de Brebisson, Cyrille (Calculator Division) writes:
> [snip] trying to re-code, using inline assembly goto *jump[*progc++]
> I used inline assembly to do:
> Ldrh instr, [progc], #2 // note that in most cases, there is an
> // extra instruction here that allows to
> // cancel the waitstate caused by the use
> // of register instr on the next
> instruction
> ldr pc, [jump, instr, asl #2]
>
> because the compiler generates the highly unoptimized (and too large for
> the memory in my device)
> ldrh r1, [r4], #2
> ldr r8, .L2691+4
> ldr fp, [r8, r1, asl #2]
> mov pc, fp @ indirect register jump
> [/snip]
>
> >This is the crucial mistake: you can't jump out of an inline asm.
>
> So, how can I optimize my code? Is there a way to force the compiler to
> 1: put a variable in a register? As the asm ("register"); constraint
> does not seem to do a lot of forcing
Definitely: if declaring a global register variable doesn't work,
that's a bug. What exactly did you try?
> 2: get the compiler to condense the last 2 instructions in 1?
I'm not sure why gcc generates that sequence. Forwarding to Richard
Earnshaw for comment.
Andrew.
> -----Original Message-----
> From: Andrew Haley [mailto:aph@xxxxxxxxxxx]
> Sent: 08 December 2006 09:43
> To: de Brebisson, Cyrille (Calculator Division)
> Cc: gcc-help@xxxxxxxxxxx
> Subject: Re: Help needed: Optimization of bytecode interpreter for ARM
> paltform
>
> de Brebisson, Cyrille (Calculator Division) writes:
> > Hello,
> >
> > I hope that this is the best location to ask this question, if not,
> please accept my apologize and redirect me where needed.
> >
> > I am trying to write a fast byte code interpreter, but the compiler
> optimizer just 'does not get it' and generates bad code (it does not
> realize that they are jumps everywhere and optimizes out the code
> out)...
> >
> > Here is a simplified version of the code:
> >
> > static int rom[]=
> > { 0, 1, 2, 3, 4, 5, 6, 7, 8,
> > 9, 10, 11, 12, 13, 14, }; // the 'program'
> >
> >
> > void execute()
> >
> > {
> >
> > const void * const jumps[] =
> > { &&ins000, &&ins001, &&ins002, &&ins003,
> > &&ins004, &&ins005, &&ins006, &&ins007 }; // table of jumps
> >
> > register int carry asm ("r0");
> > register int instr asm("r1"); // currently executed instruction
> > register int *pc asm ("r4"); // program counter, points on next
> instr.
> > register const void * const * jm asm ("r5") = jumps; //pointer jump
> table
> >
> > int a=0, b=0; // virtual machine registers
> >
> > // this macro does a fast carry=0; goto *jumps[*pc++];
> > #define next asm ("ldrh %2, [%0], #2\n\t" \
> > "mov %1, #0\n\t" \
> > "ldr pc, [%4, %2, asl #2]" :
> > "=r" (pc),
> > "=r" (carry),
> > "=r" (instr):
> > "0" (pc),
> > "r" (jm))
> >
> > // this macro does a fast goto *jumps[*pc++];
> > #define nextnocarry asm ("ldrh %1, [%0], #2\n\t"\
> > "ldr pc, [%3, %1, asl #2]" :
> > "=r" (pc),
> > "=r" (instr) :
> > "0" (pc),
> > "r" (jm))
>
> This is the crucial mistake: you can't jump out of an inline asm.
>
> Andrew.
>
