Unexpected Behaviour of `@btime` - cubic splines evaluation

[cortner]

The oddest thing happens when I benchmark gradient1 for cubic splines in Interpolations.jl:

The following benchmark script

using Interpolations, ForwardDiff 
rr = range(0.0, stop=5.0, length=10_000)
spl = cubic_spline_interpolation(rr, sin.(rr))
r = rand() * 5
print(" gradient1: "); @btime Interpolations.gradient1($spl, $r)
@btime Interpolations.gradient1($spl, $r)

produces output

 gradient1:   12.345 ns (0 allocations: 0 bytes)
  56.682 ns (0 allocations: 0 bytes)

System info:

julia> versioninfo()
Julia Version 1.8.2
Commit 36034abf26 (2022-09-29 15:21 UTC)
Platform Info:
  OS: macOS (arm64-apple-darwin21.6.0)
  CPU: 10 × Apple M1 Max
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-13.0.1 (ORCJIT, apple-m1)
  Threads: 8 on 8 virtual cores
Environment:
  JULIA_EDITOR = code
  JULIA_NUM_THREADS = 8

[cortner]

I've filed this issue originally with Interpolations.jl but I think it is really a benchmarking issue.

[roflmaostc]

using Interpolations, ForwardDiff, BenchmarkTools
rr = range(0.0, stop=5.0, length=10_000);
spl = cubic_spline_interpolation(rr, sin.(rr));
r = rand() * 5;
print(" gradient1: "); @btime Interpolations.gradient1($spl, $r)
@btime Interpolations.gradient1($spl, $r)

Gives me:

julia> print(" gradient1: "); @btime Interpolations.gradient1($spl, $r)
 gradient1: 
  24.252 ns (0 allocations: 0 bytes)
0.8967853309275732

julia> @btime Interpolations.gradient1($spl, $r)
  24.152 ns (0 allocations: 0 bytes)
0.8967853309275732

julia> versioninfo()
Julia Version 1.8.5
Commit 17cfb8e65ea (2023-01-08 06:45 UTC)
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 24 × AMD Ryzen 9 5900X 12-Core Processor
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-13.0.1 (ORCJIT, znver3)
  Threads: 12 on 24 virtual cores
Environment:
  JULIA_NUM_THREADS = 12

(jl_T8IZmP) pkg> st
Status `/tmp/jl_T8IZmP/Project.toml`
  [6e4b80f9] BenchmarkTools v1.3.2
  [f6369f11] ForwardDiff v0.10.35
  [a98d9a8b] Interpolations v0.14.7

[cortner]

I cannot reproduce on latest versions of packages either.