M

MultiDomain Cache Daemon

Settings | Report Duplicate

0

I Use This!

Inactive

Commits : Listings

Analyzed about 4 hours ago. based on code collected about 5 hours ago.

Commit Message	Contributor	Files Modified	Lines Added	Lines Removed	Code Location	Date
May 14, 2023 — May 14, 2024 Showing page 2 of 4 Search / Filter on:
High-performance async benchmark client	ivoras	More...				about 16 years ago
Fix network thread binding if there are no worker threads	ivoras	More...				about 16 years ago
Code cleanup after cqueue removal	ivoras	More...				about 16 years ago
Revert cqueue work - apparently the library is still unstable	ivoras	More...				about 16 years ago
Add locking around struct md_conn	ivoras	More...				about 16 years ago
Catch packet chaining earlier	ivoras	More...				about 16 years ago
More verbose log message for packet chaining	ivoras	More...				about 16 years ago
Don't rely on unlocked variables when testing lockers	ivoras	More...				about 16 years ago
Converted server to use cqueue instead of linked lists (expecting mild performance increase)	ivoras	More...				about 16 years ago
Circular queue implementation & test	ivoras	More...				about 16 years ago
Circular queue implementation	ivoras	More...				about 16 years ago
Code cleanup	ivoras	More...				about 16 years ago
Introduce binding of network threads to user threads.	ivoras	More...				about 16 years ago
Update cache_dump_running flag	ivoras	More...				about 16 years ago
Made cache dump thread detached	ivoras	More...				about 16 years ago
Cache dump on SIGUSR1 implemented	ivoras	More...				about 16 years ago
Attempt at optimization: When selecting a worker for the network thread iterations, spin through worker threads until an available worker is found.	ivoras	More...				about 16 years ago
Bound network thread iterations to worker threads	ivoras	More...				about 16 years ago
(non-working) worker mutex spin variant	ivoras	More...				about 16 years ago
Early implementation of the "second" design : MPED-2	ivoras	More...				about 16 years ago
Design notes	ivoras	More...				about 16 years ago
Increase listen(2) backlog also for unix sockets	ivoras	More...				about 16 years ago
Small code rearragements	ivoras	More...				about 16 years ago
Another fixup in ev_remove_fd: remove both EVFILT_READ and EVFILT_WRITE events.	ivoras	More...				about 16 years ago
Fix important kevent() call typo at ev_remove_fd	ivoras	More...				about 16 years ago
Minilibevent conversion	ivoras	More...				about 16 years ago
Modified server to use minilibevent	ivoras	More...				about 16 years ago
"Mini libevent" - a lightweight header-library for async IO	ivoras	More...				about 16 years ago
Sprinkle documentation comments	ivoras	More...				over 16 years ago
Added memcached benchmark, mimicking as exactly as possible the mdcached benchmark. Here's the preliminary comparison of performance on dual-core Athlon64, i386 mode, FreeBSD 8 (both client and server on the same system): mdcached: 2 worker threads, 2 client threads, unix sockets: 63800 ops/s memcached: threadless, 2 client threads, TCP/localhost: 35100 ops/s With two 4-core systems connected via gigabit Ethernet (one client, one server), mdcached can perform > 110,000 ops/s.	ivoras	More...				over 16 years ago

←
1
2
3
4
→