Optimize clear_brick_table. (#65837)

PeterSolMS · web-flow · commit 02e46674ccf3 · 2022-03-01T17:34:00.000+01:00
I observed that the inlined call to clear_brick_table in clear_region_info took more CPU samples than necessary - it's about 7x faster to call memset than it is to code a straighforward loop.
diff --git a/src/coreclr/gc/gc.cpp b/src/coreclr/gc/gc.cpp
@@ -8034,8 +8034,9 @@ uint8_t* gc_heap::brick_address (size_t brick)
 
 void gc_heap::clear_brick_table (uint8_t* from, uint8_t* end)
 {
-    for (size_t i = brick_of (from);i < brick_of (end); i++)
-        brick_table[i] = 0;
+    size_t from_brick = brick_of (from);
+    size_t end_brick = brick_of (end);
+    memset (&brick_table[from_brick], 0, sizeof(brick_table[from_brick])*(end_brick-from_brick));
 }
 
 //codes for the brick entries:

Original file line number	Diff line number	Diff line change
`@@ -8034,8 +8034,9 @@ uint8_t* gc_heap::brick_address (size_t brick)`
`8034`	`8034`
`8035`	`8035`	`void gc_heap::clear_brick_table (uint8_t* from, uint8_t* end)`
`8036`	`8036`	`{`
`8037`		`- for (size_t i = brick_of (from);i < brick_of (end); i++)`
`8038`		`- brick_table[i] = 0;`
	`8037`	`+ size_t from_brick = brick_of (from);`
	`8038`	`+ size_t end_brick = brick_of (end);`
	`8039`	`+ memset (&brick_table[from_brick], 0, sizeof(brick_table[from_brick])*(end_brick-from_brick));`
`8039`	`8040`	`}`
`8040`	`8041`
`8041`	`8042`	`//codes for the brick entries:`