numam-dpdk/lib/mempool
Morten Brørup b77f58604a mempool: align cache objects on cache lines
Add __rte_cache_aligned to the objs array.

It makes no difference in the general case, but if get/put operations are
always 32 objects, it will reduce the number of memory (or last level
cache) accesses from five to four 64 B cache lines for every get/put
operation.

For readability reasons, an example using 16 objects follows:

Currently, with 16 objects (128B), we access to 3
cache lines:

      ┌────────┐
      │len     │
cache │********│---
line0 │********│ ^
      │********│ |
      ├────────┤ | 16 objects
      │********│ | 128B
cache │********│ |
line1 │********│ |
      │********│ |
      ├────────┤ |
      │********│_v_
cache │        │
line2 │        │
      │        │
      └────────┘

With the alignment, it is also 3 cache lines:

      ┌────────┐
      │len     │
cache │        │
line0 │        │
      │        │
      ├────────┤---
      │********│ ^
cache │********│ |
line1 │********│ |
      │********│ |
      ├────────┤ | 16 objects
      │********│ | 128B
cache │********│ |
line2 │********│ |
      │********│ v
      └────────┘---

However, accessing the objects at the bottom of the mempool cache is a
special case, where cache line0 is also used for objects.

Consider the next burst (and any following bursts):

Current:
      ┌────────┐
      │len     │
cache │        │
line0 │        │
      │        │
      ├────────┤
      │        │
cache │        │
line1 │        │
      │        │
      ├────────┤
      │        │
cache │********│---
line2 │********│ ^
      │********│ |
      ├────────┤ | 16 objects
      │********│ | 128B
cache │********│ |
line3 │********│ |
      │********│ |
      ├────────┤ |
      │********│_v_
cache │        │
line4 │        │
      │        │
      └────────┘
4 cache lines touched, incl. line0 for len.

With the proposed alignment:
      ┌────────┐
      │len     │
cache │        │
line0 │        │
      │        │
      ├────────┤
      │        │
cache │        │
line1 │        │
      │        │
      ├────────┤
      │        │
cache │        │
line2 │        │
      │        │
      ├────────┤
      │********│---
cache │********│ ^
line3 │********│ |
      │********│ | 16 objects
      ├────────┤ | 128B
      │********│ |
cache │********│ |
line4 │********│ |
      │********│_v_
      └────────┘
Only 3 cache lines touched, incl. line0 for len.

Credits go to Olivier Matz for the nice ASCII graphics.

Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
2022-10-30 10:07:58 +01:00
..
mempool_trace_points.c lib: remove librte_ prefix from directory names 2021-04-21 14:04:09 +02:00
meson.build mempool: add telemetry endpoint 2021-10-22 22:40:59 +02:00
rte_mempool_ops_default.c lib: remove librte_ prefix from directory names 2021-04-21 14:04:09 +02:00
rte_mempool_ops.c dev: hide driver object 2022-09-23 16:14:34 +02:00
rte_mempool_trace_fp.h lib: remove librte_ prefix from directory names 2021-04-21 14:04:09 +02:00
rte_mempool_trace.h lib: remove librte_ prefix from directory names 2021-04-21 14:04:09 +02:00
rte_mempool.c mempool: fix cache flushing algorithm 2022-10-26 12:09:13 +02:00
rte_mempool.h mempool: align cache objects on cache lines 2022-10-30 10:07:58 +01:00
version.map version: 22.11-rc0 2022-07-21 12:13:48 +02:00