oscar64

Optimizing small space C Compiler Assembler and Runtime for C64

History and motivation

It is a sad fact that the 6502 used in the Commodore64 and other home computers of the 80s has a poor code density when it comes to 16 bit code. The C standard requires computations to be made with ints which work best if they have the same size as a pointer.

The 6502 also has a very small stack of 256 bytes which cannot be easily addressed and thus cannot be used for local variables. Therefore a second stack for variables has to be maintained, resulting in costly indexing operations.

A C compiler for the 6502 thus generates large binaries if it translates to native machine code. The idea for the oscar64 compiler is to translate the C source to an intermediate 16 bit byte code with the option to use native machine code for crucial functions. Using embedded assembly for runtime libraries or critical code should also be possible.

The resulting compiler is a frankenstein constructed from a converted javascript parser a intermediate code optimizer based on a 15 year old compiler for 64bit x86 code and some new components for the backend.

The performance of interpreted code is clearly not as good as native machine code but the penalty for 16bit code is around 40-50% and less than 10% for floating point. Code that can use 8bit my suffer up to a factor of 10 to 20.

The goal is to implement the actual C standard and not some subset for performance reasons. So the compiler must support:

• Floating point
• Recursion
• Multi dimensional arrays
• Pointer to structs

Limits and Errors

After four weeks, the compiler has now matured significantly. There are still several open areas.

Language

• No long integer
• No struct function return
• Missing const checks for structs and enums
• Missing warnings for all kind of abuses

Linker

• No explicit sections for code, data bss or stack
• No media file import

Standard Libraries

• No file functions

Runtime

• No INF and NaN support for floats
• Underflow in float multiply and divide not checked
• Basic zero page variables not restored on stop/restore

Optimizing

• All global variables are considered volatile
• Simple loop opmtimization
• Poor bookeeping of callee saved registers
• Partial block domination analysis
• No register use for arguments
• Auto variables placed on fixed stack for known call sequence

Intermediate code generation

• No check for running out of temporary registers
• Wasted 7 codes for far jumps

Native code generation

• More byte operation optimisation required
• Simple loop detection and optimisation not complete

Compiler arguments

The compiler is command line driven, and creates an executable .prg file.

oscar64 {-i=includePath} [-o=output.prg] [-cr=runtime.c] [-e] [-n] [-dSYMBOL[=value]] {source.c}

• -i : additional include paths
• -o : optional output file name
• -cr : alternative runtime library, replaces the crt.c
• -e : execute the result in the integrated emulator
• -n : create pure native code for all functions
• -d : define a symbol (e.g. NOFLOAT to avoid float code in printf)

A list of source files can be provided.

Implementation Details

The compiler does a full program compile, the linker step is part of the compilation. It knows all functions during the compilation run and includes only reachable code in the output. Source files are added to the build with the help of a pragma:

#pragma compile("stdio.c")

The character map for string and char constants can be changed with a pragma to match a custon character set or PETSCII.

#pragma charmap(char, code [,count])

The byte code interpreter is compiled by the compiler itself and placed in the source file "crt.c". Functions implementing byte codes are marked with a pragma:

#pragma	bytecode(BC_CONST_P8, inp_const_p8)

The functions are written in 6502 assembly with the __asm keyword

__asm inp_const_p8
{
lda	(ip), y
tax
iny
lda	(ip), y
sta	$00, x
lda	#0
sta	$01, x
iny
jmp	startup.exec
}

The current byte code program counter is (ip),y. The interpreter loop guarantees that y is always <= 128 and can thus be used to index the additional byte code arguments without the need to check the 16 bit pointer. The interpreter loop itself is quite compact and takes 21 cycles (including the final jump of the byte code function itself). Moving it to zero page would reduce this by another two cycles but is most likely not worth the waste of temporary space.

exec:
    lda	(ip), y
    sta	execjmp + 1
    iny		
    bmi	incip	
execjmp:
    jmp 	(0x0900)

The intermediate code generator assumes a large number of registers so the zero page is used for this purpose. The allocation is not yet final:

• 0x02-0x02 spilling of y register
• 0x03-0x09 workspace for mul/div and floating point routines
• 0x19-0x1a instruction pointer
• 0x1b-0x1e integer and floating point accumulator
• 0x1f-0x22 pointers for indirect addressing
• 0x23-0x24 stack pointer
• 0x25-0x26 frame pointer
• 0x43-0x52 caller saved registers
• 0x53-0x8f callee saved registers

Routines can be marked to be compiled to 6502 machine code with the native pragma:

void Plot(int x, int y)
{
	(*Bitmap)[y >> 3][x >> 3][y & 7] |= 0x80 >> (x & 7);
}

#pragma native(Plot)