Chiebot-Mirror/grpc - grpc - Gitea: Git with a cup of tea

Commit Graph

Author	SHA1	Message	Date
Na-Na Pang	a2daa4ff08	Clean format'	6 years ago
Na-Na Pang	1ea651aee3	Add assertion	6 years ago
Na-Na Pang	762e58b574	Change client context allocation	6 years ago
Na-Na Pang	070902b871	Merge bm_callback_cq to bm_cq	6 years ago
Na-Na Pang	3fc702510f	Reuse reactor to send new RPC	6 years ago
Na-Na Pang	2d5a9750a0	Manually add echo.proto to pass Portability build test	6 years ago
Na-Na Pang	1ba5f5c701	Modify build file	6 years ago
Na-Na Pang	32e10e618a	address the reference arguments	6 years ago
Na-Na Pang	714e13b426	Delete log	6 years ago
Na-Na Pang	c905f76a5b	Clang format	6 years ago
Na-Na Pang	2787dedd70	Modify dependency of callback_test_service	6 years ago
Na-Na Pang	927c2f2c61	Change format	6 years ago
Na-Na Pang	5748665bc5	Add callback completion queue and modify callback streaming ping pong	6 years ago
Esun Kim	e18ed03c04	Made gRPC inialized after entering main function in microbenchmarks.	6 years ago
Na-Na Pang	25128d18c1	Modify unary ping pong to send next rpc in callback function	6 years ago
Na-Na Pang	7d3fdec445	Add microbenchmark for callback completion queue	6 years ago
Na-Na Pang	e70d507abe	Changes based on comment	6 years ago
Na-Na Pang	b028141f01	Change streaming ping pong args and add comment	6 years ago
Na-Na Pang	8bf138d799	Add copyright	6 years ago
Na-Na Pang	875d2df399	Modify build file	6 years ago
Na-Na Pang	bbd4eb5028	Add microbenchmark for callback unary ping pong and bidistreaming ping pong	6 years ago
Yash Tibrewal	9345b7e276	s/atleast/at\ least	6 years ago
Arjun Roy	8ce42f67b2	Shrink arena size by 40 bytes and add additional alignment options (for cache-alignment). We shrink by: 1) Removing an unnecessary zone pointer. 2) Replacing gpr_mu (40 bytes when using pthread_mutex_t) with std::atomic_flag. We also header-inline the fastpath alloc (ie. when not doing a zone alloc) and move the malloc() for a zone alloc outside of the mutex critical zone, which allows us to replace the mutex with a spinlock. We also cache-align created arenas.	6 years ago
Yash Tibrewal	a2316142b2	Reviewer comments	6 years ago
Mark D. Roth	ce3ff86763	Convert call_combiner to C++.	6 years ago
Karthik Ravi Shankar	858b5cca20	Fix clang_format_code.sh issues and move the internal calls to new name	6 years ago
Yash Tibrewal	4309a98b66	Health checking service name to be passed as a channel arg for now	6 years ago
Nicolas "Pixel" Noble	b9a279c030	Resolving ambiguous call to CreateCustomChannel.	6 years ago
Yash Tibrewal	116ce0fb24	Use health check parser	6 years ago
Karthik Ravi Shankar	973fa6950e	Fold Channel into grpc_impl from grpc namespace	6 years ago
Karthik Ravi Shankar	dad3f34207	Revert "Revert "Revert "Folding the Channel class into the grpc_impl namespace."""	6 years ago
Guantao Liu	872d2787a0	Avoid using grpc_core::Executor when the background poller is available. Instead, run closures in the background poller. This will generally avoid the thread hop in the gRPC runtime.	6 years ago
Karthik Ravi Shankar	4e0923e802	Fix errors from clang_format_code.sh	6 years ago
Soheil Hassas Yeganeh	1014fe507f	Use const ref for grpc_slice. We are copying the slice on every call creation, which is hurting ping/pong traffic.	6 years ago
Bill Feng	d93959853f	Enabled Windows Bazel build for cpp tests	6 years ago
Karthik Ravi Shankar	4260fe1147	More fixes	6 years ago
Soheil Hassas Yeganeh	ad1b3e5094	Introduce grpc_byte_buffer_reader_peek and use it for Protobuf parsing. grpc_byte_buffer_reader_next() copies and references the slice. This is not always necessary since the caller will not use the slice after destroying the byte buffer. A prominent example is the protobuf parser, which calls grpc_byte_buffer_reader_next() and immediately unrefs the slice after the call. This ref() and unref() calls can be very expensive in the hot path. This commit introduces grpc_byte_buffer_reader_peek() which essentialy return a pointer to the slice in the buffer, i.e., no copies, and no refs. QPS of 1MiB 1 Channel callback benchmark increases by 5%. More importantly insructions per cycle is increased by 10%. Also add tests and benchmarks for byte_buffer_reader_peek() This commit reaplies `509e77a5a3`	6 years ago
Karthik Ravi Shankar	2a8f3f79ab	Fix more namespace stuff	6 years ago
Karthik Ravi Shankar	f66b654795	Revert "Revert "Fold opencensus into grpc_impl namespace""	6 years ago
Karthik Ravi Shankar	2bf934f97d	Revert "Fold opencensus into grpc_impl namespace"	6 years ago
Jan Tattermusch	cf6a311761	Revert "Windows builds for gRPC C++ tests"	6 years ago
billfeng327	9e102ea8b0	excluded non-compatible test	6 years ago
billfeng327	4241edeaa4	renamed tag to no_windows in conformation with Bazel and TensorFlow	6 years ago
Soheil Hassas Yeganeh	18b19105f2	Implement TCP_INQ for gRPC in Linux TCP_INQ is a socket option we added to Linux to report pending bytes on the socket as a control message. Using TCP_INQ we can accurately decide whether to continue read or not. Add an urgent parameter, when we do not want to wait for EPOLLIN. This commit improves the latency of 1 RPC unary (minimal benchmark) significantly: Before: l_50: 61.3584984733 l_90: 94.8328711277 l_99: 126.211351174 l_999: 158.722406029 After: l_50: 51.3546011488 (-16%) l_90: 72.3420731581 (-23%) l_99: 103.280218974 (-18%) l_999: 130.905689996 (-17%)	6 years ago
billfeng327	ab06853fc9	C++ Windows test builds	6 years ago
Soheil Hassas Yeganeh	c060d55cc7	Revert "Introduce grpc_byte_buffer_reader_peek and use it for Protobuf parsing."	6 years ago
Karthik Ravi Shankar	b606cad6cc	Fold opencensus into grpc_impl namespace Moving opencensus into grpc_impl namespace..	6 years ago
Soheil Hassas Yeganeh	509e77a5a3	Introduce grpc_byte_buffer_reader_peek and use it for Protobuf parsing. grpc_byte_buffer_reader_next() copies and references the slice. This is not always necessary since the caller will not use the slice after destroying the byte buffer. A prominent example is the protobuf parser, which calls grpc_byte_buffer_reader_next() and immediately unrefs the slice after the call. This ref() and unref() calls can be very expensive in the hot path. This commit introduces grpc_byte_buffer_reader_peek() which essentialy return a pointer to the slice in the buffer, i.e., no copies, and no refs. QPS of 1MiB 1 Channel callback benchmark increases by 5%. More importantly insructions per cycle is increased by 10%. Also add tests and benchmarks for byte_buffer_reader_peek()	6 years ago
Mark D. Roth	251d66aac6	Convert client channel factory to C++	6 years ago
Nicolas Noble	91ad888497	Revert "Revert "Folding the Channel class into the grpc_impl namespace.""	6 years ago

1 2 3 4 5 ...

752 Commits (3f46d6897587e29157ef5adb3ecce2c3725a2e81)