - I'd taken the project over when I was hired

Top Amstrad techie post

[voodooflux]

[NickFitz]

Top Amstrad techie post

Eye thank ewe

"Recondite but true" - that's my motto

And a hint of "totally useless knowledge nowadays" thrown in for good measure

[EqualOpportunities]

The Amstrad PC1512 (and PC1640) preserved registers across BIOS calls...

I hope you realise, Nick, that whilst I consider myself to be somewhat smarter than that average bear, I haven't so much as the first clue what that means?

Maybe we should have a 'bowing to your superior technical knowledge' smiley

[NickFitz]

I hope you realise, Nick, that whilst I consider myself to be somewhat smarter than that average bear, I haven't so much as the first clue what that means?

Maybe we should have a 'bowing to your superior technical knowledge' smiley

An 8086 processor, as used in AMS's PC range, has registers called AX, BX, CX, DX, DI, SI, BP, SP, and IP (if I remember correctly) - they are the in-CPU memory that the processor keeps values in, whether such values be the address of the next instruction to be executed (IP, for Instruction Pointer) after the current instruction has finished executing, the offset into the stack where the current stack frame (including the return address from the current subroutine) is located (either SP alone, or a combination of SP and BP if one used full-blown stack frames for passing parameters and for local variables), or simply data.

Being on the chip and therefore directly accessible to the ALU and control unit, registers are the single most valuable piece of memory in a computer.

It's always an important consideration, at machine code level, as to whether registers should be preserved across calls to BIOS or OS functions. AFAIK, the reason that the original IBM PC BIOS didn't preserve registers across functions was mainly to reduce code size; the additional instructions for pushing the values onto the stack on entry and restoring them on exit occupied significant ROM space in 1981.

By the mid-80s, there was plenty of space available on the average ROM chip, so it wasn't a problem to guarantee register integrity across BIOS calls. However, there was in fact no point doing so, as the vast majority of the installed base of PC/MS-DOS machines didn't do so, and therefore software couldn't take advantage of this new capability and still expect to run on any standard PC hardware.

Still, somewhere down the line, Alan Sugar paid for the time it took for a programmer to ensure register integrity across BIOS calls, and there was no point doing so except from a sense of elegance.

It's funny to think that probably the only Amstrad product component ever to receive wholly unnecessary investment was a useless refinement of the machine code in the BIOS of a long-defunct computer

[BrilloPad]

morning all

[BrilloPad]

Population
1816 Revenues
419650 € Unemployment 1 %
Transport 99 %
Crime rate 2 %
Pollution 1 %

rank=310

[cailin maith]

Morning