Save\Load Data

[parfenyev]

Hello everyone,

I am studying an example from the documentation and I have a question: how can I save the result of calculation to a file and then load it from the file? I want to work with the time series of velocity field in the future (calculate the spectra, average them over time, etc.). An example of writing the results to a file and then reading them would help me a lot.

Thank you in advance!

[navidcy]

@parfenyev you raise a good point! This is related to FourierFlows/FourierFlows.jl#276. I'll try to update the FourierFlows.jl docs to sort that and update an example in GeophysicalFlows.jl to demonstrate this functionality.

[parfenyev]

I have solved the problem by converting the solution to a normal array before saving it to a file. In terms of that example one needs to replace get_sol(prob) = prob.sol by get_sol(prob) = Array(prob.sol). Perhaps this will also be useful to someone.

[navidcy]

Sorry for slacking in replying...

[navidcy]

@glwagner, I thought JLD2 can save CuArrays.... Am I wrong?

[navidcy]

@liasiegelman did you have this problem when running on GPU and saving to .jld2 files?

[glwagner]

@navidcy as far as I know saving as CuArray will work, though it's not always what you want because I don't think you'll be able to load the data without a working CUDA.

[liasiegelman]

Nop

…

Sent from my iPhone

On May 23, 2022, at 1:13 PM, Navid C. Constantinou ***@***.***> wrote:  @liasiegelman did you have this problem when running on GPU and saving to .jld2 files? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.

[navidcy]

@parfenyev could you post the output of:

julia> using Pkg; Pkg.status()

and

julia> versioninfo()

[parfenyev]

In an attempt to solve the problem, I updated the packages, now they are

      Status `~/.julia/environments/v1.7/Project.toml`
  [052768ef] CUDA v3.10.0
  [44ee3b1c] GeophysicalFlows v0.14.0
  [7073ff75] IJulia v1.23.3
  [033835bb] JLD2 v0.4.22
  [de0858da] Printf
  [9a3f8284] Random
  [10745b16] Statistics

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: Intel(R) Core(TM) i5-10600 CPU @ 3.30GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake)

The error message has also changed. The minimal example that reproduces the error has nothing to do with GeophysicalFlows.jl, so sorry to bother you. First, I have created an array on GPU and save it to file:

using JLD2, CUDA
x = CUDA.fill(1.0f0, 10)
jldsave("temp.jld2"; x)

then, I can read this file like file = jldopen("temp.jld2") that results in:

JLDFile /home/students/parfeniev/!numeric/2d_turbulence/centers/temp.jld2 (read-only)
 └─🔢 x 

But when I try to get acсess to x via file["x"] I see the error message:

UndefRefError: access to undefined reference

Stacktrace:
  [1] current_device
    @ ~/.julia/packages/CUDA/fAEDi/lib/cudadrv/devices.jl:24 [inlined]
  [2] context!(ctx::CuContext)
    @ CUDA ~/.julia/packages/CUDA/fAEDi/lib/cudadrv/state.jl:150
  [3] #context!#63
    @ ~/.julia/packages/CUDA/fAEDi/lib/cudadrv/state.jl:162 [inlined]
  [4] context!
    @ ~/.julia/packages/CUDA/fAEDi/lib/cudadrv/state.jl:161 [inlined]
  [5] unsafe_copyto!(dest::Vector{Float32}, doffs::Int64, src::CuArray{Float32, 1, CUDA.Mem.DeviceBuffer}, soffs::Int64, n::Int64)
    @ CUDA ~/.julia/packages/CUDA/fAEDi/src/array.jl:406
  [6] copyto!
    @ ~/.julia/packages/CUDA/fAEDi/src/array.jl:360 [inlined]
  [7] copyto!
    @ ~/.julia/packages/CUDA/fAEDi/src/array.jl:364 [inlined]
  [8] copyto_axcheck!
    @ ./abstractarray.jl:1104 [inlined]
  [9] Vector{Float32}(x::CuArray{Float32, 1, CUDA.Mem.DeviceBuffer})
    @ Base ./array.jl:563
 [10] Array
    @ ./boot.jl:481 [inlined]
 [11] convert
    @ ./array.jl:554 [inlined]
 [12] adapt_storage
    @ ~/.julia/packages/GPUArrays/Zecv7/src/host/abstractarray.jl:45 [inlined]
 [13] adapt_structure
    @ ~/.julia/packages/Adapt/wASZA/src/Adapt.jl:42 [inlined]
 [14] adapt
    @ ~/.julia/packages/Adapt/wASZA/src/Adapt.jl:40 [inlined]
 [15] print_array
    @ ~/.julia/packages/GPUArrays/Zecv7/src/host/abstractarray.jl:48 [inlined]
 [16] show(io::IOContext{IOBuffer}, #unused#::MIME{Symbol("text/plain")}, X::CuArray{Float32, 1, CUDA.Mem.DeviceBuffer})
    @ Base ./arrayshow.jl:396
 [17] limitstringmime(mime::MIME{Symbol("text/plain")}, x::CuArray{Float32, 1, CUDA.Mem.DeviceBuffer})
    @ IJulia ~/.julia/packages/IJulia/AQu2H/src/inline.jl:43
 [18] display_mimestring
    @ ~/.julia/packages/IJulia/AQu2H/src/display.jl:71 [inlined]
 [19] display_dict(x::CuArray{Float32, 1, CUDA.Mem.DeviceBuffer})
    @ IJulia ~/.julia/packages/IJulia/AQu2H/src/display.jl:102
 [20] #invokelatest#2
    @ ./essentials.jl:716 [inlined]
 [21] invokelatest
    @ ./essentials.jl:714 [inlined]
 [22] execute_request(socket::ZMQ.Socket, msg::IJulia.Msg)
    @ IJulia ~/.julia/packages/IJulia/AQu2H/src/execute_request.jl:112
 [23] #invokelatest#2
    @ ./essentials.jl:716 [inlined]
 [24] invokelatest
    @ ./essentials.jl:714 [inlined]
 [25] eventloop(socket::ZMQ.Socket)
    @ IJulia ~/.julia/packages/IJulia/AQu2H/src/eventloop.jl:8
 [26] (::IJulia.var"#15#18")()
    @ IJulia ./task.jl:423