2446 Commits (310e49e7f34d7f19a3c04e0af5c96e7a75716e5c)

 

All Branches   

Search

Author	SHA1	Message	Date
winlin	6a49ae4ffb	update readme, for 10k performance	10 years ago
winlin	a3de1c71fc	reset the mw_latency to 350.	10 years ago
winlin	ffacf1b774	update 2.0 history for 1.0 released.	10 years ago
winlin	8c1023d816	fix #251, 10k+ clients, use queue cond wait and fast vector. 2.0.67	10 years ago
winlin	71155fc3d2	update readme for 1.0release	10 years ago
winlin	ed93c50218	update readme. 10k clients.	10 years ago
winlin	1311b6fe65	fix #251, support 10k clients. 2.0.67	10 years ago
winlin	55d98fceed	for bug #251, add min msgs for queue cond wait.	10 years ago
winlin	438ee30ac1	for bug #251, refine the queue fast vector. 2.0.66	10 years ago
winlin	5a2b8afead	for bug #251, add queue fast vector. 2.0.66	10 years ago
winlin	8f72f79504	for bug #251, add queue fast cache. 2.0.65	10 years ago
winlin	1bbdae6632	refine code.	10 years ago
winlin	bba6063492	for bug #251, the shared ptr message share the header. 2.0.64	10 years ago
winlin	330819fb74	for bug #251, refine the mic algorithm. 2.0.63	10 years ago
winlin	d53fd7f570	for bug #251, support mic(message iovs cache). 2.0.61	10 years ago
winlin	32d537b96b	for bug #251, refine code before mic.	10 years ago
winlin	d827928eeb	for bug #251, merge the performance refines.	10 years ago
winlin	a6599f2250	remove the qt modify file.	10 years ago
winlin	9892b92258	Revert "for bug #251, somhc(session-oriented message-header cache). 2.0.61" ... It hurt performance, should never use it. This reverts commit d073adde58.	10 years ago
winlin	d073adde58	for bug #251, somhc(session-oriented message-header cache). 2.0.61	10 years ago
winlin	de993b6465	for bug #251, refine the directly access ptrs for hls,dvr,forwarder,consumer.	10 years ago
winlin	d3a103184a	refine protocol sdk send msg, the chunk header generate.	10 years ago
winlin	f9b9a60de7	for bug #251, remove the SrsMessage, use SrsCommonMessage or SrsSharedPtrMessage. 2.0.60	10 years ago
winlin	528ae1e9b1	refine source and queue dump msgs.	10 years ago
winlin	e80c8603d4	fix #251, revert changes, for the cond wait and fast cache queue is no use. 2.0.59	10 years ago
winlin	92ecdf088b	for bug #251, use macro to define the fast cache and cond wait. 2.0.58	10 years ago
winlin	8aff915538	release SRS 1.0	10 years ago
winlin	cdb1984ba4	update readme	10 years ago
winlin	cd317859db	add mw sleep time and msgs/iovs table.	10 years ago
winlin	9ee138746f	for bug #251, 9k+ clients, use fast cache for msgs queue. 2.0.57	10 years ago
winlin	dde05c6315	for bug #251, refine the send use cond wait.	10 years ago
winlin	4c1d5c0d1e	refine the log for mw and mr. 2.0.55	10 years ago
winlin	fad6074a25	set send socket buf. mw/mr sleep default to 350.	10 years ago
winlin	4d1c162ebd	fix utest failed bug.	10 years ago
winlin	4c6cf959c1	refine fast buffer, check available bytes to read.	10 years ago
winlin	98647d6e67	limit the user-space buffer size to 128KB, 128MB for 1k publishers.	10 years ago
winlin	76af04c55d	refine the recv buffer for mr.	10 years ago
winlin	09101b8d5e	remove the increase timeout for read, for there is no small timeout.	10 years ago
winlin	279b4ff4c9	refine macros	10 years ago
winlin	b1d7fbe668	fix #241, add mw(merged-write) config. 2.0.53	10 years ago
winlin	72fa33d9fb	update performance to 4k, when mr_sleep to 2000ms. 2.0.52	10 years ago
winlin	f1192a8302	fix bug for bug #241, default to 32k for socket buffer.	10 years ago
winlin	5589b13d2e	for bug #241, support mr(merged-read) config and reload. 2.0.52.	10 years ago
winlin	57f844b636	for bug #249, simplify the macro, 0 to disable chunk stream cache.	10 years ago
winlin	159ef3823c	for bug#241, simplify the buffer, donot realloc it.	10 years ago
winlin	a23b5b9a4b	for bug#241, simplify the buffer, donot change its size.	10 years ago
winlin	d026861e23	for bug #241, simplify the buffer, donot reset when chunk size change.	10 years ago
winlin	4b09531e2f	for bug #241, simplify the merged read config macros.	10 years ago
winlin	310f51e6a1	update readme.	10 years ago
winlin	f57801eb46	fix #249, cache the chunk headers info to +5% or +10% performance. 2.0.51	10 years ago