Move norecompile of ODEFunction to DiffEqBase #143

ChrisRackauckas · 2022-03-22T10:03:01Z

No description provided.

codecov · 2022-03-22T10:05:45Z

Codecov Report

Merging #143 (8e61339) into master (18ffccc) will decrease coverage by 16.33%.
The diff coverage is 0.00%.

@@             Coverage Diff             @@
##           master     #143       +/-   ##
===========================================
- Coverage   53.21%   36.87%   -16.34%     
===========================================
  Files          40       40               
  Lines        3497     3460       -37     
===========================================
- Hits         1861     1276      -585     
- Misses       1636     2184      +548

Impacted Files	Coverage Δ
src/problems/ode_problems.jl	`15.38% <0.00%> (-35.35%)`	⬇️
src/scimlfunctions.jl	`38.79% <ø> (-14.91%)`	⬇️
src/problems/rode_problems.jl	`0.00% <0.00%> (-100.00%)`	⬇️
src/problems/noise_problems.jl	`0.00% <0.00%> (-100.00%)`	⬇️
src/tabletraits.jl	`3.57% <0.00%> (-69.77%)`	⬇️
src/operators/common_defaults.jl	`20.58% <0.00%> (-41.18%)`	⬇️
src/problems/steady_state_problems.jl	`40.00% <0.00%> (-40.00%)`	⬇️
src/solutions/rode_solutions.jl	`35.00% <0.00%> (-36.67%)`	⬇️
src/ensemble/ensemble_analysis.jl	`43.85% <0.00%> (-35.66%)`	⬇️
... and 20 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

@time

Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master?

@time

Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master?

@time

Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master?

@time

This is a continuation of #736 instead using OpaqueClosures. As such, it also requires: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 But now the MWE doesn't run: ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ```julia [ Info: Precompiling OrdinaryDiffEq [1dea7af3-3e70-54e6-95c3-0bf5283fa5ed] fatal: error thrown and no exception handler available. ErrorException("unimplemented: serialization of MethodInstances for OpaqueClosure") jl_error at /cygdrive/c/buildbot/worker/package_win64/build/src\rtutils.c:41 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:686 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:526 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:661 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:675 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:834 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_datatype at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:310 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:605 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:357 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:345 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_save_incremental at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:2289 jl_write_compiler_output at /cygdrive/c/buildbot/worker/package_win64/build/src\precompile.c:65 jl_atexit_hook at /cygdrive/c/buildbot/worker/package_win64/build/src\init.c:211 jl_repl_entrypoint at /cygdrive/c/buildbot/worker/package_win64/build/src\jlapi.c:702 mainCRTStartup at /cygdrive/c/buildbot/worker/package_win64/build/cli\loader_exe.c:42 BaseThreadInitThunk at C:\WINDOWS\System32\KERNEL32.DLL (unknown line) RtlUserThreadStart at C:\WINDOWS\SYSTEM32\ntdll.dll (unknown line) ``` The precompilation from SciML/OrdinaryDiffEq.jl#1627 seems to not work with opaque closures. Do they not support precompilation? @Keno

@time

Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master?

@time

This is a continuation of #736 instead using OpaqueClosures. As such, it also requires: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 But now the MWE doesn't run: ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ```julia [ Info: Precompiling OrdinaryDiffEq [1dea7af3-3e70-54e6-95c3-0bf5283fa5ed] fatal: error thrown and no exception handler available. ErrorException("unimplemented: serialization of MethodInstances for OpaqueClosure") jl_error at /cygdrive/c/buildbot/worker/package_win64/build/src\rtutils.c:41 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:686 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:526 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:661 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:675 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:834 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_datatype at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:310 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:605 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:357 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:345 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_save_incremental at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:2289 jl_write_compiler_output at /cygdrive/c/buildbot/worker/package_win64/build/src\precompile.c:65 jl_atexit_hook at /cygdrive/c/buildbot/worker/package_win64/build/src\init.c:211 jl_repl_entrypoint at /cygdrive/c/buildbot/worker/package_win64/build/src\jlapi.c:702 mainCRTStartup at /cygdrive/c/buildbot/worker/package_win64/build/cli\loader_exe.c:42 BaseThreadInitThunk at C:\WINDOWS\System32\KERNEL32.DLL (unknown line) RtlUserThreadStart at C:\WINDOWS\SYSTEM32\ntdll.dll (unknown line) ``` The precompilation from SciML/OrdinaryDiffEq.jl#1627 seems to not work with opaque closures. Do they not support precompilation? @Keno

@time

This is a continuation of #736 instead using OpaqueClosures. As such, it also requires: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 But now the MWE doesn't run: ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ```julia [ Info: Precompiling OrdinaryDiffEq [1dea7af3-3e70-54e6-95c3-0bf5283fa5ed] fatal: error thrown and no exception handler available. ErrorException("unimplemented: serialization of MethodInstances for OpaqueClosure") jl_error at /cygdrive/c/buildbot/worker/package_win64/build/src\rtutils.c:41 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:686 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:526 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:661 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:675 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:834 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_datatype at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:310 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:605 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:357 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:345 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_save_incremental at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:2289 jl_write_compiler_output at /cygdrive/c/buildbot/worker/package_win64/build/src\precompile.c:65 jl_atexit_hook at /cygdrive/c/buildbot/worker/package_win64/build/src\init.c:211 jl_repl_entrypoint at /cygdrive/c/buildbot/worker/package_win64/build/src\jlapi.c:702 mainCRTStartup at /cygdrive/c/buildbot/worker/package_win64/build/cli\loader_exe.c:42 BaseThreadInitThunk at C:\WINDOWS\System32\KERNEL32.DLL (unknown line) RtlUserThreadStart at C:\WINDOWS\SYSTEM32\ntdll.dll (unknown line) ``` The precompilation from SciML/OrdinaryDiffEq.jl#1627 seems to not work with opaque closures. Do they not support precompilation? @Keno

@time

This is a continuation of #736 instead using OpaqueClosures. As such, it also requires: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 But now the MWE doesn't run: ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ```julia [ Info: Precompiling OrdinaryDiffEq [1dea7af3-3e70-54e6-95c3-0bf5283fa5ed] fatal: error thrown and no exception handler available. ErrorException("unimplemented: serialization of MethodInstances for OpaqueClosure") jl_error at /cygdrive/c/buildbot/worker/package_win64/build/src\rtutils.c:41 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:686 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:526 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:661 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:675 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:870 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:834 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_datatype at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:310 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:605 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:357 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_module at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:345 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:733 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:565 jl_serialize_value_ at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:510 [inlined] jl_save_incremental at /cygdrive/c/buildbot/worker/package_win64/build/src\dump.c:2289 jl_write_compiler_output at /cygdrive/c/buildbot/worker/package_win64/build/src\precompile.c:65 jl_atexit_hook at /cygdrive/c/buildbot/worker/package_win64/build/src\init.c:211 jl_repl_entrypoint at /cygdrive/c/buildbot/worker/package_win64/build/src\jlapi.c:702 mainCRTStartup at /cygdrive/c/buildbot/worker/package_win64/build/cli\loader_exe.c:42 BaseThreadInitThunk at C:\WINDOWS\System32\KERNEL32.DLL (unknown line) RtlUserThreadStart at C:\WINDOWS\SYSTEM32\ntdll.dll (unknown line) ``` The precompilation from SciML/OrdinaryDiffEq.jl#1627 seems to not work with opaque closures. Do they not support precompilation? @Keno

ChrisRackauckas · 2022-08-20T12:51:52Z

No longer required.

@time

Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master?

@time

* Setup FunctionWrappersWrappers norecompile mode Needs: - SciML/SciMLBase.jl#143 - SciML/OrdinaryDiffEq.jl#1627 ```julia using OrdinaryDiffEq function f(du, u, p, t) du[1] = 0.2u[1] du[2] = 0.4u[2] end u0 = ones(2) tspan = (0.0, 1.0) prob = ODEProblem{true,false}(f, u0, tspan, Float64[]) function lorenz(du, u, p, t) du[1] = 10.0(u[2] - u[1]) du[2] = u[1] * (28.0 - u[3]) - u[2] du[3] = u[1] * u[2] - (8 / 3) * u[3] end lorenzprob = ODEProblem{true,false}(lorenz, [1.0; 0.0; 0.0], (0.0, 1.0), Float64[]) typeof(prob) === typeof(lorenzprob) # true @time sol = solve(prob, Rosenbrock23(autodiff=false)) @time sol = solve(prob, Rosenbrock23(chunk_size=1)) ``` ``` 2.763588 seconds (10.32 M allocations: 648.718 MiB, 4.92% gc time, 99.89% compilation time) 10.577789 seconds (45.44 M allocations: 2.760 GiB, 4.87% gc time, 99.97% compilation time) ``` While the types of `prob` are exactly the same, there is still a significant amount of compile time, even with that exact same time being called in `using` at OrdinaryDiffEq. Maybe this needs to be run on master? * with extra specialization! * try `@nospecialize` on the arguments * force precompilation? * try more * no precompile file * add sys * do not require SciMLBase changes and support oop * fix oop handling

ChrisRackauckas mentioned this pull request Mar 22, 2022

Setup FunctionWrappersWrappers norecompile mode SciML/DiffEqBase.jl#736

Merged

ChrisRackauckas force-pushed the norecompile branch 2 times, most recently from e6b3650 to 4494b6d Compare March 27, 2022 20:23

ChrisRackauckas force-pushed the norecompile branch from 4494b6d to 7f5c4a2 Compare April 22, 2022 11:23

ChrisRackauckas mentioned this pull request Apr 22, 2022

Setup OpaqueClosures norecompile mode SciML/DiffEqBase.jl#745

Closed

Move norecompile of ODEFunction to DiffEqBase

8e61339

ChrisRackauckas force-pushed the norecompile branch from 7f5c4a2 to 8e61339 Compare August 20, 2022 03:37

ChrisRackauckas closed this Aug 20, 2022

ChrisRackauckas deleted the norecompile branch August 20, 2022 12:51

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Move norecompile of ODEFunction to DiffEqBase #143

Move norecompile of ODEFunction to DiffEqBase #143

Uh oh!

ChrisRackauckas commented Mar 22, 2022

Uh oh!

codecov bot commented Mar 22, 2022 •

edited

Loading

Uh oh!

ChrisRackauckas commented Aug 20, 2022

Uh oh!

Uh oh!

Uh oh!

Move norecompile of ODEFunction to DiffEqBase #143

Move norecompile of ODEFunction to DiffEqBase #143

Uh oh!

Conversation

ChrisRackauckas commented Mar 22, 2022

Uh oh!

codecov bot commented Mar 22, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

ChrisRackauckas commented Aug 20, 2022

Uh oh!

Uh oh!

codecov bot commented Mar 22, 2022 •

edited

Loading