boost/fiber
2
0
Fork 0
mirror of https://github.com/boostorg/fiber.git synced 2026-02-18 01:52:24 +00:00
Code Issues Projects Releases Wiki Activity

Proofread new discussion of fiber migration in Rationale.

Browse Source
This commit is contained in:
Nat Goodspeed
2015-08-13 07:38:30 -04:00
parent 5112f9b4b5
commit 4d36df2085
1 changed files with 11 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
doc/rational.qbk
View File

@@ -48,17 +48,18 @@ this description].)
[heading migrating fibers between threads]
The support for migrating fibers between threads was removed.
Especially in the case of NUMA-architectures it is not always advised to migrate
data between threads. Suppose fiber ['f] is running on logical CPU ['cpu0] which
belongs to NUMA node ['node0]. The data of ['f] are allocated on the physical
memory located at ['node0]. Migrating the fiber from ['cpu0] to another logical
CPU ['cpuX] which is part of NUMA node ['nodeX] will reduce the performance of
the application because of increased latency of memory access.
Support for migrating fibers between threads was removed.
Especially in the case of NUMA-architectures, it is not always advisable to
migrate data between threads. Suppose fiber ['f] is running on logical CPU
['cpu0] which belongs to NUMA node ['node0]. The data of ['f] are allocated on
the physical memory located at ['node0]. Migrating the fiber from ['cpu0] to
another logical CPU ['cpuX] which is part of a different NUMA node ['nodeX]
will reduce the performance of the application because of increased latency of
memory access.
A more [*important] apsect is the problem with thread-local-storage([*TLS]).
Compilers might, as an optimization, reuse a previously computed address of an
TLS variable, instead to recomputing the address again
A more [*important] aspect is the problem with thread-local-storage ([*TLS]).
Instead of recomputing the address of a TLS variable, a compiler might, as an
optimization, reuse its previously-computed address
[footnote dicussed in [@http://www.crystalclearsoftware.com/soc/coroutine/coroutine/coroutine_thread.html
'Interaction between coroutines and threads' by Giovanni P. Deretta,
[@https://groups.google.com/a/isocpp.org/forum/#!topic/std-proposals/3g6ZIWedGJ8