Benchmarks for #123158 (stephentoub)

[EgorBot]

Processing dotnet/runtime#123158 (comment) command:

Command

-amd -intel -arm

using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Benchmarks).Assembly).Run(args);

[MemoryDiagnoser]
public class Benchmarks
{
    private string[] _strings = [.. Enumerable.Range(0, 10).Select(i => i.ToString())];

    [Benchmark]
    [Arguments("a", "b")]
    public string Join(string s1, string s2) => string.Join("", s1, s2);
    
    [Benchmark]
    public string Join() => string.Join("", _strings);
}

(EgorBot will reply in this issue)

[EgorBot]

Benchmark results on linux_azure_genoa

BenchmarkDotNet v0.15.0, Linux Ubuntu 24.04.3 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 8 logical and 4 physical cores
  Job-JUWTMB : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  Job-YXUOIA : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Method	Toolchain	s1	s2	Mean	Error	Ratio	Gen0	Allocated	Alloc Ratio
Join	Main	?	?	41.33 ns	0.061 ns	1.00	0.0029	48 B	1.00
Join	PR	?	?	40.92 ns	0.069 ns	0.99	0.0029	48 B	1.00
Join	Main	a	b	12.30 ns	0.117 ns	1.00	0.0019	32 B	1.00
Join	PR	a	b	13.26 ns	0.095 ns	1.08	0.0019	32 B	1.00

Artifacts.zip

cc @stephentoub (agent_logs.txt). EgorBot manual: link.

[EgorBot]

Benchmark results on linux_azure_cobalt100

BenchmarkDotNet v0.15.0, Linux Ubuntu 24.04.3 LTS (Noble Numbat)
Neoverse-N2, 8 physical cores
  Job-SBERQX : .NET 11.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD
  Job-EUHZPC : .NET 11.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD

Method	Toolchain	s1	s2	Mean	Error	Ratio	Gen0	Allocated	Alloc Ratio
Join	Main	?	?	45.39 ns	0.059 ns	1.00	0.0007	48 B	1.00
Join	PR	?	?	46.90 ns	0.052 ns	1.03	0.0007	48 B	1.00
Join	Main	a	b	14.15 ns	0.065 ns	1.00	0.0005	32 B	1.00
Join	PR	a	b	14.68 ns	0.020 ns	1.04	0.0005	32 B	1.00

Artifacts.zip

cc @stephentoub (agent_logs.txt). EgorBot manual: link.

[EgorBot]

Benchmark results on linux_azure_cascadelake

BenchmarkDotNet v0.15.0, Linux Ubuntu 24.04.3 LTS (Noble Numbat)
Intel Xeon Platinum 8370C CPU 2.80GHz, 1 CPU, 8 logical and 4 physical cores
  Job-PYWOHP : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  Job-VJOKNW : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Method	Toolchain	s1	s2	Mean	Error	Ratio	Gen0	Allocated	Alloc Ratio
Join	Main	?	?	47.20 ns	0.267 ns	1.00	0.0019	48 B	1.00
Join	PR	?	?	48.48 ns	0.061 ns	1.03	0.0019	48 B	1.00
Join	Main	a	b	14.75 ns	0.139 ns	1.00	0.0013	32 B	1.00
Join	PR	a	b	15.55 ns	0.082 ns	1.05	0.0013	32 B	1.00

Artifacts.zip

cc @stephentoub (agent_logs.txt). EgorBot manual: link.

[EgorBo]

@EgorBot -amd -profiler

using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Benchmarks).Assembly).Run(args);

[MemoryDiagnoser]
public class Benchmarks
{
    private string[] _strings = [.. Enumerable.Range(0, 10).Select(i => i.ToString())];

    [Benchmark]
    public string JoinBench() => string.Join("", _strings);
}

[EgorBot]

Benchmark results on linux_aws_genoa

BenchmarkDotNet v0.15.0, Linux Ubuntu 24.04 LTS (Noble Numbat)
AMD EPYC 9R14, 1 CPU, 16 logical and 16 physical cores
  Job-YTZZTA : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  Job-ZCDCRA : .NET 11.0.0 (42.42.42.42424), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Method	Toolchain	Mean	Error	Ratio	Gen0	Allocated	Alloc Ratio
JoinBench	Main	40.92 ns	0.134 ns	1.00	0.0029	48 B	1.00
JoinBench	PR	40.50 ns	0.230 ns	0.99	0.0029	48 B	1.00

Artifacts.zip

Perf profile for Benchmarks_JoinBench:

Speedscope: Main vs PR 🔥
Svg: Main vs PR
Hot asm: Main vs PR
HW counters: Main vs PR

cc @EgorBo (agent_logs.txt). EgorBot manual: link.