[llvm-commits] CVS: llvm/lib/Target/X86/README.txt

[Evan Cheng]

Changes in directory llvm/lib/Target/X86:

README.txt updated: 1.57 -> 1.58
---
Log message:

Some updates

---
Diffs of the changes:  (+31 -9)

 README.txt |   40 +++++++++++++++++++++++++++++++---------
 1 files changed, 31 insertions(+), 9 deletions(-)


Index: llvm/lib/Target/X86/README.txt
diff -u llvm/lib/Target/X86/README.txt:1.57 llvm/lib/Target/X86/README.txt:1.58
--- llvm/lib/Target/X86/README.txt:1.57	Thu Feb 16 22:20:13 2006
+++ llvm/lib/Target/X86/README.txt	Mon Feb 20 13:58:27 2006
@@ -39,6 +39,19 @@
 This should use fiadd on chips where it is profitable:
 double foo(double P, int *I) { return P+*I; }
 
+We have fiadd patterns now but the followings have the same cost and
+complexity. We need a way to specify the later is more profitable.
+
+def FpADD32m  : FpI<(ops RFP:$dst, RFP:$src1, f32mem:$src2), OneArgFPRW,
+                    [(set RFP:$dst, (fadd RFP:$src1,
+                                     (extloadf64f32 addr:$src2)))]>;
+                // ST(0) = ST(0) + [mem32]
+
+def FpIADD32m : FpI<(ops RFP:$dst, RFP:$src1, i32mem:$src2), OneArgFPRW,
+                    [(set RFP:$dst, (fadd RFP:$src1,
+                                     (X86fild addr:$src2, i32)))]>;
+                // ST(0) = ST(0) + [mem32int]
+
 //===---------------------------------------------------------------------===//
 
 The FP stackifier needs to be global.  Also, it should handle simple permutates
@@ -386,11 +399,6 @@
 
 //===---------------------------------------------------------------------===//
 
-The x86 backend currently supports dynamic-no-pic. Need to add asm
-printer support for static and PIC.
-
-//===---------------------------------------------------------------------===//
-
 We should generate bts/btr/etc instructions on targets where they are cheap or
 when codesize is important.  e.g., for:
 
@@ -419,10 +427,6 @@
 
 //===---------------------------------------------------------------------===//
 
-Use fisttp to do FP to integer conversion whenever it is available.
-
-//===---------------------------------------------------------------------===//
-
 Instead of the following for memset char*, 1, 10:
 
 	movl $16843009, 4(%edx)
@@ -475,3 +479,21 @@
 
 which is probably slower, but it's interesting at least :)
 
+//===---------------------------------------------------------------------===//
+
+Currently the x86 codegen isn't very good at mixing SSE and FPStack
+code:
+
+unsigned int foo(double x) { return x; }
+
+foo:
+	subl $20, %esp
+	movsd 24(%esp), %xmm0
+	movsd %xmm0, 8(%esp)
+	fldl 8(%esp)
+	fisttpll (%esp)
+	movl (%esp), %eax
+	addl $20, %esp
+	ret
+
+This will be solved when we go to a dynamic programming based isel.