From 2e3f22fc115c75cc76bff7ce1b902e4923484813 Mon Sep 17 00:00:00 2001
From: Per Hedbor <ph@opera.com>
Date: Sat, 16 Jun 2001 15:49:40 +0200
Subject: [PATCH] Added optimized case for (int)string when string->len < 10 &&
string->size_shift == 1. This speed up the sum a column of integers test on
the language shootout site by 30%.
Rev: src/opcodes.c:1.110
---
src/opcodes.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/src/opcodes.c b/src/opcodes.c
index 1c82e60ebd..ceed10830c 100644
--- a/src/opcodes.c
+++ b/src/opcodes.c
@@ -27,7 +27,7 @@
#include "bignum.h"
#include "operators.h"
-RCSID("$Id: opcodes.c,v 1.109 2001/06/05 10:12:55 hubbe Exp $");
+RCSID("$Id: opcodes.c,v 1.110 2001/06/16 13:49:40 per Exp $");
void index_no_free(struct svalue *to,struct svalue *what,struct svalue *ind)
{
@@ -269,7 +269,21 @@ void o_cast(struct pike_type *type, INT32 run_time_type)
we really want to reduce the number of number parsers
around here. :) /Noring */
#ifdef AUTO_BIGNUM
- convert_stack_top_string_to_inumber(10);
+
+
+ /* The generic function is rather slow, so I added this
+ * code for benchmark purposes. :-) /per
+ */
+ if( sp[-1].u.string->len < 10 &&
+ !sp[-1].u.string->size_shift )
+ {
+ i=atoi(sp[-1].u.string->str);
+ free_string(sp[-1].u.string);
+ sp[-1].type=T_INT;
+ sp[-1].u.integer=i;
+ }
+ else
+ convert_stack_top_string_to_inumber(10);
return; /* FIXME: OK to return? Cast tests below indicates
we have to do this, at least for now... /Noring */
/* Yes, it is ok to return, it is actually an optimization :)
--
GitLab