docs / benchmarks / concurrent
Benchmarks index
aot
codec
codec / csv
codec / json
collection
- Overview
- ArrayList
- BPlusTree
- Cache
- Collectors
- HashMap
- HashSet
- Heap
- ImmutableList
- ImmutableMap
- ImmutableSet
- LinkedList
- RedBlackTree
- Sort
collection / ltm
concurrent
gfx
hash
http
io
io / file
io / net
io / net / uri
lang
- Overview
- Allocation and ownership
- Calls and dispatch
- Exceptions
- Guid
- Loops
- Math and SIMD
- Optional
- Pair
- Slice
- String
- StringBuilder
- Whole programs: the Benchmarks Game
lang / stream
llm
math
math / fft
math / linalg
math / npio
math / poly
math / random
math / stats
ml
nucleo
primavera
process
reflect
search
search / distance
search / fuzzy
search / ngram
time
toolchain
toolchain / kernel-startup
wire
xpu
xpu / mesh
FiberLocal
cajeta.concurrent.FiberLocal — ambient per-request state.
Benchmarks
| id | workload | against | check |
|---|---|---|---|
concurrent/fiberlocal-get-set | 1e8 passes of where(value, () -> get()) on one fiber (FiberLocal binds by scope; there is no set) | C++ thread_local, Rust thread_local!, Go context.Value, Java ScopedValue.where(..).run(..) and ThreadLocal, Python contextvars | sum |
Results
One chart per benchmark, for the machine selected at the top of the page. Languages are the bars, best first; Cajeta’s bar is in color and the competitors’ are gray. A skipped competitor keeps its row with the reason in place of a bar. A benchmark with no row on the selected machine says so and shows nothing else.
Not yet measured on Proton.
Reproduce
cd bench
./bench.sh --bench fiberlocal-get-set # Cajeta
./bench.sh --bench fiberlocal-get-set --competitors # every pinned competitor
Every row records the compiler and competitor versions, flags, machine, OS and driver; METHODOLOGY states the rules a number must satisfy before it is published.
See also
- Reference: FiberLocal
- Source:
runtime/src/cajeta/concurrent/FiberLocal.cajeta - Benchmark source:
bench/src/main/cajeta/bench/concurrent/