[
https://issues.apache.org/jira/browse/DISPATCH-1372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16876728#comment-16876728
]
ASF GitHub Bot commented on DISPATCH-1372:
------------------------------------------
franz1981 commented on pull request #525: DISPATCH-1372 alloc_pool intrusive
linked list can be replaced by a linked stack
URL: https://github.com/apache/qpid-dispatch/pull/525#discussion_r299326869
##########
File path: src/alloc_pool.c
##########
@@ -44,20 +46,142 @@ DEQ_DECLARE(qd_alloc_type_t, qd_alloc_type_list_t);
#define PATTERN_BACK 0xbabecafe
struct qd_alloc_item_t {
- DEQ_LINKS(qd_alloc_item_t);
uint32_t sequence;
#ifdef QD_MEMORY_DEBUG
qd_alloc_type_desc_t *desc;
uint32_t header;
#endif
};
-DEQ_DECLARE(qd_alloc_item_t, qd_alloc_item_list_t);
+//128 has been chosen because many CPUs arch use an
+//adiacent line prefetching optimization that load
+//2*cache line bytes in batch
+#define CHUNK_SIZE 128/sizeof(void*)
+struct qd_alloc_chunk_t {
+ qd_alloc_chunk_t *prev; //do not use DEQ_LINKS here:
field position could affect access cost
+ qd_alloc_item_t *items[CHUNK_SIZE];
+ qd_alloc_chunk_t *next;
+};
+
+struct qd_alloc_linked_stack_t {
+ //the base
+ qd_alloc_chunk_t *top_chunk;
+ uint32_t top; //qd_alloc_item* top_item =
top_chunk->items[top+1] <-> top > 0
+ uint64_t size;
+ qd_alloc_chunk_t base_chunk;
+};
+
+static inline void init_stack(qd_alloc_linked_stack_t *stack)
+{
+ stack->top_chunk = &stack->base_chunk;
+ stack->top_chunk->next = NULL;
+ stack->top = 0;
+ stack->size = 0;
+}
+
+static inline void prev_chunk_stack(qd_alloc_linked_stack_t *const stack)
+{
+ assert(stack->top == 0);
+ assert(stack->size != 0);
+ assert(stack->top_chunk != &stack->base_chunk);
+ qd_alloc_chunk_t *prev = stack->top_chunk->prev;
+ //TODO(franz): stack->top_chunk could be passed externally and walked its
nexts
+ // to recycle the last chunk.
+ // Just need to pay attention to null out
released_chunk->prev->next
+ // to make it unreachable from the stack
+ stack->top_chunk = prev;
+ stack->top = CHUNK_SIZE;
+}
+
+static inline qd_alloc_item_t *pop_stack(qd_alloc_linked_stack_t *const stack)
+{
+ if (stack->top == 0) {
+ if (stack->size == 0) {
+ assert(stack->top_chunk == &stack->base_chunk);
+ return NULL;
+ }
+ prev_chunk_stack(stack);
+ }
+ stack->top--;
+ assert(stack->top >= 0 && stack->top < CHUNK_SIZE);
+ stack->size--;
+ assert(stack->size >= 0);
+ qd_alloc_item_t *item = stack->top_chunk->items[stack->top];
+ assert(item != NULL);
+ return item;
+}
+
+static inline void free_stack_chunks(qd_alloc_linked_stack_t *stack)
+{
+ assert(stack->size == 0);
+ //the assumption here is that next is always correctly set
+ qd_alloc_chunk_t *chunk = stack->base_chunk.next;
+ while (chunk != NULL) {
+ qd_alloc_chunk_t *next = chunk->next;
+ free(chunk);
+ chunk = next;
+ }
+}
+
+static inline void next_chunk_stack(qd_alloc_linked_stack_t *const stack)
+{
+ assert(stack->top == CHUNK_SIZE);
+ qd_alloc_chunk_t *top = stack->top_chunk->next;
+ if (top == NULL) {
+ top = NEW(qd_alloc_chunk_t);
Review comment:
@astitcher @kgiusti I was thiking to allocate new chunkcs aligned to cache
line size, to help `unordered_move_stack` `memcpy`s, wdyt?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> alloc_pool intrusive linked list can be replaced by a linked stack
> ------------------------------------------------------------------
>
> Key: DISPATCH-1372
> URL: https://issues.apache.org/jira/browse/DISPATCH-1372
> Project: Qpid Dispatch
> Issue Type: Improvement
> Components: Routing Engine
> Affects Versions: 1.8.0
> Reporter: Francesco Nigro
> Priority: Major
> Attachments: DOOM-3-BFG-Technical-Note.pdf,
> image-2019-06-21-11-08-17-015.png, image-2019-06-21-11-09-02-228.png,
> linked_list_misses.svg, stack_list_misses.svg
>
>
> alloc_pool is currently using a intrusive linked list approach to reduce the
> need of external data structures to hold data, saving expensive pointer
> chasing, but on modern architectures the data dependency between a current
> node and next/prev prevent the CPU prefetcher to stream nodes speculatively.
> There are different approaches that could benefit of prefetcing, but need to
> decouple the data stored from its container eg a linked stack.
> A linked stack is composed by doubly-linked chunks (allocated lazily) that
> make possible for the CPU to prefetch next/prev pointers given that those are
> already contained in the current chunk (if any).
> Although it seems counter-intuitive (given that introduce 1 more hop to reach
> the data), such data-structure is much more cache-friendly on modern
> architectures: I will attach some cache misses analysis to show it.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org
