Ken Payson
42bd87e376
Adds gRPC Experimental CQ DoThenAsyncNext lambda API
7 years ago
Yuchen Zeng
ca31256597
Remove redundant WaitForConnected
7 years ago
Vijay Pai
088112fea1
Create inproc version of QPS test and add a few simple tests of this to standard testing suite.
7 years ago
Craig Tiller
c18ad11837
Wait until all clients connected before starting streams
7 years ago
Craig Tiller
f6cd77c48d
Stability fixes
7 years ago
Craig Tiller
2897235609
Expose stats into qps_driver
7 years ago
Vijay Pai
6defb325d9
Remove some pre-C++11-style workarounds
7 years ago
Vijay Pai
c38955501b
Move some internal includes to the right places
7 years ago
Vijay Pai
0f95fa4909
Add idiomatic C++ API for grpc::Slice construction that doesn't
...
require using grpc_slice
7 years ago
Guantao Liu
abe3cf51f0
Add a new metric 'Queries/CPU-sec'.
...
Enable internal credential types.
7 years ago
Jan Tattermusch
7897ae9308
auto-fix most of licenses
8 years ago
Yuxuan Li
8698f2bbee
fix extern c location
8 years ago
Yuxuan Li
87827e03aa
use lock instand of atomics, fix include position
8 years ago
Yuxuan Li
f8365cd87b
clean up, fix minor issue
8 years ago
Yuxuan Li
28efff3e1e
clang-format
8 years ago
Yuxuan Li
999ac157e6
initial implementation.
8 years ago
Vijay Pai
eea8cf0fe3
Add QPS tests for one-sided streaming
8 years ago
yang-g
d0084c220e
Include x.pb.h instead of x.grpc.pb.h for message-only protos
8 years ago
Alexander Polcyn
6ff1ca4871
re-run generate project to build qps json driver
8 years ago
Alexander Polcyn
8a0ef63dea
remove LimitCores in c++ benchmark
8 years ago
Yuxuan Li
98c0bd7c82
add setting channel args functionality to performance tesing
...
setting channel args --draft
clang-format
8 years ago
Vijay Pai
90102c2bfc
Eliminate unnecessary uses of new[]/delete[] that can be replaced
...
with vector
Also start eliminating uses of plain-old delete that are not helpful
8 years ago
Vijay Pai
c0b2acb1a0
Use C++11 final and override
8 years ago
Craig Tiller
d41a4a720f
s/gpr_slice/grpc_slice, and move around tests, impls
8 years ago
Craig Tiller
0895545a58
Complete reporting for failure rates
8 years ago
Craig Tiller
032f9bfb58
Start to log errors from benchmark
8 years ago
Craig Tiller
0c80c7d12c
Collecting failure status progress
8 years ago
yang-g
e267f86c43
Properly init variable
8 years ago
Craig Tiller
9e1c2d2358
Dont start requests until all clients have gotten connected
8 years ago
Craig Tiller
531976a6d4
Add an initial mark that clients could use to delay startup of requests
8 years ago
Vijay Pai
f50020ce03
Appease the const gods, improve readability, stop using 0 and 1 as
...
proxies for false and true.
8 years ago
Vijay Pai
d02988d6b5
clang-format
8 years ago
Vijay Pai
25128f1adf
Better ending for open-loop tests: never wait more than 1 second
...
if we are in termination mode
8 years ago
Vijay Pai
773ecd62dd
Dramatically reduce time required to complete sync test when running
...
with lots of threads (by parallelizing shutdown of course)
8 years ago
Craig Tiller
77c7f9fd62
Merge pull request #7407 from ctiller/delayed-write
...
Benchmark fixes
8 years ago
Vijay Pai
5fde20d9f0
clang-format
8 years ago
Vijay Pai
40317fd720
Resolve pernicious race between destructor and thread functions by insisting that destructor is invoked after the class has gone back to being a harmless single-threaded thing.
8 years ago
Vijay Pai
f373f2cf8b
Stop holding histogram for a long time
9 years ago
Craig Tiller
e7479f5efc
Revert accidental change
9 years ago
Craig Tiller
ea0dee5be9
Fix leaks
9 years ago
vjpai
74e6e135e9
I was trigger-happy with changes to qps_worker in the hope of improving
...
stability. I believe that this change was unneeded and actually
may hurt matters by holding a lock for too long.
Partially undoes #5444 , particular commit e7042b5
9 years ago
Craig Tiller
52237e3a96
clang-format
9 years ago
Craig Tiller
63d24f1c7d
Make qps workers connect before even warmup stage begins
9 years ago
Craig Tiller
97244d5810
clang-format, make code more self-documenting
9 years ago
Craig Tiller
5c123fd22b
Add a shard channel arg to client channel construction
...
C core automatically shares subchannels between channels. Therefore our
multiple channel performance tests were really testing single channel
performance.
9 years ago
vjpai
87bac959ee
Eradicate Uniform, Deterministic, and Pareto interarrival
...
distributions since we don't use them and it's not sensible
to add them
9 years ago
Craig Tiller
6169d5f7b0
Update copyrights
9 years ago
Vijay Pai
c47c9497fa
clang-format
9 years ago
vjpai
e7042b5ddf
Simplify Mark processing code
9 years ago
vjpai
e2d39e08f9
Timer->UsageTimer consistently
9 years ago