Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Switch CI to Julia v1.10 #1562

Merged
merged 28 commits into from
Apr 29, 2024
Merged

Switch CI to Julia v1.10 #1562

merged 28 commits into from
Apr 29, 2024

Conversation

ranocha
Copy link
Member

@ranocha ranocha commented Jul 7, 2023

No description provided.

@ranocha ranocha added the testing label Jul 7, 2023
@ranocha
Copy link
Member Author

ranocha commented Jul 15, 2023

Allocations with the P4estMesh are caused by JuliaLang/julia#50562

@sloede
Copy link
Member

sloede commented Aug 18, 2023

While on vacation I noticed that --check-bounds=no leads again to crazy long compile times,eg, for the optimized Ranocha turbo Euler setup. Do you get the same issues?

@ranocha
Copy link
Member Author

ranocha commented Aug 18, 2023

While on vacation I noticed that --check-bounds=no leads again to crazy long compile times,eg, for the optimized Ranocha turbo Euler setup. Do you get the same issues?

I haven't tested it. It would be great if you could report this issue. So far, I have concentrated on the other issues with Julia v1.10:

There is evidence for other problems, e.g., p4est_part1 timed out (https://github.com/trixi-framework/Trixi.jl/actions/runs/5898667882/job/16000054066?pr=1562#step:7:3884).

@sloede
Copy link
Member

sloede commented Aug 18, 2023

OK, so first evidence that something is wrong with Julia v1.10 and --check-bounds=no (measured on Roci):

julia-1.10 --project=.:

julia> @time using Revise, Trixi, OrdinaryDiffEq
  3.221956 seconds (3.73 M allocations: 248.451 MiB, 7.26% gc time, 8.06% compilation time)
julia> @time trixi_include(default_example())
...
 13.680299 seconds (24.82 M allocations: 1.517 GiB, 4.26% gc time, 99.81% compilation time: <1% of which was recompilation)

julia-1.10 --project=. --check-bounds=no:

julia> @time using Revise, Trixi, OrdinaryDiffEq
  10.682994 seconds (25.34 M allocations: 5.952 GiB, 6.10% gc time, 3.66% compilation time: 10% of which was recompilation)
julia> @time trixi_include(default_example())
...
 69.902309 seconds (239.64 M allocations: 6.916 GiB, 1.20% gc time, 16.05% compilation time: <1% of which was recompilation)
julia> @time trixi_include(default_example())
...
 58.920417 seconds (214.77 M allocations: 5.291 GiB, 0.63% gc time)

And it's not just a compilation issue as one can see, it must be a runtime issue (type instabilities again?) 😢

@sloede
Copy link
Member

sloede commented Aug 19, 2023

Seems like StaticArrays.jl and/or StrideArrays.jl is involved:

With --check-bounds=no:

@code_warntype Trixi.get_node_vars(u, equations, dg, i, j, element)

MethodInstance for Trixi.get_node_vars(::StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, ::LinearScalarAdvectionEquation2D{Float64}, ::DGSEM{LobattoLegendreBasis{Float64, 3, SVector{3, Float64}, Matrix{Float64}, Matrix{Float64}, Matrix{Float64}}, Trixi.LobattoLegendreMortarL2{Float64, 3, Matrix{Float64}, Matrix{Float64}}, SurfaceIntegralWeakForm{FluxLaxFriedrichs{typeof(max_abs_speed_naive)}}, VolumeIntegralWeakForm}, ::Int64, ::Int64, ::Int64)
  from get_node_vars(u, equations, solver::DG, indices...) @ Trixi ~/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:465
Arguments
  #self#::Core.Const(Trixi.get_node_vars)
  u::StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}
  equations::LinearScalarAdvectionEquation2D{Float64}
  solver::DGSEM{LobattoLegendreBasis{Float64, 3, SVector{3, Float64}, Matrix{Float64}, Matrix{Float64}, Matrix{Float64}}, Trixi.LobattoLegendreMortarL2{Float64, 3, Matrix{Float64}, Matrix{Float64}}, SurfaceIntegralWeakForm{FluxLaxFriedrichs{typeof(max_abs_speed_naive)}}, VolumeIntegralWeakForm}
  indices::Tuple{Int64, Int64, Int64}
Locals
  #307::Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}}
Body::Any
1nothing%2  = Trixi.:(var"#307#320")::Core.Const(Trixi.var"#307#320")
│   %3  = Core.typeof(u)::Core.Const(StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}})
│   %4  = Core.typeof(indices)::Core.Const(Tuple{Int64, Int64, Int64})
│   %5  = Core.apply_type(%2, %3, %4)::Core.Const(Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}})
│         (#307 = %new(%5, u, indices))%7  = #307::Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}}%8  = Trixi.nvariables(equations)::Core.Const(1)
│   %9  = Trixi.Val(%8)::Core.Const(Val{1}())
│   %10 = Trixi.ntuple(%7, %9)::Tuple{Any}%11 = Trixi.SVector(%10)::Any
└──       return %11

@code_native Trixi.get_node_vars(u, equations, dg, i, j, element)

	.text
	.file	"get_node_vars"
	.globl	julia_get_node_vars_2212        # -- Begin function julia_get_node_vars_2212
	.p2align	4, 0x90
	.type	julia_get_node_vars_2212,@function
julia_get_node_vars_2212:               # @julia_get_node_vars_2212
; ┌ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:465 within `get_node_vars`
# %bb.0:                                # %top
	push	rbp
	mov	rbp, rsp
	push	r15
	push	r14
	push	r13
	push	r12
	push	rbx
	and	rsp, -32
	sub	rsp, 128
	vxorps	xmm0, xmm0, xmm0
	mov	qword ptr [rsp + 96], rdi       # 8-byte Spill
	mov	qword ptr [rsp + 88], r9        # 8-byte Spill
	#APP
	mov	rax, qword ptr fs:[0]
	#NO_APP
	mov	r13, rcx
	lea	rcx, [rsp + 32]
; │ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:476 within `get_node_vars`
; │┌ @ ntuple.jl:48 within `ntuple`
; ││┌ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:476 within `#307`
; │││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:941 within `getindex`
; ││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:930 within `_offset_ptr`
; │││││┌ @ tuple.jl:322 within `map`
	movabs	r12, offset "j_-_2214"
	mov	edi, 1
	mov	rbx, r8
	vmovaps	ymmword ptr [rsp + 32], ymm0
	mov	qword ptr [rsp + 72], 0
	mov	qword ptr [rsp + 64], 0
	mov	r14, qword ptr [rax - 8]
; │││││└
; │││││ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:929 within `_offset_ptr`
; │││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:484 within `pointer`
	mov	qword ptr [rsp + 32], 16
	mov	rax, qword ptr [r14]
	mov	qword ptr [rsp + 40], rax
	mov	qword ptr [r14], rcx
; │││││└
; │││││ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:930 within `_offset_ptr`
; │││││┌ @ tuple.jl:322 within `map`
	vzeroupper
	call	r12
; ││││││ @ tuple.jl:322 within `map` @ tuple.jl:322
	mov	rdi, r13
; ││││││ @ tuple.jl:322 within `map`
	mov	r15, rax
	mov	qword ptr [rsp + 72], rax
; ││││││ @ tuple.jl:322 within `map` @ tuple.jl:322
	call	r12
; ││││││ @ tuple.jl:322 within `map` @ tuple.jl:322 @ tuple.jl:319
	mov	rdi, rbx
; ││││││ @ tuple.jl:322 within `map` @ tuple.jl:322
	mov	r13, rax
	mov	qword ptr [rsp + 64], rax
; ││││││ @ tuple.jl:322 within `map` @ tuple.jl:322 @ tuple.jl:319
	call	r12
	mov	rdi, qword ptr [rsp + 88]       # 8-byte Reload
	mov	rbx, rax
	mov	qword ptr [rsp + 56], rax
	call	r12
; │││││└
; │││││ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:933 within `_offset_ptr`
; │││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:9 within `invpermtuple`
; ││││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:9 within `macro expansion`
	mov	qword ptr [rsp], r15
	mov	qword ptr [rsp + 8], r13
	mov	qword ptr [rsp + 16], rbx
	mov	qword ptr [rsp + 48], rax
	mov	qword ptr [rsp + 24], rax
	movabs	rax, offset jl_f_tuple
	mov	rsi, rsp
	xor	edi, edi
	mov	edx, 4
	call	rax
	mov	r12, qword ptr [rsp + 96]       # 8-byte Reload
; │││││└└
	movabs	rbx, offset ijl_gc_pool_alloc
	mov	esi, 1136
	mov	edx, 16
; │││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:9 within `invpermtuple`
; ││││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:9 within `macro expansion`
	mov	r15, rax
	mov	r13, qword ptr [r12 + 8]
	mov	qword ptr [rsp + 56], rax
; │││││└└
	mov	rdi, qword ptr [r14 + 16]
	call	rbx
	mov	rcx, rbx
	mov	rbx, rax
	movabs	rdi, 140223945836848
	mov	esi, 1136
	mov	edx, 16
	mov	qword ptr [rbx - 8], rdi
	mov	rax, qword ptr [r12]
	mov	qword ptr [rsp + 64], rbx
	mov	qword ptr [rbx], rax
	mov	rdi, qword ptr [r14 + 16]
	call	rcx
	movabs	rcx, 140214323774288
	movabs	rdi, 140214480650128
	mov	edx, 3
	mov	qword ptr [rax - 8], rcx
	mov	qword ptr [rsp], rbx
	mov	rbx, rsp
	mov	qword ptr [rsp + 8], r15
	movabs	r15, offset ijl_apply_generic
	mov	qword ptr [rax], r13
	mov	qword ptr [rsp + 48], rax
	mov	qword ptr [rsp + 16], rax
	mov	rsi, rbx
	call	r15
	movabs	rdi, 140214480669152
; ││││└
	mov	rsi, rbx
	mov	edx, 1
	mov	qword ptr [rsp + 48], rax
	mov	qword ptr [rsp], rax
	call	r15
	mov	qword ptr [rsp + 48], rax
; ││└└
	mov	qword ptr [rsp], rax
	xor	edi, edi
	mov	rsi, rbx
	mov	edx, 1
	movabs	rax, offset jl_f_tuple
	call	rax
	movabs	rdi, 140214497112368
; │└
	mov	rsi, rbx
	mov	edx, 1
	mov	qword ptr [rsp + 48], rax
	mov	qword ptr [rsp], rax
	call	r15
	mov	rcx, qword ptr [rsp + 40]
	mov	qword ptr [r14], rcx
	lea	rsp, [rbp - 40]
	pop	rbx
	pop	r12
	pop	r13
	pop	r14
	pop	r15
	pop	rbp
	ret
.Lfunc_end0:
	.size	julia_get_node_vars_2212, .Lfunc_end0-julia_get_node_vars_2212
; └
                                        # -- End function
	.section	".note.GNU-stack","",@progbits

Without --check-bounds=no:

@code_warntype Trixi.get_node_vars(u, equations, dg, i, j, element)

MethodInstance for Trixi.get_node_vars(::StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, ::LinearScalarAdvectionEquation2D{Float64}, ::DGSEM{LobattoLegendreBasis{Float64, 3, SVector{3, Float64}, Matrix{Float64}, Matrix{Float64}, Matrix{Float64}}, Trixi.LobattoLegendreMortarL2{Float64, 3, Matrix{Float64}, Matrix{Float64}}, SurfaceIntegralWeakForm{FluxLaxFriedrichs{typeof(max_abs_speed_naive)}}, VolumeIntegralWeakForm}, ::Int64, ::Int64, ::Int64)
  from get_node_vars(u, equations, solver::DG, indices...) @ Trixi ~/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:465
Arguments
  #self#::Core.Const(Trixi.get_node_vars)
  u::StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}
  equations::LinearScalarAdvectionEquation2D{Float64}
  solver::DGSEM{LobattoLegendreBasis{Float64, 3, SVector{3, Float64}, Matrix{Float64}, Matrix{Float64}, Matrix{Float64}}, Trixi.LobattoLegendreMortarL2{Float64, 3, Matrix{Float64}, Matrix{Float64}}, SurfaceIntegralWeakForm{FluxLaxFriedrichs{typeof(max_abs_speed_naive)}}, VolumeIntegralWeakForm}
  indices::Tuple{Int64, Int64, Int64}
Locals
  #307::Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}}
Body::SVector{1, Float64}
1nothing%2  = Trixi.:(var"#307#320")::Core.Const(Trixi.var"#307#320")
│   %3  = Core.typeof(u)::Core.Const(StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}})
│   %4  = Core.typeof(indices)::Core.Const(Tuple{Int64, Int64, Int64})
│   %5  = Core.apply_type(%2, %3, %4)::Core.Const(Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}})
│         (#307 = %new(%5, u, indices))%7  = #307::Trixi.var"#307#320"{StrideArraysCore.PtrArray{Float64, 4, (1, 2, 3, 4), Tuple{Static.StaticInt{1}, Static.StaticInt{3}, Static.StaticInt{3}, Int64}, NTuple{4, Nothing}, NTuple{4, Static.StaticInt{1}}}, Tuple{Int64, Int64, Int64}}%8  = Trixi.nvariables(equations)::Core.Const(1)
│   %9  = Trixi.Val(%8)::Core.Const(Val{1}())
│   %10 = Trixi.ntuple(%7, %9)::Tuple{Float64}%11 = Trixi.SVector(%10)::SVector{1, Float64}
└──       return %11

@code_native Trixi.get_node_vars(u, equations, dg, i, j, element)

	.text
	.file	"get_node_vars"
	.globl	julia_get_node_vars_6674        # -- Begin function julia_get_node_vars_6674
	.p2align	4, 0x90
	.type	julia_get_node_vars_6674,@function
julia_get_node_vars_6674:               # @julia_get_node_vars_6674
; ┌ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:465 within `get_node_vars`
# %bb.0:                                # %top
	push	rbp
; │ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:476 within `get_node_vars`
; │┌ @ ntuple.jl:48 within `ntuple`
; ││┌ @ /home/mschlott/.julia/packages/Trixi/CdZpe/src/solvers/dg.jl:476 within `#307`
; │││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:941 within `getindex`
; ││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:933 within `_offset_ptr`
; │││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:911 within `_offset_ptr_dense`
; ││││││┌ @ pointer.jl:282 within `+`
; │││││││┌ @ boot.jl:793 within `UInt64`
	mov	rdx, qword ptr [rdi]
; ││││││└└
; ││││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:902 within `_offset_dense` @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:902 @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:901
; │││││││┌ @ promotion.jl:423 within `*` @ int.jl:88
	lea	rax, [r9 + 2*r9]
	mov	rbp, rsp
; │││││││└
; │││││││ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:902 within `_offset_dense` @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:902
; │││││││┌ @ int.jl:87 within `+`
	add	rax, r8
; │││││││└
; │││││││┌ @ promotion.jl:423 within `*` @ int.jl:88
	lea	rax, [rax + 2*rax]
; │││││││└
; │││││││ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:902 within `_offset_dense`
; │││││││┌ @ int.jl:87 within `+`
	add	rax, rcx
; ││││└└└└
; ││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:719 within `pload`
; │││││┌ @ /home/mschlott/.julia/packages/StrideArraysCore/COJRJ/src/ptr_array.jl:719 within `macro expansion`
; ││││││┌ @ pointer.jl:119 within `unsafe_load` @ pointer.jl:119
	vmovsd	xmm0, qword ptr [rdx + 8*rax - 104] # xmm0 = mem[0],zero
; │└└└└└└
	pop	rbp
	ret
.Lfunc_end0:
	.size	julia_get_node_vars_6674, .Lfunc_end0-julia_get_node_vars_6674
; └
                                        # -- End function
	.section	".note.GNU-stack","",@progbits

@sloede
Copy link
Member

sloede commented Aug 19, 2023

Yep, seems like StrideArrays.jl is required to break it. If I do something like

u_array = similar(Array{Float64}, u)
u_array .= u

and then call

@code_warntype Trixi.get_node_vars(u_array, equations, dg, i, j, element)

there are no type instabilities anymore.

@sloede
Copy link
Member

sloede commented Aug 19, 2023

xref JuliaSIMD/StrideArrays.jl#78

@sloede
Copy link
Member

sloede commented Aug 20, 2023

xref JuliaLang/julia#50985

Copy link

codecov bot commented Apr 23, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 96.08%. Comparing base (73a1513) to head (efa9e95).

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #1562      +/-   ##
==========================================
+ Coverage   89.59%   96.08%   +6.49%     
==========================================
  Files         450      450              
  Lines       36155    36168      +13     
==========================================
+ Hits        32390    34749    +2359     
+ Misses       3765     1419    -2346     
Flag Coverage Δ
unittests 96.08% <ø> (+6.49%) ⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

@DanielDoehring
Copy link
Contributor

As far as I understand, the remaining errors are caused due to some methods being overwritten.
What is the recommended way of handling this?

test/test_trixi.jl Outdated Show resolved Hide resolved
@DanielDoehring DanielDoehring self-requested a review April 25, 2024 09:45
DanielDoehring
DanielDoehring previously approved these changes Apr 25, 2024
Copy link
Contributor

@DanielDoehring DanielDoehring left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good to got from my side!

@sloede sloede changed the title WIP: check CI on Julia v1.10 pre-release version Switch CI to Julia v1.10 Apr 25, 2024
Copy link
Member

@sloede sloede left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks a lot! Overall, there are fewer changes than I feared.

I noticed that for some tests the difference in the reference values is quite large, showing up already at the 6-7 significant digits (i.e., well within single precision accuracy). Do you have an idea why those particular cases failed, and did you run maybe one or two of them longer and compare them manually to a v1.9 result to ensure that the simulation remains stable?

One commonality I noticed is that most tests you had to modify are either AMR or shock capturing tests. These are sometimes tricky to test, since if values are close to a cut-off threshold, even a small rounding error (as could happen with different floating optimizations) can result in a non-negligible change: A cell that is refined (or not), an element where a pure DG or a mixed DG/FV solution is used etc.

.github/workflows/Documenter.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Show resolved Hide resolved
test/test_parabolic_1d.jl Show resolved Hide resolved
test/test_parabolic_2d.jl Show resolved Hide resolved
test/test_tree_2d_euler.jl Show resolved Hide resolved
.github/workflows/benchmark.yml Outdated Show resolved Hide resolved
Co-authored-by: Michael Schlottke-Lakemper <michael@sloede.com>
Copy link
Member

@sloede sloede left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry, I missed a few spots in my initial review. They are not directly related to v1.10, but should hopefully not require any other changes and we'd have cleaned up our CI a little 😊

Once the final convo is resolved, this is good to merge from my side

.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Outdated Show resolved Hide resolved
.github/workflows/ci.yml Show resolved Hide resolved
test/test_tree_2d_eulermulti.jl Show resolved Hide resolved
DanielDoehring and others added 3 commits April 28, 2024 08:21
sloede
sloede previously approved these changes Apr 28, 2024
Copy link
Member

@sloede sloede left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thank you very much for tackling this! I'd like to get a second review from our CI wizard though...

@sloede
Copy link
Member

sloede commented Apr 28, 2024

@ranocha Could you please take a final look as well?

@ranocha
Copy link
Member Author

ranocha commented Apr 28, 2024

Not everything works on ARM:

LoadError: User-defined reduction operators are currently not supported on non-Intel architectures.
See JuliaParallel/MPI.jl#404 for more details.

see https://github.com/trixi-framework/Trixi.jl/actions/runs/8865649949/job/24342153523?pr=1562#step:7:1124

We need to update our CI settings to reflect this - can we still use older versions/architectures for some MacOS tests?

@sloede
Copy link
Member

sloede commented Apr 28, 2024

We need to update our CI settings to reflect this - can we still use older versions/architectures for some MacOS tests?

Yes. Since this is unrelated to the purpose of this PR, we should proceed here and fix the failing tests on another PR

.github/workflows/ci.yml Outdated Show resolved Hide resolved
@sloede sloede enabled auto-merge (squash) April 29, 2024 06:58
Copy link
Member

@sloede sloede left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks again @DanielDoehring for tackling this! 🙏

@sloede sloede merged commit 1ea9e3e into main Apr 29, 2024
38 checks passed
@sloede sloede deleted the hr/ci_julia110 branch April 29, 2024 07:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

4 participants