Skip to content

Instantly share code, notes, and snippets.

View chriscoey's full-sized avatar

Chris Coey chriscoey

94 followers · 366 following
View GitHub Profile
@chriscoey
chriscoey / invhessprod.jl
Last active July 14, 2021 05:16
benchmarking inverse hessian product oracle for EpiPerSepSpectral examples
# inv hess prod oracle vs explicit hess factorization and solve
# timing and numerics checks at final point already loaded in cone
@assert stat == MOI.OPTIMAL # has to be feasible point
println("\noracle timings results")
d = inst.d
println("d = $d")
# get the EpiPerSepSpectral cone
cone = JuMP.backend(model).optimizer.model.cones[end]
@assert cone isa Hypatia.Cones.EpiPerSepSpectral
@chriscoey
chriscoey / 0717.csv
Last active August 17, 2020 21:55
hess nbhd
We can make this file beautiful and searchable if this error is corrected: Unclosed quoted field in line 4.
example,inst_set,count,inst_data,solver,extender,total_time,build_time,status,solve_time,iters,prim_obj,dual_obj,obj_diff,compl,x_viol,y_viol,z_viol,n,p,q,cones
MatrixRegressionJuMP,nat,1,"(18, 3, 15, 0, 0.2, 0, 0, 0)",Hypatia,,51.679719601,0.987383304,Optimal,10.940243005752563,9,7.503599744366309,7.503599627367746,1.1699856372615614e-7,5.027543133780554e-7,2.5024205818624523e-8,0.0,9.682654988907302e-8,47,0,93,"[""EpiPerSquare"", ""EpiNormSpectral""]"
MatrixRegressionJuMP,nat,2,"(30, 5, 25, 0, 0.2, 0, 0, 0)",Hypatia,,1.006911795,0.000775735,Optimal,0.019622087478637695,10,23.553193227354782,23.553193014466586,2.1288819596065878e-7,4.819228041955053e-6,1.130074476218823e-7,0.0,4.5691979688733397e-7,127,0,253,"[""EpiPerSquare"", ""EpiNormSpectral""]"
MatrixRegressionJuMP,nat,3,"(60, 10, 50, 0, 0.2, 0, 0, 0)",Hypatia,,1.002435584,0.005848642,Optimal,0.3052389621734619,13,58.14409839992734,58.14409846254449,-6.26171541284748e-8,3.835005893699096e-6,4.639028516884025e-8,0.0,1.26212742035392e-7,502,0,1003,"[""Epi
@chriscoey
chriscoey / 0826.csv
Last active July 27, 2020 04:49
We can make this file beautiful and searchable if this error is corrected: It looks like row 5 should actually have 21 columns, instead of 7. in line 4.
example,inst_set,real_T,count,inst_data,extender,test_time,build_time,status,solve_time,iters,prim_obj,dual_obj,obj_diff,compl,x_viol,y_viol,z_viol,ext_n,ext_p,ext_q
PortfolioJuMP,bench1,Float64,1,"(10, false, true)",,43.873566497,0.998168859,Optimal,10.122204065322876,14,-0.999904658725398,-0.9999046604315086,1.7061105861415626e-9,2.136649684379739e-9,1.7496815107875818e-10,1.1102230246251565e-16,9.877252789811996e-11,10,1,32
PortfolioJuMP,bench1,Float64,2,"(10, false, true)",ClassicConeOptimizer,13.01301585,0.000134566,Optimal,0.00415802001953125,16,-0.9999046580112835,-0.9999046643897858,6.378502281911835e-9,9.207296309539824e-9,2.386874691277697e-10,0.0,1.746000068871279e-10,20,1,51
PortfolioJuMP,bench1,Float64,3,"(500, false, true)",,0.775170642,0.058109306,Optimal,0.6779279708862305,36,-0.9999046588341749,-0.9999046613561056,2.521930664300953e-9,3.683510973579582e-9,2.3675506000131463e-12,0.0,2.0589768742252233e-11,500,1,1502
PortfolioJuMP,bench1,Float64,4,"(500, false, true)",ClassicConeOptimizer,1.962
@chriscoey
chriscoey / 7e442f4_native_mac.txt
Created March 3, 2020 18:46
hypatia native tests at 7e442f4
native tests total time: 2.22e+02 seconds
1599×7 DataFrame
│ Row │ inst_name │ sys_solver │ real_T │ preprocess │ init_use_indirect │ reduce │ test_time │
│ │ String │ String │ String │ Bool │ Bool │ Bool │ Float64 │
├──────┼──────────────────────────┼─────────────────┼──────────┼────────────┼───────────────────┼────────┼─────────────┤
│ 1 │ dimension1 │ NaiveElimDense │ Float64 │ 1 │ 0 │ 1 │ 0.0625103 │
│ 2 │ dimension1 │ NaiveElimDense │ Float32 │ 1 │ 0 │ 1 │ 0.0658439 │
│ 3 │ dimension1 │ NaiveElimDense │ BigFloat │ 1 │ 0 │ 1 │ 0.066387 │
│ 4 │ dimension1 │ NaiveDense │ Float64 │ 1 │ 0 │ 1 │ 0.00145362 │
@chriscoey
chriscoey / fd6bda7_native_mac.txt
Created March 3, 2020 18:25
hypatia native tests at fd6bda7
native tests total time: 2.13e+02 seconds
1597×7 DataFrame
│ Row │ inst_name │ sys_solver │ real_T │ preprocess │ init_use_indirect │ reduce │ test_time │
│ │ String │ String │ String │ Bool │ Bool │ Bool │ Float64 │
├──────┼──────────────────────────┼─────────────────┼──────────┼────────────┼───────────────────┼────────┼─────────────┤
│ 1 │ dimension1 │ NaiveElimDense │ Float64 │ 1 │ 0 │ 1 │ 0.0779543 │
│ 2 │ dimension1 │ NaiveElimDense │ Float32 │ 1 │ 0 │ 1 │ 0.0742461 │
│ 3 │ dimension1 │ NaiveElimDense │ BigFloat │ 1 │ 0 │ 1 │ 0.0731489 │
│ 4 │ dimension1 │ NaiveDense │ Float64 │ 1 │ 0 │ 1 │ 0.00195822 │
@chriscoey
chriscoey / 44a7540_mac.txt
Created March 3, 2020 02:14
hypatia commit 44a7540 full examples run on chris mac pro
439×10 DataFrame
│ Row │ model │ example │ inst │ inst_data │ test_time │ solve_time │ iters │ status │ prim_obj │ dual_obj │
│ │ String │ String │ Int64 │ String │ Float64 │ Float64 │ Int64 │ Symbol │ Float64 │ Float64 │
├─────┼────────┼───────────────────┼───────┼──────────────────────────────────────────────────────────┼────────────┼─────────────┼───────┼──────────────────┼──────────────┼──────────────┤
│ 1 │ native │ densityest │ 1 │ (:iris_data, 4, true, true, true) │ 0.50123 │ 0.047962 │ 30 │ Optimal │ -0.159389 │ -0.159389 │
│ 2 │ native │ densityest │ 2 │ (:iris_data, 5, true, true, true) │ 1.95068 │ 0.361243 │ 40 │ Optimal │ -0.207288 │ -0.207288 │
│ 3 │ native │ densityest │ 3 │ (:iris_data, 6, true,
@chriscoey
chriscoey / 13807f8_mac.txt
Created March 3, 2020 01:41
hypatia commit 13807f8 full examples run on chris mac pro
439×10 DataFrame
│ Row │ model │ example │ inst │ inst_data │ test_time │ solve_time │ iters │ status │ prim_obj │ dual_obj │
│ │ String │ String │ Int64 │ String │ Float64 │ Float64 │ Int64 │ Symbol │ Float64 │ Float64 │
├─────┼────────┼───────────────────┼───────┼──────────────────────────────────────────────────────────┼────────────┼─────────────┼───────┼──────────────────┼──────────────┼──────────────┤
│ 1 │ native │ densityest │ 1 │ (:iris_data, 4, true, true, true) │ 0.472703 │ 0.056597 │ 34 │ Optimal │ -0.159389 │ -0.159389 │
│ 2 │ native │ densityest │ 2 │ (:iris_data, 5, true, true, true) │ 1.50063 │ 0.406062 │ 48 │ Optimal │ -0.207288 │ -0.207288 │
│ 3 │ native │ densityest │ 3 │ (:iris_data, 6, true,
@chriscoey
chriscoey / gist:12793cee656ba916cd403cb6278dfcbb
Created May 5, 2019 00:38
SumOfSquares + DynamicPolynomials bug
using DynamicPolynomials
using SemialgebraicSets
using SumOfSquares
using PolyJuMP
d = (n = 2, C2 = [0.0, 0.0], C1 = [0.0, 200.0], C0 = [0.0, 0.0], Pl = [-3.5, 0.0], Pu = [-3.5, 6.0], Ql = [3.5, -4.0], Qu = [3.5, 4.0], Y = Complex{Float64}[0.9615384615384615 - 4.8076923076923075im 0.0 + 0.0im; -1.923076923076923 + 9.615384615384615im 0.9615384615384615 - 4.8076923076923075im])
realify(p) = Polynomial(real.(p.a), p.x)
@chriscoey
chriscoey / gist:c429dcc703cbe32fddf59b33af78e478
Created May 4, 2019 20:07
PowerModels precompile failure (Julia 1.3 master)
[ Info: Precompiling PowerModels [c36e90e8-916a-50a6-bd94-075b64ef4655]
ERROR: Error while loading expression starting at /home/coey/.julia/packages/JuMP/PbnIJ/src/JuMP.jl:229
caused by [exception 1]
syntax: malformed "import" statement
Stacktrace:
[1] top-level scope at /home/coey/.julia/packages/JuMP/PbnIJ/src/JuMP.jl:229
[2] eval at ./boot.jl:330 [inlined]
[3] eval(::Expr) at /home/coey/.julia/packages/JuMP/PbnIJ/src/JuMP.jl:11
[4] top-level scope at /home/coey/.julia/packages/JuMP/PbnIJ/src/JuMP.jl:230
[5] include at ./boot.jl:328 [inlined]