Let's accept keeping up to 8 MiB of HeapBlock memory cached. This allows the GC allocator to reuse memory instead of asking the kernel for more.
@@ -20,7 +20,7 @@ public:
void deallocate_block(void*);
private:
- static constexpr size_t max_cached_blocks = 64;
+ static constexpr size_t max_cached_blocks = 512;
Vector<void*, max_cached_blocks> m_blocks;
};