Add prefetching on strings before hashing

Turns out that the hotspots of nasm are mainly on string hashing
when accessing memory. A simple performance improvement is to
prefetch the first cacheline of a string to be hashed. Ran 50 tests
on an i9-12900 building intel-ipsec-mb that heavily uses nasm and
improved wall clock build times from 56.1 seconds to 53.2 seconds or
around 5% speed improvement.

Signed-off-by: Colin Ian King <colin.i.king@intel.com>

Author: Colin Ian King

nasmlib/hashtbl.c

@@ -80,6 +80,8 @@ void **hash_findb(struct hash_table *head, const void *key,
 void **hash_find(struct hash_table *head, const char *key,
                  struct hash_insert *insert)
 {
+    __builtin_prefetch(key);
+
     return hash_findb(head, key, strlen(key)+1, insert);
 }
 
@@ -124,6 +126,8 @@ void **hash_findib(struct hash_table *head, const void *key, size_t keylen,
 void **hash_findi(struct hash_table *head, const char *key,
                   struct hash_insert *insert)
 {
+    __builtin_prefetch(key);
+
     return hash_findib(head, key, strlen(key)+1, insert);
 }
 

nasmlib/strlist.c

@@ -48,6 +48,8 @@ strlist_add(struct strlist *list, const char *str)
 	struct hash_insert hi;
 	size_t size;
 
+	__builtin_prefetch(str);
+
 	if (!list)
 		return NULL;