project-base/server-skynet-source-3rd-jemalloc
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Improve rtree cache with a two-level cache design.

Browse Source 
Two levels of rcache is implemented: a direct mapped cache as L1, combined with
a LRU cache as L2.  The L1 cache offers low cost on cache hit, but could suffer
collision under circumstances.  This is complemented by the L2 LRU cache, which
is slower on cache access (overhead from linear search + reordering), but solves
collison of L1 rather well.
This commit is contained in:
Qi Wang
2017-04-14 11:05:38 -07:00
committed by Qi Wang
parent d16f1e53df
commit 3c9c41edb2
5 changed files with 97 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
src/prof.c
View File

@@ -310,6 +310,7 @@ prof_leave(tsd_t *tsd, prof_tdata_t *tdata) {
}
#ifdef JEMALLOC_PROF_LIBUNWIND
JEMALLOC_ALIGNED(CACHELINE)
void
prof_backtrace(prof_bt_t *bt) {
int nframes;

37
src/rtree.c
View File

@@ -256,6 +256,16 @@ rtree_leaf_elm_lookup_hard(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
leaf = rtree->root;
#endif
if (config_debug) {
uintptr_t leafkey = rtree_leafkey(key);
for (unsigned i = 0; i < RTREE_CTX_NCACHE; i++) {
assert(rtree_ctx->cache[i].leafkey != leafkey);
}
for (unsigned i = 0; i < RTREE_CTX_NCACHE_L2; i++) {
assert(rtree_ctx->l2_cache[i].leafkey != leafkey);
}
}
#define RTREE_GET_CHILD(level) { \
assert(level < RTREE_HEIGHT-1); \
if (level != 0 && !dependent && \
@@ -277,20 +287,30 @@ rtree_leaf_elm_lookup_hard(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
dependent); \
} \
}
/*
* Cache replacement upon hard lookup (i.e. L1 & L2 rtree cache miss):
* (1) evict last entry in L2 cache; (2) move the collision slot from L1
* cache down to L2; and 3) fill L1.
*/
#define RTREE_GET_LEAF(level) { \
assert(level == RTREE_HEIGHT-1); \
if (!dependent && unlikely(!rtree_leaf_valid(leaf))) { \
return NULL; \
} \
if (RTREE_CTX_NCACHE > 1) { \
memmove(&rtree_ctx->cache[1], \
&rtree_ctx->cache[0], \
if (RTREE_CTX_NCACHE_L2 > 1) { \
memmove(&rtree_ctx->l2_cache[1], \
&rtree_ctx->l2_cache[0], \
sizeof(rtree_ctx_cache_elm_t) * \
(RTREE_CTX_NCACHE-1)); \
(RTREE_CTX_NCACHE_L2 - 1)); \
} \
size_t slot = rtree_cache_direct_map(key); \
rtree_ctx->l2_cache[0].leafkey = \
rtree_ctx->cache[slot].leafkey; \
rtree_ctx->l2_cache[0].leaf = \
rtree_ctx->cache[slot].leaf; \
uintptr_t leafkey = rtree_leafkey(key); \
rtree_ctx->cache[0].leafkey = leafkey; \
rtree_ctx->cache[0].leaf = leaf; \
rtree_ctx->cache[slot].leafkey = leafkey; \
rtree_ctx->cache[slot].leaf = leaf; \
uintptr_t subkey = rtree_subkey(key, level); \
return &leaf[subkey]; \
}
@@ -433,6 +453,11 @@ rtree_ctx_data_init(rtree_ctx_t *ctx) {
cache->leafkey = RTREE_LEAFKEY_INVALID;
cache->leaf = NULL;
}
for (unsigned i = 0; i < RTREE_CTX_NCACHE_L2; i++) {
rtree_ctx_cache_elm_t *cache = &ctx->l2_cache[i];
cache->leafkey = RTREE_LEAFKEY_INVALID;
cache->leaf = NULL;
}
}
bool