RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 11:42:57 +01:00
Code Issues Projects Releases Wiki Activity

move an entry, add some notes, remove a completed item (IMPLICIT_DEF)

Browse Source 
llvm-svn: 60821
This commit is contained in:
Chris Lattner 2008-12-10 01:30:48 +00:00
parent e2b5854e41
commit 3987712b2d
1 changed files with 38 additions and 37 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

75
lib/Target/README.txt
View File

@ -2,13 +2,6 @@ Target Independent Opportunities:
//===---------------------------------------------------------------------===//
We should make the various target's "IMPLICIT_DEF" instructions be a single
target-independent opcode like TargetInstrInfo::INLINEASM. This would allow
us to eliminate the TargetInstrDesc::isImplicitDef() method, and would allow
us to avoid having to define this for every target for every register class.
//===---------------------------------------------------------------------===//
With the recent changes to make the implicit def/use set explicit in
machineinstrs, we should change the target descriptions for 'call' instructions
so that the .td files don't list all the call-clobbered registers as implicit
@ -30,7 +23,10 @@ Make the PPC branch selector target independant
//===---------------------------------------------------------------------===//
Get the C front-end to expand hypot(x,y) -> llvm.sqrt(x*x+y*y) when errno and
precision don't matter (ffastmath). Misc/mandel will like this. :)
precision don't matter (ffastmath). Misc/mandel will like this. :) This isn't
safe in general, even on darwin. See the libm implementation of hypot for
examples (which special case when x/y are exactly zero to get signed zeros etc
right).
//===---------------------------------------------------------------------===//
@ -166,6 +162,9 @@ Expand these to calls of sin/cos and stores:
Doing so could allow SROA of the destination pointers. See also:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17687
This is now easily doable with MRVs. We could even make an intrinsic for this
if anyone cared enough about sincos.
//===---------------------------------------------------------------------===//
Scalar Repl cannot currently promote this testcase to 'ret long cst':
@ -511,6 +510,8 @@ int i;
}
}
BasicAA also doesn't do this for add. It needs to know that &A[i+1] != &A[i].
//===---------------------------------------------------------------------===//
We should investigate an instruction sinking pass. Consider this silly
@ -925,35 +926,6 @@ vec2d foo () {
//===---------------------------------------------------------------------===//
This C++ file:
void g(); struct A { int n; int m; A& operator++(void) { ++n; if (n == m) g();
return *this; } A() : n(0), m(0) { } friend bool operator!=(A const& a1,
A const& a2) { return a1.n != a2.n; } }; void testfunction(A& iter) { A const
end; while (iter != end) ++iter; }
Compiles down to:
bb: ; preds = %bb3.backedge, %bb.nph
%.rle = phi i32 [ %1, %bb.nph ], [ %7, %bb3.backedge ] ; <i32> [#uses=1]
%4 = add i32 %.rle, 1 ; <i32> [#uses=2]
store i32 %4, i32* %0, align 4
%5 = load i32* %3, align 4 ; <i32> [#uses=1]
%6 = icmp eq i32 %4, %5 ; <i1> [#uses=1]
br i1 %6, label %bb1, label %bb3.backedge
bb1: ; preds = %bb
tail call void @_Z1gv()
br label %bb3.backedge
bb3.backedge: ; preds = %bb, %bb1
%7 = load i32* %0, align 4 ; <i32> [#uses=2]
The %7 load is partially redundant with the store of %4 to %0, GVN's PRE
should remove it, but it doesn't apply to memory objects.
//===---------------------------------------------------------------------===//
Better mod/ref analysis for scanf would allow us to eliminate the vtable and a
bunch of other stuff from this example (see PR1604):
@ -1432,3 +1404,32 @@ void foo (int a, struct T b)
//===---------------------------------------------------------------------===//
This C++ file:
void g(); struct A { int n; int m; A& operator++(void) { ++n; if (n == m) g();
return *this; } A() : n(0), m(0) { } friend bool operator!=(A const& a1,
A const& a2) { return a1.n != a2.n; } }; void testfunction(A& iter) { A const
end; while (iter != end) ++iter; }
Compiles down to:
bb: ; preds = %bb3.backedge, %bb.nph
%.rle = phi i32 [ %1, %bb.nph ], [ %7, %bb3.backedge ] ; <i32> [#uses=1]
%4 = add i32 %.rle, 1 ; <i32> [#uses=2]
store i32 %4, i32* %0, align 4
%5 = load i32* %3, align 4 ; <i32> [#uses=1]
%6 = icmp eq i32 %4, %5 ; <i1> [#uses=1]
br i1 %6, label %bb1, label %bb3.backedge
bb1: ; preds = %bb
tail call void @_Z1gv()
br label %bb3.backedge
bb3.backedge: ; preds = %bb, %bb1
%7 = load i32* %0, align 4 ; <i32> [#uses=2]
The %7 load is partially redundant with the store of %4 to %0, GVN's PRE
should remove it, but it doesn't apply to memory objects.
//===---------------------------------------------------------------------===//