NVTX

CliMA · Sep 20, 2024 · 5d1ae5a · 5d1ae5a
1 parent d31df8d
commit 5d1ae5a
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 3 deletions.
diff --git a/src/GPU/DiagnosticKernel.jl b/src/GPU/DiagnosticKernel.jl
@@ -177,7 +177,7 @@ end
   end
 end
 
-function FcnPrepareGPU!(U,FE,Metric,Phys,Cache,Exchange,Global,Param,DiscType)
+NVTX.@annotate function FcnPrepareGPU!(U,FE,Metric,Phys,Cache,Exchange,Global,Param,DiscType)
   backend = get_backend(U)
   dXdxI = Metric.dXdxI
   X = Metric.X

diff --git a/src/GPU/FcnGPU.jl b/src/GPU/FcnGPU.jl
@@ -136,7 +136,7 @@ function FcnAdvectionGPU!(F,U,time,FE,Metric,Phys,Cache,Exchange,Global,Param,Pr
   @views Parallels.ExchangeData3DRecvGPU!(F[:,:,1:1+NumV],Exchange)
 end
 
-function FcnGPU!(F,U,FE,Metric,Phys,Cache,Exchange,Global,Param,Equation::Models.CompressibleShallow)
+NVTX.@annotate function FcnGPU!(F,U,FE,Metric,Phys,Cache,Exchange,Global,Param,Equation::Models.CompressibleShallow)
 
   backend = get_backend(F)
   FT = eltype(F)

diff --git a/src/GPU/GPU.jl b/src/GPU/GPU.jl
@@ -10,6 +10,7 @@ using StaticArrays
 using KernelAbstractions
 using KernelAbstractions: @atomic, @atomicswap, @atomicreplace
 using KernelAbstractions.Extras
+using NVTX
 
 include("OperatorKernel.jl")
 include("FcnGPU.jl")

diff --git a/src/Integration/SchurSolve.jl b/src/Integration/SchurSolve.jl
@@ -174,7 +174,7 @@ end
   end    
 end
 
-function SchurSolveGPU!(k,v,J,fac,Cache,Global)
+NVTX.@annotate function SchurSolveGPU!(k,v,J,fac,Cache,Global)
   backend = get_backend(k)
   FT = eltype(k)