[prev in list] [next in list] [prev in thread] [next in thread]
List: llvm-commits
Subject: [llvm-commits] [llvm] r60286 - /llvm/trunk/lib/Target/X86/README.txt
From: Eli Friedman <eli.friedman () gmail ! com>
Date: 2008-11-30 7:52:27
Message-ID: 200811300752.mAU7qSoL016414 () zion ! cs ! uiuc ! edu
[Download RAW message or body]
Author: efriedma
Date: Sun Nov 30 01:52:27 2008
New Revision: 60286
URL: http://llvm.org/viewvc/llvm-project?rev=60286&view=rev
Log:
A couple small cleanups, plus a new potential optimization.
Modified:
llvm/trunk/lib/Target/X86/README.txt
Modified: llvm/trunk/lib/Target/X86/README.txt
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/README.txt?rev=60286&r1=60285&r2=60286&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/README.txt (original)
+++ llvm/trunk/lib/Target/X86/README.txt Sun Nov 30 01:52:27 2008
@@ -394,6 +394,10 @@
However, if we care more about code size, then imull is better. It's two bytes
shorter than movl + leal.
+On a Pentium M, both variants have the same characteristics with regard
+to throughput; however, the multiplication has a latency of four cycles, as
+opposed to two cycles for the movl+lea variant.
+
//===---------------------------------------------------------------------===//
__builtin_ffs codegen is messy.
@@ -628,9 +632,9 @@
to grab the bytes from the next cacheline.
532 532 0x3cfc movb (1809(%esp, %esi), %bl <<<--- spans 2 64 byte lines
-942 942 0x3d03 movl %dh, (1809(%esp, %esi) \
-937 937 0x3d0a incl %esi
-3 3 0x3d0b cmpb %bl, %dl
+942 942 0x3d03 movl %dh, (1809(%esp, %esi)
+937 937 0x3d0a incl %esi
+3 3 0x3d0b cmpb %bl, %dl
27 27 0x3d0d jnz 0x000062db <main+11707>
//===---------------------------------------------------------------------===//
@@ -1790,3 +1794,25 @@
(v8i8 (MMX_MOVDQ2Qrr VR128:$src))>;
There are other cases in various td files.
+
+//===---------------------------------------------------------------------===//
+
+Take something like the following on x86-32:
+unsigned a(unsigned long long x, unsigned y) {return x % y;}
+
+We currently generate a libcall, but we really shouldn't: the expansion is
+shorter and likely faster than the libcall. The expected code is something
+like the following:
+
+ movl 12(%ebp), %eax
+ movl 16(%ebp), %ecx
+ xorl %edx, %edx
+ divl %ecx
+ movl 8(%ebp), %eax
+ divl %ecx
+ movl %edx, %eax
+ ret
+
+A similar code sequence works for division.
+
+//===---------------------------------------------------------------------===//
_______________________________________________
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic