SpatialConvolution multithread fix

BatchNormalization.lua

@@ -16,6 +16,7 @@ function BatchNormalization:__init(nFeature, eps, momentum, affine)
    self.eps = eps or 1e-5
    self.train = true
    self.momentum = momentum or 0.1
+   self.iSize = torch.LongStorage(self.nDim):fill(0)
 
    self.running_mean = torch.zeros(nFeature)
    self.running_std = torch.ones(nFeature)
@@ -45,7 +46,7 @@ function BatchNormalization:createIODescriptors(input)
           'Only CUDA tensors are supported for cudnn.BatchNormalization!')
    if not self.iDesc or not self.oDesc or not input:isSize(self.iSize) then
       local nFeature = self.running_mean:numel()
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       self.output:resizeAs(input)
       self.gradInput:resizeAs(input)
       self.iDesc = cudnn.toDescriptor(input)

Pooling.lua

@@ -49,10 +49,11 @@ function Pooling:createIODescriptors(input)
       batch = false
    end
    assert(input:dim() == 4 and input:isContiguous());
+   self.iSize = self.iSize or torch.LongStorage(4):fill(0)
    if not self.iDesc or not self.oDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       -- resize gradInput
       self.gradInput:resizeAs(input)
       -- resize output

Pooling3D.lua

@@ -53,11 +53,12 @@ function Pooling:createIODescriptors(input)
       batch = false
    end
    assert(input:dim() == 5 and input:isContiguous());
+   self.iSize = self.iSize or torch.LongStorage(5):fill(0)
    if not self.iDesc or not self.oDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4]
    or input:size(5) ~= self.iSize[5] then
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       -- resize gradInput
       self.gradInput:resizeAs(input)
       -- resize output

SpatialConvolution.lua

@@ -23,6 +23,7 @@ function SpatialConvolution:__init(nInputPlane, nOutputPlane,
            'nOutputPlane should be divisible by nGroups')
     self.weight = torch.Tensor(nOutputPlane, nInputPlane/self.groups, kH, kW)
     self.gradWeight = torch.Tensor(nOutputPlane, nInputPlane/self.groups, kH, kW)
+    self.iSize = torch.LongStorage(4):fill(0)
     self:reset()
     -- should nil for serialization, the reset will still work
     self.reset = nil
@@ -103,7 +104,7 @@ function SpatialConvolution:createIODescriptors(input)
     if not self.iDesc or not self.oDesc or
         input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
     or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-        self.iSize = input:size()
+        self.iSize:copy(input:size())
 
         assert(self.nInputPlane == input:size(2), 'input has to contain: '
                    .. self.nInputPlane

SpatialCrossMapLRN.lua

@@ -8,6 +8,7 @@ function LRN:__init(size, alpha, beta, k)
    self.alpha = alpha or 1e-4
    self.beta = beta or 0.75
    self.k = k or 1.0
+   self.iSize = torch.LongStorage(4):fill(0)
    assert(self.size >= 1 and self.size <= 16, "size has to be between 1 and 16")
    assert(self.k >= 1e-5, "k has to be greater than 1e-5")
    assert(self.beta >= 0.01, "Beta has to be > 0.01")
@@ -32,10 +33,11 @@ function LRN:createIODescriptors(input)
       batch = false
    end
    assert(input:dim() == 4 and input:isContiguous());
+   self.iSize = self.iSize or torch.LongStorage(4):fill(0)
    if not self.iDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       self.gradInput:resizeAs(input)
       self.output:resizeAs(input)
 

SpatialDivisiveNormalization.lua

@@ -8,6 +8,7 @@ function DivisiveNorm:__init(size, alpha, beta, K)
    self.alpha = alpha or 1e-4
    self.beta = beta or 0.75
    self.K = K or 2.0
+   self.iSize = torch.LongStorage(4):fill(0)
    assert(self.size >= 1 and self.size <= 16, "size has to be between 1 and 16")
    assert(self.K >= 1e-5, "K has to be greater than 1e-5")
    assert(self.beta >= 0.01, "Beta has to be > 0.01")
@@ -32,10 +33,11 @@ function DivisiveNorm:createIODescriptors(input)
       batch = false
    end
    assert(input:dim() == 4 and input:isContiguous());
+   self.iSize = self.iSize or torch.LongStorage(4):fill(0)
    if not self.iDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       self.gradInput:resizeAs(input)
       self.output:resizeAs(input)
 

SpatialFullConvolution.lua

@@ -8,6 +8,11 @@ autotunerCache[1] = {} -- forward
 autotunerCache[2] = {} -- backwardFilter
 autotunerCache[3] = {} -- backwardData
 
+function SpatialFullConvolution:__init(...)
+    parent.__init(self, ...)
+    self.iSize = torch.LongStorage(4):fill(0)
+end
+
 -- if you change the configuration of the module manually, call this
 function SpatialFullConvolution:resetWeightDescriptors()
     assert(torch.typename(self.weight) == 'torch.CudaTensor',
@@ -73,7 +78,7 @@ function SpatialFullConvolution:createIODescriptors(input)
     if not self.iDesc or not self.oDesc or
         input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
     or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-        self.iSize = input:size()
+        self.iSize:copy(input:size())
 
         -- resize gradInput
         if self.gradInput then self.gradInput:resizeAs(input); end

SpatialSoftMax.lua

@@ -8,6 +8,7 @@ function SpatialSoftMax:__init(fast)
    else
       self.algorithm = 'CUDNN_SOFTMAX_ACCURATE'
    end
+   self.iSize = torch.LongStorage(4):fill(0)
 end
 
 function SpatialSoftMax:createIODescriptors(input)
@@ -34,7 +35,7 @@ function SpatialSoftMax:createIODescriptors(input)
    if not self.iDesc or not self.oDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
-      self.iSize = input:size()
+      self.iSize:copy(input:size())
       self.gradInput:resizeAs(input)
       self.output:resizeAs(input)
       self.iDesc = cudnn.toDescriptor(input)

TemporalConvolution.lua

@@ -19,14 +19,17 @@ function TemporalConvolution:__init(inputFrameSize, outputFrameSize,
     self.gradWeight = self.gradWeight:view(outputFrameSize, inputFrameSize*kH)
 --self.dW and self.kW now have different meaning than in nn.TemporalConvolution, because
 --W and H are switched in temporal and spatial
+    self.iSize = torch.LongStorage(4):fill(0)
 end
 
 function TemporalConvolution:createIODescriptors(input)
+    self.iSize = self.iSize or torch.LongStorage(4):fill(0)
     local sizeChanged = false
     if not self.iDesc or not self.oDesc or
         input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
     or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4] then
        sizeChanged = true
+       self.iSize:copy(input:size())
     end
     cudnn.SpatialConvolution.createIODescriptors(self,input)
     if sizeChanged then

VolumetricConvolution.lua

@@ -3,6 +3,11 @@ local VolumetricConvolution, parent
 local ffi = require 'ffi'
 local errcheck = cudnn.errcheck
 
+function VolumetricConvolution:__init(...)
+   parent.__init(self, ...)
+   self.iSize = torch.LongStorage(5):fill(0)
+end
+
 -- if you change the configuration of the module manually, call this
 function VolumetricConvolution:resetWeightDescriptors()
    assert(torch.typename(self.weight) == 'torch.CudaTensor',
@@ -30,7 +35,7 @@ end
 function VolumetricConvolution:fastest(mode)
    if mode == nil then mode = true end
    self.fastest_mode = mode
-   self.iSize = self.iSize or torch.LongStorage(4)
+   self.iSize = self.iSize or torch.LongStorage(5)
    self.iSize:fill(0)
    return self
 end
@@ -45,7 +50,7 @@ function VolumetricConvolution:setMode(fmode, bdmode, bwmode)
    if bwmode ~= nil then
       self.bwmode = bwmode
    end
-   self.iSize = self.iSize or torch.LongStorage(4)
+   self.iSize = self.iSize or torch.LongStorage(5)
    self.iSize:fill(0)
    return self
 end
@@ -65,12 +70,12 @@ function VolumetricConvolution:createIODescriptors(input)
       batch = false
    end
    assert(input:dim() == 5 and input:isContiguous());
-   self.iSize = self.iSize or torch.LongStorage(4):fill(0)
+   self.iSize = self.iSize or torch.LongStorage(5):fill(0)
    if not self.iDesc or not self.oDesc or
       input:size(1) ~= self.iSize[1] or input:size(2) ~= self.iSize[2]
    or input:size(3) ~= self.iSize[3] or input:size(4) ~= self.iSize[4]
    or input:size(5) ~= self.iSize[5] then
-         self.iSize = input:size()
+         self.iSize:copy(input:size())
          -- resize gradInput
          if self.gradInput then self.gradInput:resizeAs(input); end
          -- create input descriptor