Improvements to benchmark outputs

[penelopeysm]

• Prints more info during the run itself
• Sorts the table more carefully so that options that are similar to each other are grouped together
• Adds extra lines to help guide the eye
• Adds tgrad / tref column so that we can tell how much something improves gradients without having to do mental multiplication

[github-actions]

Benchmark Report

• this PR's head: 4e57712b1813861d11e157b2fc093a353c5a0414
• base branch: 8553e401182b894a653b638e5978f00ae42ae031

Computer Information

Julia Version 1.11.7
Commit f2b3dbda30a (2025-09-08 12:10 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 4 × AMD EPYC 7763 64-Core Processor
  WORD_SIZE: 64
  LLVM: libLLVM-16.0.6 (ORCJIT, znver3)
Threads: 1 default, 0 interactive, 1 GC (on 4 virtual cores)

Benchmark Results

┌───────────────────────┬───────┬─────────────┬───────────────────┬────────┬─────────────────────────────────┬───────────────────────────┬─────────────────────────────────┐
│                       │       │             │                   │        │        t(eval) / t(ref)         │     t(grad) / t(eval)     │        t(grad) / t(ref)         │
│                       │       │             │                   │        │ ──────────┬───────────┬──────── │ ──────┬─────────┬──────── │ ──────────┬───────────┬──────── │
│                 Model │   Dim │  AD Backend │           VarInfo │ Linked │      base │   this PR │ speedup │  base │ this PR │ speedup │      base │   this PR │ speedup │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│               Dynamic │    10 │    mooncake │             typed │   true │    519.33 │    380.85 │    1.36 │  8.96 │   11.58 │    0.77 │   4651.05 │   4410.28 │    1.05 │
│                   LDA │    12 │ reversediff │             typed │   true │   2945.40 │   2704.08 │    1.09 │  1.98 │    1.97 │    1.00 │   5822.08 │   5321.36 │    1.09 │
│   Loop univariate 10k │ 10000 │    mooncake │             typed │   true │ 175512.86 │ 142763.28 │    1.23 │  5.00 │    5.78 │    0.87 │ 878051.65 │ 824952.49 │    1.06 │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│    Loop univariate 1k │  1000 │    mooncake │             typed │   true │  15848.71 │  14577.39 │    1.09 │  5.19 │    5.45 │    0.95 │  82213.87 │  79396.63 │    1.04 │
│      Multivariate 10k │ 10000 │    mooncake │             typed │   true │  35443.18 │  28195.34 │    1.26 │  9.02 │   10.11 │    0.89 │ 319647.48 │ 284970.67 │    1.12 │
│       Multivariate 1k │  1000 │    mooncake │             typed │   true │   3884.77 │   3351.04 │    1.16 │  8.41 │    8.93 │    0.94 │  32685.99 │  29912.61 │    1.09 │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│ Simple assume observe │     1 │ forwarddiff │             typed │  false │     18.08 │     16.52 │    1.09 │  1.89 │    1.83 │    1.03 │     34.26 │     30.27 │    1.13 │
│           Smorgasbord │   201 │ forwarddiff │             typed │  false │   2669.01 │   2411.90 │    1.11 │ 85.58 │   45.30 │    1.89 │ 228402.83 │ 109269.90 │    2.09 │
│           Smorgasbord │   201 │ forwarddiff │       simple_dict │   true │  24579.98 │  23757.61 │    1.03 │ 25.83 │   26.24 │    0.98 │ 634861.41 │ 623463.26 │    1.02 │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│           Smorgasbord │   201 │ forwarddiff │ simple_namedtuple │   true │   1086.36 │    990.94 │    1.10 │ 86.99 │   78.38 │    1.11 │  94507.29 │  77673.10 │    1.22 │
│           Smorgasbord │   201 │      enzyme │             typed │   true │   2648.14 │   2420.15 │    1.09 │  5.43 │    5.08 │    1.07 │  14375.94 │  12306.01 │    1.17 │
│           Smorgasbord │   201 │    mooncake │             typed │   true │   2657.08 │   2414.65 │    1.10 │  5.65 │    5.83 │    0.97 │  15021.01 │  14074.25 │    1.07 │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│           Smorgasbord │   201 │ reversediff │             typed │   true │   2639.04 │   2433.91 │    1.08 │ 56.44 │   56.41 │    1.00 │ 148942.12 │ 137293.07 │    1.08 │
│           Smorgasbord │   201 │ forwarddiff │      typed_vector │   true │   2700.46 │   2437.90 │    1.11 │ 48.02 │   45.26 │    1.06 │ 129688.24 │ 110349.57 │    1.18 │
│           Smorgasbord │   201 │ forwarddiff │           untyped │   true │   2370.40 │   2167.86 │    1.09 │ 48.52 │   44.90 │    1.08 │ 115021.81 │  97339.44 │    1.18 │
├───────────────────────┼───────┼─────────────┼───────────────────┼────────┼───────────┼───────────┼─────────┼───────┼─────────┼─────────┼───────────┼───────────┼─────────┤
│           Smorgasbord │   201 │ forwarddiff │    untyped_vector │   true │   2389.78 │   2195.51 │    1.09 │ 49.17 │   48.19 │    1.02 │ 117499.70 │ 105791.06 │    1.11 │
│              Submodel │     1 │    mooncake │             typed │   true │     26.44 │     25.05 │    1.06 │  5.25 │    5.27 │    0.99 │    138.66 │    132.09 │    1.05 │
└───────────────────────┴───────┴─────────────┴───────────────────┴────────┴───────────┴───────────┴─────────┴───────┴─────────┴─────────┴───────────┴───────────┴─────────┘

[codecov]

Codecov Report

❌ Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (breaking@8553e40). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
src/test_utils/ad.jl	0.00%	2 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             breaking    #1146   +/-   ##
===========================================
  Coverage            ?   81.67%           
===========================================
  Files               ?       42           
  Lines               ?     3930           
  Branches            ?        0           
===========================================
  Hits                ?     3210           
  Misses              ?      720           
  Partials            ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.