Merge pull request #120 from FourierFlows/ImplementCUDAapi

Implement CUDAapi

Author: glwagner

.gitignore

@@ -12,3 +12,4 @@
 docs/build/
 docs/site/
 coverage/
+docs/Manifest.toml

.travis.yml

@@ -5,7 +5,7 @@ os:
   - linux
 
 julia:
-  - 1.1
+  - 1.2
   - nightly
 
 matrix:
@@ -19,7 +19,7 @@ after_success:
 jobs:
   include:
     - stage: "Documentation"
-      julia: 1.1
+      julia: 1.2
       os: linux
       script:
         - julia --project=docs/ -e 'using Pkg; Pkg.instantiate();

Manifest.toml

@@ -6,6 +6,12 @@ git-tree-sha1 = "380e36c66edfa099cd90116b24c1ce8cafccac40"
 uuid = "621f4979-c628-5d54-868e-fcf4e3e8185c"
 version = "0.4.1"
 
+[[Adapt]]
+deps = ["LinearAlgebra"]
+git-tree-sha1 = "82dab828020b872fa9efd3abec1152b075bc7cbf"
+uuid = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
+version = "1.0.0"
+
 [[AxisAlgorithms]]
 deps = ["LinearAlgebra", "Random", "SparseArrays", "WoodburyMatrices"]
 git-tree-sha1 = "a4d07a1c313392a77042855df46c5f534076fab9"
@@ -21,6 +27,35 @@ git-tree-sha1 = "c7361ce8a2129f20b0e05a89f7070820cfed6648"
 uuid = "b99e7846-7c00-51b0-8f62-c81ae34c0232"
 version = "0.5.6"
 
+[[CEnum]]
+git-tree-sha1 = "62847acab40e6855a9b5905ccb99c2b5cf6b3ebb"
+uuid = "fa961155-64e5-5f13-b03f-caf6b980ea82"
+version = "0.2.0"
+
+[[CSTParser]]
+deps = ["Tokenize"]
+git-tree-sha1 = "c69698c3d4a7255bc1b4bc2afc09f59db910243b"
+uuid = "00ebfdb7-1f24-5e51-bd34-a7502290713f"
+version = "0.6.2"
+
+[[CUDAapi]]
+deps = ["Libdl", "Logging"]
+git-tree-sha1 = "9b2b4b71d6b7f946c9689bb4dea03ff92e3c7091"
+uuid = "3895d2a7-ec45-59b8-82bb-cfc6a382f9b3"
+version = "1.1.0"
+
+[[CUDAdrv]]
+deps = ["CUDAapi", "Libdl", "Printf"]
+git-tree-sha1 = "9ce99b5732c70e06ed97c042187baed876fb1698"
+uuid = "c5f51814-7f29-56b8-a69c-e4d8f6be1fde"
+version = "3.1.0"
+
+[[CUDAnative]]
+deps = ["Adapt", "CUDAapi", "CUDAdrv", "DataStructures", "InteractiveUtils", "LLVM", "Libdl", "Logging", "Printf", "TimerOutputs"]
+git-tree-sha1 = "0a00bef482b7c9127495c7f4a2a85e73b13b5af8"
+uuid = "be33ccc6-a3ff-5ff2-a52e-74243cff1e17"
+version = "2.3.0"
+
 [[CodecZlib]]
 deps = ["BinaryProvider", "Libdl", "TranscodingStreams"]
 git-tree-sha1 = "05916673a2627dd91b4969ff8ba6941bc85a960e"
@@ -45,6 +80,18 @@ git-tree-sha1 = "f9780daf3fea51ad2d7b7ed4f480ac34ab36587f"
 uuid = "a2441757-f6aa-5fb2-8edb-039e3f45d037"
 version = "0.9.2"
 
+[[Crayons]]
+deps = ["Test"]
+git-tree-sha1 = "f621b8ef51fd2004c7cf157ea47f027fdeac5523"
+uuid = "a8cc5b0e-0ffa-5ad4-8c14-923d3ee1735f"
+version = "4.0.0"
+
+[[CuArrays]]
+deps = ["AbstractFFTs", "Adapt", "CUDAapi", "CUDAdrv", "CUDAnative", "GPUArrays", "LinearAlgebra", "MacroTools", "NNlib", "Printf", "Random", "Requires", "SparseArrays", "TimerOutputs"]
+git-tree-sha1 = "46b48742a84bb839e74215b7e468a4a1c6ba30f9"
+uuid = "3a865a2d-5b23-5a0f-bc46-62713ec82fae"
+version = "1.2.1"
+
 [[DataStructures]]
 deps = ["InteractiveUtils", "OrderedCollections"]
 git-tree-sha1 = "0809951a1774dc724da22d26e4289bbaab77809a"
@@ -75,11 +122,23 @@ git-tree-sha1 = "351f001a78aa1b7ad2696e386e110b5abd071c71"
 uuid = "5789e2e9-d7fb-5bc7-8068-2c6fae9b9549"
 version = "1.0.7"
 
+[[FillArrays]]
+deps = ["LinearAlgebra", "Random", "SparseArrays"]
+git-tree-sha1 = "8fba6ddaf66b45dec830233cea0aae43eb1261ad"
+uuid = "1a297f60-69ca-5386-bcde-b61e274b549b"
+version = "0.6.4"
+
+[[GPUArrays]]
+deps = ["Adapt", "FFTW", "FillArrays", "LinearAlgebra", "Printf", "Random", "Serialization", "StaticArrays", "Test"]
+git-tree-sha1 = "dd169c636d1d3656a9faca772f5bd7c226a61254"
+uuid = "0c68f7d7-f131-5f86-a1c3-88cf8149b2d7"
+version = "1.0.1"
+
 [[HTTP]]
 deps = ["Base64", "Dates", "IniFile", "MbedTLS", "Sockets"]
-git-tree-sha1 = "03ddc88af7f2d963fac5aa9f3ac8e11914d68a78"
+git-tree-sha1 = "c4a527dba1d26add0e85946e1a53f42a1b343acc"
 uuid = "cd3eb016-35fb-5094-929b-558a96fad6f3"
-version = "0.8.4"
+version = "0.8.5"
 
 [[IniFile]]
 deps = ["Test"]
@@ -109,6 +168,12 @@ git-tree-sha1 = "b34d7cef7b337321e97d22242c3c2b91f476748e"
 uuid = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
 version = "0.21.0"
 
+[[LLVM]]
+deps = ["CEnum", "Libdl", "Printf", "Unicode"]
+git-tree-sha1 = "52cfea426bd248a427aace7d88eb5d45b84ea297"
+uuid = "929cbde3-209d-540e-8aea-75f648917ca0"
+version = "1.2.0"
+
 [[LibGit2]]
 uuid = "76f85450-5226-5b5a-8eaa-529ad045b433"
 
@@ -122,19 +187,31 @@ uuid = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 [[Logging]]
 uuid = "56ddb016-857b-54e1-b83d-db4d58db5568"
 
+[[MacroTools]]
+deps = ["CSTParser", "Compat", "DataStructures", "Test", "Tokenize"]
+git-tree-sha1 = "d6e9dedb8c92c3465575442da456aec15a89ff76"
+uuid = "1914dd2f-81c6-5fcd-8719-6d5c9610ff09"
+version = "0.5.1"
+
 [[Markdown]]
 deps = ["Base64"]
 uuid = "d6f4376e-aef5-505a-96c1-9c027394607a"
 
 [[MbedTLS]]
-deps = ["BinaryProvider", "Dates", "Distributed", "Libdl", "Random", "Sockets", "Test"]
-git-tree-sha1 = "2d94286a9c2f52c63a16146bb86fd6cdfbf677c6"
+deps = ["BinaryProvider", "Dates", "Libdl", "Random", "Sockets"]
+git-tree-sha1 = "85f5947b53c8cfd53ccfa3f4abae31faa22c2181"
 uuid = "739be429-bea8-5141-9913-cc70e7f3736d"
-version = "0.6.8"
+version = "0.7.0"
 
 [[Mmap]]
 uuid = "a63ad114-7e13-5084-954f-fe012c677804"
 
+[[NNlib]]
+deps = ["Libdl", "LinearAlgebra", "Requires", "Statistics", "TimerOutputs"]
+git-tree-sha1 = "0c667371391fc6bb31f7f12f96a56a17098b3de8"
+uuid = "872c559c-99b0-510c-b3b7-b6c96a88d5cd"
+version = "0.6.0"
+
 [[OffsetArrays]]
 git-tree-sha1 = "1af2f79c7eaac3e019a0de41ef63335ff26a0a57"
 uuid = "6fe1bfb0-de20-5000-8ca7-80f57d26f881"
@@ -148,9 +225,9 @@ version = "1.1.0"
 
 [[Parsers]]
 deps = ["Dates", "Test"]
-git-tree-sha1 = "db2b35dedab3c0e46dc15996d170af07a5ab91c9"
+git-tree-sha1 = "ef0af6c8601db18c282d092ccbd2f01f3f0cd70b"
 uuid = "69de0a69-1ddd-5017-9359-2bf0b02dc9f0"
-version = "0.3.6"
+version = "0.3.7"
 
 [[Pkg]]
 deps = ["Dates", "LibGit2", "Markdown", "Printf", "REPL", "Random", "SHA", "UUIDs"]
@@ -217,6 +294,17 @@ uuid = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 deps = ["Distributed", "InteractiveUtils", "Logging", "Random"]
 uuid = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
+[[TimerOutputs]]
+deps = ["Crayons", "Printf", "Test", "Unicode"]
+git-tree-sha1 = "b80671c06f8f8bae08c55d67b5ce292c5ae2660c"
+uuid = "a759f4b9-e2f1-59dc-863e-4aeb61b1ea8f"
+version = "0.5.0"
+
+[[Tokenize]]
+git-tree-sha1 = "dfcdbbfb2d0370716c815cbd6f8a364efb6f42cf"
+uuid = "0796e94c-ce3b-5d07-9a54-7f471281c624"
+version = "0.5.6"
+
 [[TranscodingStreams]]
 deps = ["Random", "Test"]
 git-tree-sha1 = "7c53c35547de1c5b9d46a4797cf6d8253807108c"

Project.toml

@@ -6,11 +6,13 @@ authors = ["Gregory L. Wagner <wagner.greg@gmail.com>", "Navid C. Constantinou <
 description = "Tools for solving partial differential equations on periodic domains using Fourier-based pseudospectral methods."
 documentation = "https://fourierflows.github.io/FourierFlows.jl/dev/"
 repository = "https://github.com/FourierFlows/FourierFlows.jl"
-version ="0.3.1"
+version = "0.3.1"
 versions = ["0.0.1", "0.0.2", "0.1.0", "0.1.1", "0.1.2", "0.1.3", "0.2.0", "0.2.1", "0.2.2", "0.3.0", "0.3.1"]
 
 [deps]
+CUDAapi = "3895d2a7-ec45-59b8-82bb-cfc6a382f9b3"
 Coverage = "a2441757-f6aa-5fb2-8edb-039e3f45d037"
+CuArrays = "3a865a2d-5b23-5a0f-bc46-62713ec82fae"
 FFTW = "7a1cc6ca-52ef-59f5-83cd-3a7055c09341"
 Interpolations = "a98d9a8b-a2ab-59e6-89dd-64a1c18fca59"
 JLD2 = "033835bb-8acc-5ee8-8aae-3f567f8a3819"
@@ -22,14 +24,14 @@ Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 
 [compat]
+CuArrays = "≥ 1.2.1"
 FFTW = "≥ 0.2.4"
 Interpolations = "≥ 0.8.0"
 JLD2 = "≥ 0.1.2"
 julia = "≥ 0.7.0"
 
 [extras]
-CuArrays = "3a865a2d-5b23-5a0f-bc46-62713ec82fae"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "CuArrays"]
+test = ["Test"]

appveyor.yml

@@ -1,6 +1,6 @@
 environment:
   matrix:
-  - julia_version: 1
+  - julia_version: 1.2
   - julia_version: nightly
 
 platform:

src/FourierFlows.jl

@@ -2,8 +2,7 @@ module FourierFlows
 
 export
   # Helper variables and macros for determining if machine is CUDA-enabled.
-  HAVE_CUDA,
-  @hascuda,
+  @has_cuda,
 
   Device,
   CPU,
@@ -59,6 +58,7 @@ using
   JLD2,
   Statistics,
   Interpolations,
+  CUDAapi,
   Requires
 
 import Base: resize!, getindex, setindex!, lastindex, push!, append!
@@ -89,15 +89,18 @@ include("timesteppers.jl")
 include("diffusion.jl")
 
 # Import CUDA utilities if cuda is detected.
-const HAVE_CUDA = try
-    using CuArrays
-    true
-catch
-    false
+if has_cuda()
+  try
+    using CuArrays # we have CUDA, so this should not fail
+  catch ex
+    # something is wrong with the user's set-up (or there's a bug in CuArrays)
+    @warn "CUDA is installed, but CuArrays.jl fails to load" exception=(ex, catch_backtrace())
+  end
 end
 
-macro hascuda(ex)
-    return HAVE_CUDA ? :($(esc(ex))) : :(nothing)
+
+macro has_cuda(ex)
+  return has_cuda() ? :($(esc(ex))) : :(nothing)
 end
 
 

src/diffusion.jl

@@ -45,9 +45,9 @@ Params(dev, kappa::Number) = Params(kappa)
 Params(dev, kappa::AbstractArray) = Params(ArrayType(dev)(kappa))
 
 """
-    Equation(p, g)
+    DiffusionEquation(dev, kappa, grid)
 
-Returns the equation for constant diffusivity problem with params p and grid g.
+Returns the equation for constant diffusivity problem with diffusivity kappa and grid.
 """
 function DiffusionEquation(dev::Device, kappa::T, grid) where T<:Number
   L = zeros(dev, T, grid.nkr)

test/createffttestfunctions.jl

@@ -86,7 +86,7 @@ function create_testfuncs(g::TwoDGrid{Tg,<:Array}) where Tg
     f₁, f₂, f₁h, f₂h, f₁hr, f₂hr, f₁hr_mul, f₂hr_mul, f₁h_analytical, f₁hr_analytical, f₂h_analytical, f₂hr_analytical
 end
 
-@hascuda begin
+@has_cuda begin
   function create_testfuncs(g::OneDGrid{Tg, <:CuArray}) where Tg
     cpugrid = OneDGrid(g.nx, g.Lx)
     out = create_testfuncs(cpugrid)

test/runtests.jl

@@ -13,8 +13,8 @@ using LinearAlgebra: mul!, ldiv!, norm
 
 # the devices on which tests will run
 devices = (CPU(),)
-@hascuda devices = (CPU(), GPU())
-@hascuda using CuArrays
+@has_cuda devices = (CPU(), GPU())
+@has_cuda using CuArrays
 
 const rtol_fft = 1e-12
 const rtol_output = 1e-12