48 Commits

Author	SHA1	Message	Date
Felipe Petroski Such	38cd87cdcc	add memory region functions	2023-04-07 15:38:48 -07:00
Changho Hwang	949a9cd0a3	Optional use of gdrcopy (#48) ... Co-authored-by: Saeed Maleki <saemal@microsoft.com>	2023-04-07 13:36:59 +08:00
Saeed Maleki	cd3cd2c157	lint	2023-04-06 03:20:21 +00:00
Saeed Maleki	08275e93d7	added barrier API + pushed one after mscclppsetup	2023-04-06 03:15:54 +00:00
Saeed Maleki	1731911d00	removing extra stream and destroying created ones	2023-04-02 02:07:41 +00:00
Saeed Maleki	4c6616e7b9	lint	2023-04-01 19:20:50 +00:00
Saeed Maleki	8927dd4d72	great allgather numbers with the current binding mechanism	2023-04-01 18:54:42 +00:00
Binyang Li	af5825b474	bind numa node to communicator	2023-03-31 08:05:49 +00:00
Saeed Maleki	be5e422021	merged with main	2023-03-29 23:03:12 +00:00
Binyang2014	62279b0063	Add mscclppSetBootstrapConnTimeout (#34)	2023-03-28 14:01:56 +08:00
Saeed Maleki	33af4bfb67	no gdr copy anywhere in the code except for the files that are not compiled	2023-03-28 05:36:31 +00:00
Saeed Maleki	d9ba953fb0	gdrcopy is not initialized	2023-03-28 04:56:06 +00:00
Saeed Maleki	e7cccbf897	both head and tail are on OK to be only used by GPU	2023-03-28 04:26:39 +00:00
Saeed Maleki	952d852256	both head and tail are on OK to be only used by GPU	2023-03-28 04:24:45 +00:00
Saeed Maleki	43c52367fb	merged with main and simplified the callback requirements	2023-03-27 23:41:27 +00:00
Saeed Maleki	19bf369dc1	link format correction	2023-03-27 20:40:15 +00:00
Changho Hwang	8fc8f5b4fe	Lint	2023-03-27 14:09:26 +00:00
Changho Hwang	8e4146aba9	Add mscclppSetLogHandler	2023-03-27 13:33:07 +00:00
Saeed Maleki	0898214f0a	added mscclppGetErrorString	2023-03-24 22:57:14 +00:00
Saeed Maleki	3fb9383621	Merge pull request #24 from microsoft/madanm-apipush ... simplified API for CUDA level communication calls.	2023-03-24 10:42:07 -07:00
Saeed Maleki	56b599b5e7	a bit of api change and clean up on docs	2023-03-24 17:41:04 +00:00
Changho Hwang	551eae0ba1	Update docs	2023-03-24 09:28:12 +00:00
Ziyue Yang	f92b428cba	Port NPKit	2023-03-24 06:41:16 +00:00
Changho Hwang	05fde6c6f3	minor changes	2023-03-24 04:51:20 +00:00
Saeed Maleki	777e93ee47	merged with main	2023-03-24 02:35:15 +00:00
Madan Musuvathi	e6ee81e4fa	fixed the order of remote rank and tag in mscclppConnect API	2023-03-23 21:09:04 +00:00
Madan Musuvathi	72edabe2a6	added GetDevConn api to retrieve a connection from remoteRank and tag	2023-03-23 21:03:30 +00:00
Madan Musuvathi	896539b236	Comm owns all state including devcons	2023-03-22 22:43:32 +00:00
Saeed Maleki	270839797e	Merge branch 'main' into chhwang/dealloc	2023-03-22 21:14:42 +00:00
Madan Musuvathi	4c459aa0df	allgather_test code cleanup	2023-03-22 20:38:29 +00:00
Changho Hwang	48a23243a4	Dealloc more resources	2023-03-22 12:06:35 +00:00
Changho Hwang	9f2eef35d3	Init from a given mscclppUniqueId	2023-03-22 11:25:49 +00:00
Saeed Maleki	483b0c8433	flag is now allocated by the system	2023-03-22 05:14:24 +00:00
Saeed Maleki	b75f9e6d8a	implementing new API	2023-03-22 00:29:10 +00:00
Olli Saarikivi	0cfe2dcffb	Add allpairs allreduce test ... To support this include separate source and destination offsets in the trigger. Add functions for getting the rank and world size from a communicator.	2023-03-21 19:00:13 +00:00
Saeed Maleki	5493e22633	fixed multinode bug	2023-03-19 06:09:07 +00:00
Saeed Maleki	a485a7f238	single node works fine -- multinode is problematic	2023-03-19 01:08:05 +00:00
Saeed Maleki	9cc21f70e6	redesigning fifo	2023-03-17 22:51:11 +00:00
Changho Hwang	67dbbd1692	Thread-safe trigger	2023-03-17 09:46:23 +00:00
Changho Hwang	dc41c58769	Alloc proxy states on demand	2023-03-14 10:05:56 +00:00
Changho Hwang	e357beef00	One fifo per proxy	2023-03-13 14:19:36 +00:00
Saeed Maleki (saemal)	dced4c4c14	done with the design	2023-03-06 12:26:46 -08:00
Saeed Maleki (saemal)	b663469bcd	making a fifo for proxy threads	2023-03-06 10:58:12 -08:00
Saeed Maleki	0216ceb34e	added todos	2023-03-06 08:04:33 +00:00
Changho Hwang	9e5573f16b	Misc changes and comments	2023-03-03 08:32:47 +00:00
Changho Hwang	9674830db2	Change flags into uint64_t	2023-03-03 07:41:34 +00:00
Changho Hwang	3d051a985f	Add p2p proxy: doesn't work with cuda graph yet	2023-02-28 18:58:03 +00:00
Changho Hwang	48b81edf6d	Move some files	2023-02-22 11:07:22 +00:00