36b57f3432
This series contains several optimizations for dht::token
around its comparison functions as well as minimum_token and maximum_token definitions,
by moving them inline into dht/token.hh
This results in a nice improvement in perf-simple-query:
```
==> perf-simple-query.pre <== (21c67a5a64)
throughput: mean=95774.01 standard-deviation=1129.83 median=96243.64 median-absolute-deviation=1090.08 maximum=96864.09 minimum=94471.19
instructions_per_op: mean=41813.68 standard-deviation=16.27 median=41809.29 median-absolute-deviation=7.02 maximum=41841.64 minimum=41799.41
cpu_cycles_per_op: mean=22383.19 standard-deviation=331.01 median=22254.53 median-absolute-deviation=332.26 maximum=22744.11 minimum=21996.73
==> perf-simple-query.post.0 <== (token: move ordering operator inline)
throughput: mean=96350.01 standard-deviation=640.10 median=96228.88 median-absolute-deviation=621.45 maximum=96988.16 minimum=95478.51
instructions_per_op: mean=41627.13 standard-deviation=37.55 median=41627.06 median-absolute-deviation=2.43 maximum=41679.44 minimum=41573.31
cpu_cycles_per_op: mean=22184.65 standard-deviation=151.03 median=22163.05 median-absolute-deviation=120.83 maximum=22348.49 minimum=21967.30
==> perf-simple-query.post.1 <== (token: operator<=>: optimize the common case)
throughput: mean=96778.29 standard-deviation=1719.34 median=97021.72 median-absolute-deviation=1059.56 maximum=98300.99 minimum=93893.75
instructions_per_op: mean=41590.25 standard-deviation=5.53 median=41589.50 median-absolute-deviation=4.17 maximum=41598.39 minimum=41584.57
cpu_cycles_per_op: mean=22135.33 standard-deviation=471.98 median=21969.30 median-absolute-deviation=244.89 maximum=22905.24 minimum=21685.33
==> perf-simple-query.post.3 <== (token: always initialize data member)
throughput: mean=98264.33 standard-deviation=998.49 median=98533.02 median-absolute-deviation=780.45 maximum=99075.40 minimum=96656.51
instructions_per_op: mean=41657.61 standard-deviation=22.53 median=41648.49 median-absolute-deviation=12.89 maximum=41696.81 minimum=41642.07
cpu_cycles_per_op: mean=21808.57 standard-deviation=93.63 median=21794.56 median-absolute-deviation=75.41 maximum=21949.46 minimum=21719.55
==> perf-simple-query.post.4 <== (token: constexpr ctors, methods, and minimum/maximum_token)
throughput: mean=98095.05 standard-deviation=1333.32 median=98930.22 median-absolute-deviation=906.80 maximum=99209.38 minimum=96194.25
instructions_per_op: mean=41572.28 standard-deviation=6.04 median=41574.49 median-absolute-deviation=4.76 maximum=41579.56 minimum=41564.72
cpu_cycles_per_op: mean=21831.35 standard-deviation=169.56 median=21732.86 median-absolute-deviation=102.93 maximum=22091.66 minimum=21689.63
==> perf-simple-query.post.5 <== (token: initialize non-key tokens with min() value)
throughput: mean=99502.32 standard-deviation=1003.70 median=99744.03 median-absolute-deviation=388.87 maximum=100482.95 minimum=97813.42
instructions_per_op: mean=41593.48 standard-deviation=17.27 median=41585.25 median-absolute-deviation=8.46 maximum=41619.41 minimum=41575.86
cpu_cycles_per_op: mean=21545.90 standard-deviation=86.66 median=21578.01 median-absolute-deviation=43.17 maximum=21612.41 minimum=21395.42
```
Optimization only. No backport required
Closes scylladb/scylladb#19782
* github.com:scylladb/scylladb:
token: initialize non-key tokens with min() value
token: make kind-based ctor private
token: constexpr ctors, methods, and minimum/maximum_token
token: always initialize data member
everywhere: use dht::token is_{minimum,maximum}
token: operator<=>: optimize the common case
token: move ordering operator inline
partitioner_test: add more token-level tests
Scylla in-source tests.
For details on how to run the tests, see docs/dev/testing.md
Shared C++ utils, libraries are in lib/, for Python - pylib/
alternator - Python tests which connect to a single server and use the DynamoDB API unit, boost, raft - unit tests in C++ cql-pytest - Python tests which connect to a single server and use CQL topology* - tests that set up clusters and add/remove nodes cql - approval tests that use CQL and pre-recorded output rest_api - tests for Scylla REST API Port 9000 scylla-gdb - tests for scylla-gdb.py helper script nodetool - tests for C++ implementation of nodetool
If you can use an existing folder, consider adding your test to it. New folders should be used for new large categories/subsystems, or when the test environment is significantly different from some existing suite, e.g. you plan to start scylladb with different configuration, and you intend to add many tests and would like them to reuse an existing Scylla cluster (clusters can be reused for tests within the same folder).
To add a new folder, create a new directory, and then
copy & edit its suite.ini.