View Larger Image. Ask Seller a Question. Title: Introduction au microprocesseur Cours Publisher: Editions universitaires europeennes EUE.
|Published (Last):||27 March 2017|
|PDF File Size:||15.13 Mb|
|ePub File Size:||16.96 Mb|
|Price:||Free* [*Free Regsitration Required]|
I have documented a Z80 flags feature that noone has already done. This list contains every documented and undocumented interrupt call known. He is well-known in cyberspace for maintaining the Interrupt List.
We all appreciate his continued support. GNU GCC is fine, but it's more aimed at 32 bit micros and I get the impression that you're targetting your work for the low-end side. Writing a compiler can be a very interesting and rewarding experience in its own right. However if writing the compiler is not your primary goal i. Compilers are relatively easy to write. You can pick up the bare bones of one from many books or GPL source.
The black art is the machine code generator. Producing optimised machine code for some CPUs is very difficult. Before you set out on you epic quest, try converting the following code to assembler for your target CPU and you may get a feel for what lies ahead. Try defining some of the above variables as signed chars, unsigned chars, short ints, long ints, floats and doubles. Generate assembler that handles the mixed combinations efficiently.
Did you have any problems with allocating registers or computing addresses? Consider how your compiler is going to generate code to compute the addresses and how it is going to be potentially restricted to using your scarce registers to build an address. Is this going to conflict with keeping your result in a register.
Do you have a stack available to you to spill your registers to when you run out? How efficient are the stack addressing operations available on your target CPU? Will you need to use directly addressed dedicated RAM locations? Some CPUs have only one index register and severely restrict the use of the stack pointer, so referencing parameters, locals and global arrays becomes a serious juggling act involving the index register.
This also affects how you prepare parameters to be passed. How will you reference tables in RAM and ROM where the code and data have their own address and data buses and both overlap Harvard architecture? If you find that your target CPU is too difficult to generate code for, maybe you should consider choosing a more advanced CPU as your starting point.
Learn on that and then go back to your original CPU. I've got a copy of his articles here somewhere , should someone need them. His articles are in my permanent collection, both on disk and on the bookshelf. Wirth is always good reading, and I had forgotten that article, but will take it out again. The book by Mak is not familiar, though I do have an older one by Gries which was quite good though not reflective of the current state of technology.
MASM Documentation. Retour au sommaire. Email :?
Lectures recommandées sur ce thème :
assembleur 8086 instruction