1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-25 14:02:52 +02:00
Commit Graph

10 Commits

Author SHA1 Message Date
Jakob Stoklund Olesen
acaf9e9ce1 Be more aggressive about following hints.
RAGreedy::tryAssign will now evict interference from the preferred
register even when another register is free.

To support this, add the EvictionCost struct that counts how many hints
are broken by an eviction. We don't want to break one hint just to
satisfy another.

Rename canEvict to shouldEvict, and add the first bit of eviction policy
that doesn't depend on spill weights: Always make room in the preferred
register as long as the evictees can be split and aren't already
assigned to their preferred register.

Also make the CSR avoidance more accurate. When looking for a cheaper
register it is OK to use a new volatile register. Only CSR aliases that
have never been used before should be avoided.

llvm-svn: 134735
2011-07-08 20:46:18 +00:00
Benjamin Kramer
52ffb6ea96 BuildUDIV: If the divisor is even we can simplify the fixup of the multiplied value by introducing an early shift.
This allows us to compile "unsigned foo(unsigned x) { return x/28; }" into
	shrl	$2, %edi
	imulq	$613566757, %rdi, %rax
	shrq	$32, %rax
	ret

instead of
	movl    %edi, %eax
	imulq   $613566757, %rax, %rcx
	shrq    $32, %rcx
	subl    %ecx, %eax
	shrl    %eax
	addl    %ecx, %eax
	shrl    $4, %eax

on x86_64

llvm-svn: 127829
2011-03-17 20:39:14 +00:00
Cameron Zwarich
b7e676db6c The signed version of our "magic number" computation for the integer approximation
of a constant had a minor typo introduced when copying it from the book, which
caused it to favor negative approximations over positive approximations in many
cases. Positive approximations require fewer operations beyond the multiplication.

In the case of division by 3, we still generate code that is a single instruction
larger than GCC's code.

llvm-svn: 126097
2011-02-21 00:22:02 +00:00
Benjamin Kramer
6b3c3de09a Teach DAGCombine to fold fold (sra (trunc (sr x, c1)), c2) -> (trunc (sra x, c1+c2) when c1 equals the amount of bits that are truncated off.
This happens all the time when a smul is promoted to a larger type.

On x86-64 we now compile "int test(int x) { return x/10; }" into
  movslq  %edi, %rax
  imulq $1717986919, %rax, %rax
  movq  %rax, %rcx
  shrq  $63, %rcx
  sarq  $34, %rax <- used to be "shrq $32, %rax; sarl $2, %eax"
  addl  %ecx, %eax

This fires 96 times in gcc.c on x86-64.

llvm-svn: 124559
2011-01-30 16:38:43 +00:00
Dale Johannesen
e0fb87c3d7 Reapply 122353-122355 with fixes. 122354 was wrong;
the shift type was needed one place, the shift count
type another.  The transform in 123555 had the same
problem.

llvm-svn: 122366
2010-12-21 21:55:50 +00:00
Dale Johannesen
972aba543a Revert 122353-122355 for the moment, they broke stuff.
llvm-svn: 122360
2010-12-21 21:22:27 +00:00
Dale Johannesen
5f3e7b08f6 Get the type of a shift from the shift, not from its shift
count operand.  These should be the same but apparently are
not always, and this is cleaner anyway.  This improves the
code in an existing test.

llvm-svn: 122354
2010-12-21 20:06:19 +00:00
Chris Lattner
81815cd4db take care of some todos, transforming [us]mul_lohi into
a wider mul if the wider mul is legal.

llvm-svn: 121848
2010-12-15 06:04:19 +00:00
Chris Lattner
3bec2e7d0d merge two tests
llvm-svn: 121847
2010-12-15 05:58:59 +00:00
Chris Lattner
d17cbf803b rename test
llvm-svn: 121697
2010-12-13 08:39:40 +00:00