[llvm-commits] Global Merge Pass for ARM

[Evan Cheng]

Hi Anton,

Looks good in general. A coupld of comments.
1. Use SmallVector instead of std::vector?
2. Can you make it a target independent pass? This should be profitable to many targets whose indexed address mode.
3. MaxOffset should not be initialized when the pass is created. What if we want to support switching between ARM / Thumb mode on a function to function basis? Can it use TargetLowering::isLegalAddressingMode() or something similar instead?

Evan

On May 16, 2010, at 3:49 AM, Anton Korobeynikov wrote:

> Hello, Everyone
> 
> Please find the patch which can be viewed as some early approximation of
> "section anchors" feature seen in gcc. 
> 
> It tries to solve the following problem: consider the code touches
> several global variables at once, e.g.:
> 
> <=cut=>
> static int foo[N], bar[N], baz[N];
> 
> for (i = 0; i < N; ++i) {
>  foo[i] = bar[i] * baz[i];
> }
> <=cut=>
> 
> On ARM the addresses of 3 arrays should be kept in the registers, thus
> this code has quite large register pressure (loop body):
> 
>        ldr     r1, [r5], #4
>        ldr     r2, [r6], #4
>        mul     r1, r2, r1
>        str     r1, [r0], #4
> 
> Pass converts the code to something like:
> 
> <=cut=>
> static struct {
>  int foo[N];
>  int bar[N];
>  int baz[N];
> } merged;
> 
> for (i = 0; i < N; ++i) {
>  merged.foo[i] = merged.bar[i] * merged.baz[i];
> }
> <=cut=>
> 
> and in ARM code this becomes:
>        ldr     r0, [r5, #40]
>        ldr     r1, [r5, #80]
>        mul     r0, r1, r0
>        str     r0, [r5], #4
> 
> note that we saved 2 registers here.
> 
> This way only the address of the merged structured needs to be kept in
> the registers. For the fields accesses ldr/str with offsets are used.
> Pass correctly distinguishes constant and non-constant globals. Maximum
> size of the struct dependes on the instruction set used (it's 4095 for
> ARM/Thumb2 and 127 for Thumb1).
> 
> Maybe PPC can benefit from this pass as well, but I'm not yet sure.
> 
> Ok to commit?
> 
> -- 
> With best regards, Anton Korobeynikov.
> 
> Faculty of Mathematics & Mechanics, Saint Petersburg State University.
> <ARMGlobalMerge.cpp>_______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits