Use views to avoid thread issues

src/DAT/DAT.jl

@@ -691,7 +691,11 @@ function runLoop(dc::DATConfig, showprog)
         mapfun(CachingPool(workers()),allRanges, on_error=identity) do r
             incaches, outcaches, args = getallargs(dc)
             updateinars(dc, r, incaches)
-            innerLoop(r, args...)
+            if dc.ntr[myid()] > 1
+                innerLoop_threaded(r, args...)
+            else
+                innerLoop(r,args...)
+            end
             writeoutars(dc, r, outcaches)
             dc.do_gc && GC.gc()
         end
@@ -700,7 +704,12 @@ function runLoop(dc::DATConfig, showprog)
         mapfun = showprog ? progress_map : map
         mapfun(allRanges) do r
             updateinars(dc, r, incaches)
-            innerLoop(r, args...)
+            if dc.ntr[1] > 1
+                innerLoop_threaded(r, args...)
+            else
+                @show "Running nonthreaded"
+                innerLoop(r,args...)
+            end
             writeoutars(dc, r, outcaches)
             dc.do_gc && GC.gc()
         end
@@ -726,7 +735,6 @@ getlaxvals(a::AllLoopAxes, cI, offscur) = (
 function getallargs(dc::DATConfig)
     incache, outcache = getCubeCache(dc)
     filters = map(ic -> ic.desc.procfilter, dc.incubes)
-    inworkar, outworkar = generateworkarrays(dc)
     axvals = if dc.include_loopvars
         lax = (dc.LoopAxes...,)
         AllLoopAxes(lax)
@@ -760,7 +768,7 @@ function getallargs(dc::DATConfig)
     end
     incache,
     outcache,
-    (fu, inarsbc, outarsbc, filters, inworkar, outworkar, axvals, adda, kwa)
+    (fu, inarsbc, outarsbc, filters, axvals, adda, kwa)
 end
 
 
@@ -1078,32 +1086,23 @@ function distributeLoopRanges(block_size::NTuple{N,Int}, loopR::NTuple{N,Int}, c
     Iterators.product(allranges...)
 end
 
-function generateworkarrays(dc::DATConfig)
-    inwork = map(i -> getworkarray(i, dc.ntr[myid()]), dc.incubes)
-    outwork = map(i -> getworkarray(i, dc.ntr[myid()]), dc.outcubes)
-    inwork, outwork
-end
-
 function innercode(
     f,
     cI,
     xinBC,
     xoutBC,
     filters,
-    inwork,
-    outwork,
     axvalcreator,
     offscur,
     addargs,
     kwargs,
 )
-    ithr = Threads.threadid()
-    #Pick the correct array according to thread
-    myinwork = map(i -> i[ithr], inwork)
-    myoutwork = map(i -> i[ithr], outwork)
     #Copy data into work arrays
-    foreach(myinwork, xinBC) do iw, x
-        YAXArrayBase.getdata(iw) .= view(x, cI.I...)
+    myinwork = map(xinBC) do x
+        view(x, cI.I...)
+    end
+    myoutwork = map(xoutBC) do x
+        view(x, cI.I...)
     end
     #Apply filters
     mvs = map(docheck, filters, myinwork)
@@ -1116,57 +1115,59 @@ function innercode(
         #Finally call the function
         f(myoutwork..., myinwork..., laxval..., addargs...; kwargs...)
     end
-    #Copy data into output array
-    foreach((iw, x) -> view(x, cI.I...) .= YAXArrayBase.getdata(iw), myoutwork, xoutBC)
 end
 
 using DataStructures: OrderedDict
 using Base.Cartesian
+@noinline function innerLoop_threaded(
+    loopRanges,
+    f,
+    xinBC,
+    xoutBC,
+    filters,
+    axvalcreator,
+    addargs,
+    kwargs,
+)
+    offscur = map(i -> (first(i) - 1), loopRanges)
+    Threads.@threads for cI in CartesianIndices(map(i -> 1:length(i), loopRanges))
+        innercode(
+            f,
+            cI,
+            xinBC,
+            xoutBC,
+            filters,
+            axvalcreator,
+            offscur,
+            addargs,
+            kwargs,
+        )
+    end
+end
+
 @noinline function innerLoop(
     loopRanges,
     f,
     xinBC,
     xoutBC,
     filters,
-    inwork,
-    outwork,
     axvalcreator,
     addargs,
     kwargs,
 )
     offscur = map(i -> (first(i) - 1), loopRanges)
-    if length(inwork[1]) == 1
-        for cI in CartesianIndices(map(i -> 1:length(i), loopRanges))
-            innercode(
-                f,
-                cI,
-                xinBC,
-                xoutBC,
-                filters,
-                inwork,
-                outwork,
-                axvalcreator,
-                offscur,
-                addargs,
-                kwargs,
-            )
-        end
-    else
-        Threads.@threads :static for cI in CartesianIndices(map(i -> 1:length(i), loopRanges))
-            innercode(
-                f,
-                cI,
-                xinBC,
-                xoutBC,
-                filters,
-                inwork,
-                outwork,
-                axvalcreator,
-                offscur,
-                addargs,
-                kwargs,
-            )
-        end
+    for cI in CartesianIndices(map(i -> 1:length(i), loopRanges))
+        innercode(
+            f,
+            cI,
+            xinBC,
+            xoutBC,
+            filters,
+            axvalcreator,
+            offscur,
+            addargs,
+            kwargs,
+        )
     end
 end