Run RISC-V Debian GNU/Linux bullseye/sid via QEMU.
opensbi and u-boot-qemu packages are not found add testing apt repository (aka bullseye). Or even unstable (aka sid)sudo vi /etc/apt/sources.list
# Add testing repo (or unstable)
deb http://cdn-aws.deb.debian.org/debian testing main
deb-src http://cdn-aws.deb.debian.org/debian testing main
| /* | |
| Intel(R) Core(TM) i5-14600KF, 64G mem | |
| ❯ perf stat -e branches,branch-misses $BENCH_BIN --bench --profile-time 5 "Binary Search/Traditional" | |
| Gnuplot not found, using plotters backend | |
| Benchmarking Binary Search/Traditional: Complete (Analysis Disabled) | |
| Performance counter stats for 'target/release/deps/bs_bench-df11742845ad013f --bench --profile-time 5 Binary Search/Traditional': |
| use core::arch::aarch64::*; | |
| use std::hint::black_box; | |
| use criterion::{BenchmarkId, Criterion, criterion_group, criterion_main}; | |
| use rand::RngExt; | |
| /// we use unzip to extract even and odd indexed digits, then we can do multiplication and addition | |
| /// in parallel. for [1, 2, 3, 4, 5, 6, 7, 8, 9, 5, 4, 3, 2, 1, 0, 0] | |
| /// it will be split into even = [1, 3, 5, 7, 9, 4, 2, 0, ...] and odd = [2, 4, 6, 8, 5, 3, 1, 0, | |
| /// ...] then we can compute 1*10 + 2, 3*10 + 4, 5*10 + 6, 7*10 + 8, 9*10 + 5, 4*10 + 3, 2*10 + 1, |
| use criterion::{Criterion, criterion_group, criterion_main}; | |
| use std::arch::aarch64::*; | |
| use std::hint::black_box; | |
| #[inline(always)] | |
| unsafe fn atoi_unzip(data_ptr: *const u8) -> u64 { | |
| let zeros = vdupq_n_u8(b'0'); | |
| let reg = vsubq_u8(vld1q_u8(data_ptr), zeros); | |
| // Level 1: u8 -> u16 |
| // Benchmark comparison between atoi_simd and sonic-number | |
| use std::hint::black_box; | |
| use atoi_simd::parse as atoi_simd_parse; | |
| use criterion::{Criterion, criterion_group, criterion_main}; | |
| use sonic_number::simd_str2int; | |
| const NUMBERS: &[&str] = &[ | |
| "1234567890123", | |
| "12345678901234", |
| use criterion::{black_box, criterion_group, criterion_main, Criterion}; | |
| use std::sync::atomic::{AtomicUsize, Ordering}; | |
| /// 显式地使用 compare_exchange_weak 来模拟 compare_exchange (strong) 的语义 | |
| /// 只有当值真的不相等时才返回 Err，虚假失败会在内部循环掉 | |
| fn compare_exchange_weak_wrapped( | |
| atomic: &AtomicUsize, | |
| current: usize, | |
| new: usize, | |
| ) -> Result<usize, usize> { |
| //! A simple benchmark for comparing wasm fallback vs simd128 performance. | |
| //! | |
| //! Build & run: | |
| //! # fallback (no simd128) | |
| //! cargo build --example wasm_bench --target wasm32-wasip1 --release | |
| //! wasmtime run target/wasm32-wasip1/release/examples/wasm_bench.wasm | |
| //! | |
| //! # simd128 | |
| //! RUSTFLAGS="-C target-feature=+simd128" \ | |
| //! cargo build --example wasm_bench --target wasm32-wasip1 --release |
| /// Benchmark comparing RDTSC vs std::time::Instant for time measurements | |
| /// | |
| /// This benchmark measures the overhead of both time measurement approaches | |
| /// used in Miri's tracing infrastructure. | |
| use std::time::{Duration, Instant}; | |
| #[cfg(all(target_os = "linux", any(target_arch = "x86", target_arch = "x86_64")))] | |
| fn rdtsc() -> u64 { | |
| #[cfg(target_arch = "x86")] |
| /*-----------------------------------------------------------------------*/ | |
| /* Program: STREAM */ | |
| /* Revision: $Id: stream.c,v 5.10 2013/01/17 16:01:06 mccalpin Exp mccalpin $ */ | |
| /* Original code developed by John D. McCalpin */ | |
| /* Programmers: John D. McCalpin */ | |
| /* Joe R. Zagar */ | |
| /* */ | |
| /* This program measures memory transfer rates in MB/s for simple */ | |
| /* computational kernels coded in C. */ |
| #![allow(unused)] | |
| use std::collections::{HashMap, HashSet}; | |
| #[derive(Clone, Debug)] | |
| pub enum Expr { | |
| Bool(bool), | |
| Int(i32), | |
| // x | |
| Var(String), |