From 281ae7c2cf7e2dfd48cf50b2f68fd76f7c6ab0e2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bj=C3=B6rn=20Gustavsson?= <bjorn@erlang.org>
Date: Mon, 25 Sep 2017 07:18:56 +0200
Subject: Eliminate MY_IS_SSMALL()

For a long time, there has been the two macros IS_SSMALL() and
MY_IS_SSMALL() that do exactly the same thing.

There should only be one, and it should be called IS_SSMALL().

However, we must decide which implementation to use. When
MY_IS_SSMALL() was introduced a long time ago, it was the most
efficient. In a modern C compiler, there might not be any
difference.

To find out, I used the following small C program to examine
the code generation:

    #include <stdio.h>

    typedef unsigned int Uint32;
    typedef unsigned long Uint64;
    typedef long Sint;

    #define SWORD_CONSTANT(Const) Const##L
    #define SMALL_BITS	(64-4)
    #define MAX_SMALL	((SWORD_CONSTANT(1) << (SMALL_BITS-1))-1)
    #define MIN_SMALL	(-(SWORD_CONSTANT(1) << (SMALL_BITS-1)))

    #define MY_IS_SSMALL32(x) (((Uint32) ((((x)) >> (SMALL_BITS-1)) + 1)) < 2)
    #define MY_IS_SSMALL64(x) (((Uint64) ((((x)) >> (SMALL_BITS-1)) + 1)) < 2)
    #define MY_IS_SSMALL(x) (sizeof(x) == sizeof(Uint32) ? MY_IS_SSMALL32(x) : MY_IS_SSMALL64(x))

    #define IS_SSMALL(x)      (((x) >= MIN_SMALL) && ((x) <= MAX_SMALL))

    void original(Sint n)
    {
        if (IS_SSMALL(n)) {
            printf("yes\n");
        }
    }

    void enhanced(Sint n)
    {
        if (MY_IS_SSMALL(n)) {
            printf("yes\n");
        }
    }

gcc 7.2 produced the following code for the original() function:

.LC0:
  .string "yes"
original(long):
  movabs rax, 576460752303423488
  add rdi, rax
  movabs rax, 1152921504606846975
  cmp rdi, rax
  jbe .L4
  rep ret
.L4:
  mov edi, OFFSET FLAT:.LC0
  jmp puts

clang 5.0.0 produced the following code which is slightly better:

original(long):
  movabs rax, 576460752303423488
  add rax, rdi
  shr rax, 60
  jne .LBB0_1
  mov edi, .Lstr
  jmp puts # TAILCALL
.LBB0_1:
  ret
.Lstr:
  .asciz "yes"

However, in the context of beam_emu.c, clang could produce
similar to what gcc produced.

gcc 7.2 produced the following code when MY_IS_SSMALL() was used:

.LC0:
  .string "yes"
enhanced(long):
  sar rdi, 59
  add rdi, 1
  cmp rdi, 1
  jbe .L4
  rep ret
.L4:
  mov edi, OFFSET FLAT:.LC0
  jmp puts

clang produced similar code.

This code seems to be the cheapest. There are four instructions, and
there is no loading of huge integer constants.
---
 erts/emulator/beam/big.h | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

(limited to 'erts/emulator/beam/big.h')

diff --git a/erts/emulator/beam/big.h b/erts/emulator/beam/big.h
index 48efce20e7..7556205063 100644
--- a/erts/emulator/beam/big.h
+++ b/erts/emulator/beam/big.h
@@ -70,7 +70,20 @@ typedef Uint  dsize_t;	 /* Vector size type */
 
 /* Check for small */
 #define IS_USMALL(sgn,x)  ((sgn) ? ((x) <= MAX_SMALL+1) : ((x) <= MAX_SMALL))
-#define IS_SSMALL(x)      (((x) >= MIN_SMALL) && ((x) <= MAX_SMALL))
+
+/*
+ * It seems that both clang and gcc will generate sub-optimal code
+ * for the more obvious way to write the range check:
+ *
+ *    #define IS_SSMALL(x)  (((x) >= MIN_SMALL) && ((x) <= MAX_SMALL))
+ *
+ * Note that IS_SSMALL() may be used in the 32-bit emulator with
+ * a Uint64 argument. Therefore, we must test the size of the argument
+ * to ensure that the cast does not discard the high-order 32 bits.
+ */
+#define _IS_SSMALL32(x) (((Uint32) ((((x)) >> (SMALL_BITS-1)) + 1)) < 2)
+#define _IS_SSMALL64(x) (((Uint64) ((((x)) >> (SMALL_BITS-1)) + 1)) < 2)
+#define IS_SSMALL(x) (sizeof(x) == sizeof(Uint32) ? _IS_SSMALL32(x) : _IS_SSMALL64(x))
 
 /* The heap size needed for a bignum */
 #define BIG_NEED_SIZE(x)  ((x) + 1)
-- 
cgit v1.2.3