train forward pass

src/cuda/cudnn.jl

@@ -57,6 +57,7 @@ mutable struct RNNDesc{T}
   params::CuVector{T}
   weights::NTuple{2,CuMatrix{T}}
   bias::CuVector{T}
+  reserve::CuVector{UInt8}
   ptr::Ptr{Void}
 end
 
@@ -82,7 +83,8 @@ function RNNDesc{T}(mode::Int, input::Int, hidden::Int; layers = 1) where T
 
   w = cuzeros(T, rnnParamSize(T, d[], 10))
   ngates = [1, 1, 4, 3][mode+1]
-  rd = RNNDesc{T}(mode, input, hidden, w, params(w, input, hidden, ngates)..., d[])
+  # TODO: avoid reserve allocation here
+  rd = RNNDesc{T}(mode, input, hidden, w, params(w, input, hidden, ngates)..., CuVector{UInt8}(1), d[])
   finalizer(rd, x ->
     @check ccall((:cudnnDestroyRNNDescriptor,libcudnn),cudnnStatus_t,(Ptr{Void},),x))
   return rd
@@ -102,49 +104,64 @@ function rnnTrainingReserveSize(r::RNNDesc, seqlen, xdesc)
   return Int(size[])
 end
 
-function forwardInference(rnn::RNNDesc{T}, x, h, c = nothing) where T
+function getreserve(r::RNNDesc, seqlen, xdesc)
+  sz = rnnTrainingReserveSize(r, seqlen, xdesc)
+  sz ≤ length(r.reserve) ? r.reserve : (r.reserve = CuVector{UInt8}(sz))
+end
+
+function cudnnRNNForward(::Type{T}, rnn, seqlen, xd, x, hd, h, cd, c, wd, w, yd, y, hod, ho, cod, co, workspace, reserve=nothing) where T
+  if reserve == nothing
+    @check ccall((:cudnnRNNForwardInference, libcudnn), cudnnStatus_t,
+                 (Ptr{Void}, Ptr{Void}, Cint,
+                  Ptr{Ptr{Void}}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Ptr{Void}}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T},
+                  Ptr{Void}, Csize_t),
+                 libcudnn_handle[], rnn, seqlen,
+                 xd, x, hd, h, cd, c, wd, w, yd, y, hod, ho, cod, co,
+                 workspace, length(workspace))
+  else
+    @check ccall((:cudnnRNNForwardTraining, libcudnn), cudnnStatus_t,
+                 (Ptr{Void}, Ptr{Void}, Cint,
+                  Ptr{Ptr{Void}}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Ptr{Void}}, Ptr{T}, Ptr{Void}, Ptr{T}, Ptr{Void}, Ptr{T},
+                  Ptr{Void}, Csize_t, Ptr{Void}, Csize_t),
+                 libcudnn_handle[], rnn, seqlen,
+                 xd, x, hd, h, cd, c, wd, w, yd, y, hod, ho, cod, co,
+                 workspace, length(workspace), reserve, length(reserve))
+  end
+end
+
+function forward(rnn::RNNDesc{T}, x::CuArray{T}, h::CuArray{T}, c = nothing; train = Val{false}) where T
   @assert size(x, 1) == rnn.input
   @assert size(h, 1) == rnn.hidden
   @assert size(x, 2) == size(h, 2)
   seqLength = 1
-  xdesc = [TensorDesc(reshape(x, 1, size(x, 1), size(x, 2)))]
+  xdesc = [TensorDesc(T, (1, size(x, 1), size(x, 2)))]
   y = x isa AbstractVector ? similar(x, rnn.hidden) : similar(x, rnn.hidden, size(x, 2))
-  ydesc = [TensorDesc(reshape(y, 1, size(y, 1), size(y, 2)))]
-  hout = similar(h)
+  ydesc = [TensorDesc(T, (1, size(y, 1), size(y, 2)))]
   workspace = CuVector{UInt8}(rnnWorkspaceSize(rnn, seqLength, xdesc)) # TODO: reuse this
+  reserve = train == Val{true} ? getreserve(rnn, seqLength, xdesc) : nothing
   if c ≠ nothing
     @assert size(c, 1) == rnn.hidden
     @assert size(c, 2) == size(h, 2)
     cptr = c
-    cdesc = TensorDesc(reshape(c, size(c, 1), size(c, 2), 1))
+    cdesc = TensorDesc(T, (size(c, 1), size(c, 2), 1))
     cout = similar(c)
-    coutdesc = TensorDesc(reshape(cout, size(cout, 1), size(cout, 2), 1))
+    coutdesc = TensorDesc(T, (size(cout, 1), size(cout, 2), 1))
   else
     cptr = cdesc = cout = coutdesc = C_NULL
   end
-  @check ccall((:cudnnRNNForwardInference, libcudnn), cudnnStatus_t,
-               (Ptr{Void}, Ptr{Void}, Cint,
-                Ptr{Ptr{Void}}, Ptr{T},
-                Ptr{Void}, Ptr{T},
-                Ptr{Void}, Ptr{T},
-                Ptr{Void}, Ptr{T},
-                Ptr{Ptr{Void}}, Ptr{T},
-                Ptr{Void}, Ptr{T},
-                Ptr{Void}, Ptr{T},
-                Ptr{Void}, Csize_t),
-               libcudnn_handle[], rnn, seqLength,
-               xdesc, x,
-               TensorDesc(reshape(h, size(h, 1), size(h, 2), 1)), h,
-               cdesc, cptr,
-               TensorDesc(reshape(rnn.params, 1, 1, :)), rnn.params,
-               ydesc, y,
-               TensorDesc(reshape(hout, size(hout, 1), size(hout, 2), 1)), hout,
-               coutdesc, cout,
-               workspace, length(workspace))
+  cudnnRNNForward(T, rnn, seqLength,
+                  xdesc, x,
+                  TensorDesc(T, (size(h, 1), size(h, 2), 1)), h,
+                  cdesc, cptr,
+                  TensorDesc(T, (1, 1, length(rnn.params))), rnn.params,
+                  ydesc, y,
+                  C_NULL, C_NULL, # hout
+                  coutdesc, cout,
+                  workspace, reserve)
   if c == nothing
-    return y, hout
+    return y, y
   else
-    return y, hout, cout
+    return y, y, cout
   end
 end
 
@@ -197,17 +214,19 @@ function desc(rnn)
   return d
 end
 
+istrain(m::CuRNNs, args...) = any(x -> x isa TrackedArray, (m.Wi, m.Wh, m.b, args...))
+
 function (m::CuRNN{T})(h::CuParam{T}, x::CuParam{T}) where T <: Union{Float32,Float64}
-  y, h = forwardInference(desc(m), Flux.data(x), Flux.data(h))
+  y, h = forward(desc(m), Flux.data(x), Flux.data(h), train = Val{istrain(m, h, x)})
   return h, y
 end
 
 function (m::CuGRU{T})(h::CuParam{T}, x::CuParam{T}) where T <: Union{Float32,Float64}
-  y, h = forwardInference(desc(m), Flux.data(x), Flux.data(h))
+  y, h = forward(desc(m), Flux.data(x), Flux.data(h), train = Val{istrain(m, h, x)})
   return h, y
 end
 
 function (m::CuLSTM{T})(h::NTuple{2,CuParam{T}}, x::CuParam{T}) where T <: Union{Float32,Float64}
-  y, h, c = forwardInference(desc(m), Flux.data(x), Flux.data.(h)...)
+  y, h, c = forward(desc(m), Flux.data(x), Flux.data.(h)..., train = Val{istrain(m, h, x)})
   return (h, c), y
 end