1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
|
# go-sieve - SIEVE is simpler than LRU
## What is it?
`go-sieve` is a golang implementation of the
[SIEVE](https://yazhuozhang.com/assets/pdf/nsdi24-sieve.pdf) cache
eviction algorithm.
This implementation closely follows the paper's pseudo-code - but uses
golang generics to provide an ergonomic interface.
## Key Design Features
### Custom Memory Allocator for Reduced GC Pressure
This implementation uses a custom memory allocator designed to minimize
GC pressure:
- **Pre-allocated Node Pool**: Rather than allocating nodes
individually, the cache pre-allocates all nodes at initialization time
in a single contiguous array based on cache capacity.
- **Efficient Freelist**: Recycled nodes are managed through a
zero-overhead freelist that repurposes the existing node pointers,
avoiding the need for auxiliary data structures.
- **Single-Allocation Strategy**: By allocating all memory upfront in a
single operation, the implementation reduces heap fragmentation and
minimizes the number of objects the garbage collector must track.
## Usage
The API is designed to be simple and intuitive. See the test files for
examples of how to use the cache in your applications.
|
