From 330482177c45dceb71e5787c1a7e7fc87308c144 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Fri, 7 Oct 2016 00:04:28 +0200 Subject: [PATCH] ralloc: don't memset ralloc_header, clear it manually MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit time GALLIUM_NOOP=1 ./run shaders/private/alien_isolation/ >/dev/null Before (2 takes): real 0m8.734s 0m8.773s user 0m34.232s 0m34.348s sys 0m0.084s 0m0.056s After (2 takes): real 0m8.448s 0m8.463s user 0m33.104s 0m33.160s sys 0m0.088s 0m0.076s Average change in "real" time spent: -3.4% calloc should only do 2 things compared to malloc: - check for overflow of "n * size" - call memset I'm not sure if that explains the difference. v2: clear "parent" and "next" in the caller of add_child. Reviewed-by: Edward O'Callaghan (v1) Tested-by: Edmondo Tommasina (v1) Reviewed-by: Nicolai Hähnle (v1) --- src/util/ralloc.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/src/util/ralloc.c b/src/util/ralloc.c index 9526011..7023d79 100644 --- a/src/util/ralloc.c +++ b/src/util/ralloc.c @@ -122,13 +122,27 @@ ralloc_size(const void *ctx, size_t size) void * rzalloc_size(const void *ctx, size_t size) { - void *block = calloc(1, size + sizeof(ralloc_header)); + void *block = malloc(size + sizeof(ralloc_header)); ralloc_header *info; ralloc_header *parent; if (unlikely(block == NULL)) return NULL; + info = (ralloc_header *) block; + /* measurements have shown that calloc is slower (because of + * the multiplication overflow checking?), so clear things + * manually + */ + info->parent = NULL; + info->child = NULL; + info->prev = NULL; + info->next = NULL; + info->destructor = NULL; + + /* memset the allocation except for ralloc_header */ + memset(&info[1], 0, size); + parent = ctx != NULL ? get_header(ctx) : NULL; add_child(parent, info); -- 2.7.4