|
|
f83e21b5a0
|
Defeat shared_ptr's single-threaded optimizations
WeaselDB is always going to start multiple threads, so we don't care
about single-threaded performance
|
2025-09-11 13:13:05 -04:00 |
|
|
|
5adbf8eee2
|
Organize bench_reference.cpp with doctest
|
2025-09-11 12:32:25 -04:00 |
|
|
|
2bc17cbfe6
|
Add bench_reference.cpp
Also update snake case script for nanobench symbols
|
2025-09-11 12:22:56 -04:00 |
|
|
|
f56ed2bfbe
|
Rename ArenaAllocator -> Arena
|
2025-09-05 17:57:04 -04:00 |
|
|
|
0357a41dd8
|
Implement spend_cpu_cycles in assembly
The compiler was unrolling it previously, so we're doing assembly now for consistency.
|
2025-09-05 15:16:49 -04:00 |
|
|
|
e67e4aee17
|
Update /ok to serve dual health check/benchmarking role
|
2025-09-05 12:39:10 -04:00 |
|
|
|
0ac4c31a53
|
Measure per metric in render scale bench
|
2025-09-03 12:51:49 -04:00 |
|
|
|
52b0cb3e6e
|
Remove background thread from callback bench
|
2025-09-03 12:19:10 -04:00 |
|
|
|
76193f772c
|
Tinker with benchmarks. Looking at render performance
|
2025-09-03 12:16:56 -04:00 |
|
|
|
ff7642195b
|
Make benchmark metric families global
|
2025-08-30 17:59:24 -04:00 |
|
|
|
935bab9454
|
Make histograms atomic
E.g. count and sum should be consistent with each other
|
2025-08-29 21:05:51 -04:00 |
|
|
|
5592d065de
|
Actually have contention in benchmark
|
2025-08-29 17:05:21 -04:00 |
|
|
|
91e799aae8
|
Use plain arrays and atomic read with intrinsics for render
|
2025-08-29 15:10:10 -04:00 |
|
|
|
4fc277393e
|
Use std::latch sync in benchmarks too
|
2025-08-29 14:29:15 -04:00 |
|
|
|
a5776004de
|
Update potential misunderstanding about thread safety
|
2025-08-29 14:08:41 -04:00 |
|
|
|
62b37c067c
|
Metrics implementation, WIP
|
2025-08-29 13:43:03 -04:00 |
|
|
|
c97920c473
|
format utility improvements
|
2025-08-28 14:40:01 -04:00 |
|
|
|
7808896226
|
Add format benchmarks
|
2025-08-28 14:20:27 -04:00 |
|
|
|
6fb57619c5
|
Remove inaccurate "zero-{copy,allocation}" claims
|
2025-08-28 13:40:05 -04:00 |
|
|
|
f46a98249f
|
Change to loop_iterations
|
2025-08-28 13:34:52 -04:00 |
|
|
|
6e48a0ff9a
|
Let WaitIfUpstreamIdle spin long enough to stay saturated by load_tester
|
2025-08-26 15:38:07 -04:00 |
|
|
|
6dbf29d1e1
|
StaticThreadPipeline
|
2025-08-26 15:13:16 -04:00 |
|
|
|
1f050c861a
|
Add operator[] to batch in thread pipeline
|
2025-08-26 13:39:11 -04:00 |
|
|
|
3c3555da7a
|
Benchmark different batches and wait strategies
|
2025-08-26 13:33:16 -04:00 |
|
|
|
4e9e4d634c
|
Add initial thread pipeline benchmark
|
2025-08-26 13:11:20 -04:00 |
|
|
|
50d873e8eb
|
Actually use loopIterations
|
2025-08-25 13:18:35 -04:00 |
|
|
|
5314448480
|
Move bench_volatile_loop to benchmarks/
|
2025-08-25 13:15:59 -04:00 |
|
|
|
34b5de1744
|
Use reference parser for tests
|
2025-08-18 06:27:24 -04:00 |
|
|
|
9e397d19c9
|
Make codebase consistent with design.md
|
2025-08-17 17:00:23 -04:00 |
|
|
|
8862fdd588
|
Update documentation
|
2025-08-17 16:11:28 -04:00 |
|
|
|
ed84e4b54c
|
Fix debug warnings
|
2025-08-17 15:41:51 -04:00 |
|
|
|
67b7563f64
|
Make comparison benchmarks relative
|
2025-08-17 15:32:54 -04:00 |
|
|
|
519c9457cd
|
Only pass CommitRequest to begin_streaming_parse
|
2025-08-17 15:32:54 -04:00 |
|
|
|
032a4184cc
|
Measure per byte in benchmarks and remove some of them
|
2025-08-17 15:32:54 -04:00 |
|
|
|
fa2a2e4427
|
Decouple parser from CommitRequest
|
2025-08-17 15:32:54 -04:00 |
|
|
|
628fe466ef
|
Compare with rapidjson
|
2025-08-17 15:32:54 -04:00 |
|
|
|
8e33b477eb
|
Make allocate take a template type
So we use the right alignment
|
2025-08-15 13:31:45 -04:00 |
|
|
|
f1794bcb3e
|
Add arena debug visualization tool
|
2025-08-15 11:25:10 -04:00 |
|
|
|
469e44c6f6
|
Consolidate duplicate test data
|
2025-08-15 06:38:38 -04:00 |
|
|
|
a793db40f0
|
Compare with nlohmann/json
|
2025-08-14 23:09:29 -04:00 |
|
|
|
8daea72a22
|
Simplify Arena benchmark
|
2025-08-14 11:35:15 -04:00 |
|
|
|
281e9d728b
|
Initial arena benchmarks
|
2025-08-14 11:25:47 -04:00 |
|
