rta 0.0.3

Ṛta (ऋत) is a minimal metadata store for durable system state
Documentation 
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
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144

//! Benchmarks for `write` latency
//! Run using: `taskset -c 2,3,4 cargo bench --bench write`

use hdrhistogram::Histogram;
use rta::{RTA, Rta, RtaCfg};
use std::{sync, thread, time};

const MOD_ID: u8 = 0x00;
const THREADS: usize = 4;
const OPS: usize = 0x100_000;
const COPIES_ON_DISK: usize = 0x100;
const WARMUP_OPS: usize = OPS >> 0x0A;
const OPS_PER_THREAD: usize = OPS / THREADS;

#[repr(C)]
#[repr(align(8))]
#[derive(Debug, Default, Clone, Copy, RTA, PartialEq)]
struct Type([u64; 4]);

#[derive(Debug)]
struct BenchResult {
    hist: Histogram<u64>,
}

#[inline]
fn prep_init() -> (std::path::PathBuf, RtaCfg) {
    let path = tempfile::NamedTempFile::new().unwrap().into_temp_path().to_path_buf();
    let cfg = RtaCfg { module_id: MOD_ID, path: path.clone(), copies_on_disk: COPIES_ON_DISK };

    (path, cfg)
}

#[inline(always)]
fn record_bench(rta: &Rta<Type>, ops: usize) -> BenchResult {
    let mut hist = Histogram::<u64>::new(3).expect("new histogram");
    for _ in 0..ops {
        let start = time::Instant::now();
        unsafe {
            rta.write(|t| {
                for v in &mut t.0 {
                    *v += 1;
                }
            })
            .expect("write");
        }

        hist.record(start.elapsed().as_nanos() as u64).expect("record latency");
    }

    BenchResult { hist }
}

fn single_tx_write_latency() -> BenchResult {
    let (_path, cfg) = prep_init();
    let rta = Rta::<Type>::new(cfg).expect("new Rta");

    // warmup
    for i in 0..WARMUP_OPS {
        unsafe {
            rta.write(|t| t.0 = [i as u64; 4]).expect("warmup write");
        }
    }

    record_bench(&rta, OPS)
}

fn multi_tx_write_latency() -> BenchResult {
    let (_path, cfg) = prep_init();
    let rta = sync::Arc::new(Rta::<Type>::new(cfg).expect("new Rta"));
    let barrier = sync::Arc::new(sync::Barrier::new(THREADS));

    let mut handles = Vec::with_capacity(THREADS);
    for tid in 0..THREADS {
        let rta = sync::Arc::clone(&rta);
        let barrier = sync::Arc::clone(&barrier);

        handles.push(thread::spawn(move || {
            // warmup
            for i in 0..WARMUP_OPS {
                unsafe {
                    rta.write(|t| {
                        t.0 = [tid as u64, i as u64, 0, 0];
                    })
                    .expect("warmup write");
                }
            }

            barrier.wait();
            let result = record_bench(&rta, OPS_PER_THREAD);
            barrier.wait();

            result
        }));
    }

    let mut hist = Histogram::<u64>::new(3).expect("new histogram");
    for handle in handles {
        let result = handle.join().expect("worker should join");

        hist.add(&result.hist).expect("merge histogram");
    }

    BenchResult { hist }
}

fn print_results(single: &BenchResult, multi: &BenchResult) {
    print!("Totale measured operations: {OPS}");
    println!();
    println!("| Metric  | Single TX (µs) | Multi TX (µs) |");
    println!("|:--------|:---------------|:--------------|");
    println!(
        "| P50     | {:>14.4} | {:>13.4} |",
        single.hist.value_at_quantile(0.50) as f64 / 1000.0,
        multi.hist.value_at_quantile(0.50) as f64 / 1000.0,
    );
    println!(
        "| P90     | {:>14.4} | {:>13.4} |",
        single.hist.value_at_quantile(0.90) as f64 / 1000.0,
        multi.hist.value_at_quantile(0.90) as f64 / 1000.0,
    );
    println!(
        "| P99     | {:>14.4} | {:>13.4} |",
        single.hist.value_at_quantile(0.99) as f64 / 1000.0,
        multi.hist.value_at_quantile(0.99) as f64 / 1000.0,
    );
    println!(
        "| MEAN    | {:>14.4} | {:>13.4} |",
        single.hist.mean() as f64 / 1000.0,
        multi.hist.mean() as f64 / 1000.0,
    );
    println!(
        "| MAX     | {:>14.4} | {:>13.4} |",
        single.hist.max() as f64 / 1000.0,
        multi.hist.max() as f64 / 1000.0,
    );
    println!();
}

fn main() {
    let single = single_tx_write_latency();
    let multi = multi_tx_write_latency();

    print_results(&single, &multi);
}