Interpreting TuringBenchmarking.jl output?

Specific Domains Statistics
1

Hi,
I’m trying to benchmark a turing model as described here and it’s not really clear to me what the output means:

2-element BenchmarkTools.BenchmarkGroup:
  tags: []
  "evaluation" => 2-element BenchmarkTools.BenchmarkGroup:
          tags: []
          "linked" => Trial(110.125 μs)
          "standard" => Trial(70.416 μs)
  "gradient" => 4-element BenchmarkTools.BenchmarkGroup:
          tags: []
          "AutoReverseDiff(compile=true)" => 2-element BenchmarkTools.BenchmarkGroup:
                  tags: ["ReverseDiff [compiled]"]
                  "linked" => Trial(5.372 ms)
                  "standard" => Trial(1.294 ms)
          "AutoForwardDiff(chunksize=0)" => 2-element BenchmarkTools.BenchmarkGroup:
                  tags: ["ForwardDiff"]
                  "linked" => Trial(96.739 ms)
                  "standard" => Trial(49.003 ms)
          "AutoReverseDiff()" => 2-element BenchmarkTools.BenchmarkGroup:
                  tags: ["ReverseDiff"]
                  "linked" => Trial(29.593 ms)
                  "standard" => Trial(3.229 ms)
          "AutoZygote()" => 2-element BenchmarkTools.BenchmarkGroup:
                  tags: ["Zygote"]
                  "linked" => Trial(96.555 ms)
                  "standard" => Trial(21.833 ms)

What is the difference between linked and standard?

1 Like