mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-05-23 05:42:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Make pthread mutexes more scalable

Browse source 
pthread_mutex_lock() now uses a better algorithm which goes much faster
in multithreaded environments that have lock contention. This comes at
the cost of adding some fixed-cost overhead to mutex invocations. That
doesn't matter for Cosmopolitan because our core libraries all encode
locking operations as NOP instructions when in single-threaded mode.
Overhead only applies starting the moment you first call clone().
This commit is contained in:
Justine Tunney 2022-09-05 13:06:34 -07:00
parent 7de2f229a7
commit 7ff0ea8c13
No known key found for this signature in database
GPG key ID: BE714B4575D6E328
32 changed files with 410 additions and 112 deletions
Download patch file Download diff file Expand all files Collapse all files

9
test/libc/intrin/pthread_mutex_lock_test.c
View file

@ -220,12 +220,3 @@ TEST(pthread_spin_lock, test) {
EXPECT_EQ(THREADS * ITERATIONS, count);
EXPECT_EQ(0, pthread_spin_destroy(&slock));
}
BENCH(pthread_mutex_lock, bench) {
char schar = 0;
pthread_mutex_t lock = PTHREAD_MUTEX_INITIALIZER;
EZBENCH2("spin", donothing, pthread_spin_lock_test());
EZBENCH2("normal", donothing, pthread_mutex_lock_contention());
EZBENCH2("recursive", donothing, pthread_mutex_lock_rcontention());
EZBENCH2("errorcheck", donothing, pthread_mutex_lock_econtention());
}