solver benchmark: specifying solver-specific options

tutorials/introduction-to-solverbenchmark/Project.toml

@@ -5,10 +5,12 @@ Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 PyPlot = "d330b81b-6aea-500a-939a-2ce795aea3ee"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 SolverBenchmark = "581a75fa-a23a-52d0-a590-d6201de2218a"
+SolverCore = "ff4d7338-4cf1-434d-91df-b86cb86fb843"
 
 [compat]
 
 DataFrames = "1.3.4"
 Plots = "1.31.7"
 PyPlot = "2.10.0"
 SolverBenchmark = "0.5.3"
+SolverCore = "0.3"

tutorials/introduction-to-solverbenchmark/index.jmd

@@ -221,3 +221,23 @@ p = profile_solvers(stats, costs, costnames)
 Here is a useful tutorial on how to use the benchmark with specific solver:
 [Run a benchmark with OptimizationProblems](https://jso.dev/OptimizationProblems.jl/dev/benchmark/)
 The tutorial covers how to use the problems from `OptimizationProblems` to run a benchmark for unconstrained optimization.
+
+### Handling `solver_specific` in stats
+If a solver's GenericExecutionStats contains a `solver_specific` dictionary
+, a column is created for
+each key in that dictionary in the per-solver `DataFrame`.
+
+Here is a example showing how to set a solver-specific flag and then access it for tabulation:
+```julia
+using NLPModelsTest, DataFrames, SolverCore, SolverBenchmark
+
+function newton(nlp)
+  stats = GenericExecutionStats(nlp)
+  set_solver_specific!(stats, :isConvex, true)
+  return stats
+end
+
+solvers = Dict(:newton => newton)
+problems = [NLPModelsTest.BROWNDEN()]
+stats = bmark_solvers(solvers, problems)
+```