IO for SF and flw for fermions

src/Dirac/Diracflw.jl

 
 
-import ..YM.flw
+import ..YM.flw, ..YM.force_gauge
+
+
+
+function Nablanabla!(so, U, si, dpar::DiracParam, dws::DiracWorkspace, lp::SpaceParm{4,6,B,D}) where {B,D}
+
+    if B == BC_SF_AFWB || B == BC_SF_ORBI
+        @timeit "DwdagDw" begin
+        
+            @timeit "SF fix" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_sfbndfix!(si,si,1.0,Gamma{16},lp)
+            end
+            end
+
+            @timeit "g5Dw" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw!(st, U, si, dpar.m0, dpar.th, lp)
+            end
+            end
+
+            @timeit "SF fix" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_sfbndfix!(st,si,dpar.ct,Gamma{5},lp)
+            end
+            end
+
+            if abs(dpar.csw) > 1.0E-10
+                @timeit "Dw_improvement" begin
+                        CUDA.@sync begin
+                            CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw_impr!(st, dws.csw, dpar.csw, si,  lp)
+                        end
+                end
+            end
+
+
+            @timeit "g5Dw" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw!(so, U, st, dpar.m0, dpar.th, lp)
+            end
+            end
+
+
+            @timeit "SF fix" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_sfbndfix!(so,st,dpar.ct,Gamma{5},lp)
+            end
+            end
+
+            if abs(dpar.csw) > 1.0E-10
+                @timeit "Dw_improvement" begin
+                        CUDA.@sync begin
+                            CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw_impr!(so, dws.csw, dpar.csw, st,  lp)
+                        end
+                end
+            end               
+
+            
+        end
+
+    else
+        @timeit "DwdagDw" begin
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw!(st, U, si, dpar.m0, dpar.th, lp)
+            end
+
+            if abs(dpar.csw) > 1.0E-10
+                @timeit "Dw_improvement" begin
+                        CUDA.@sync begin
+                            CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw_impr!(st, dws.csw, dpar.csw, si,  lp)
+                        end
+                end
+            end               
+
+            CUDA.@sync begin
+                CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw!(so, U, st, dpar.m0, dpar.th, lp)
+            end
+
+            if abs(dpar.csw) > 1.0E-10
+                @timeit "Dw_improvement" begin
+                        CUDA.@sync begin
+                            CUDA.@cuda threads=lp.bsz blocks=lp.rsz krnl_g5Dw_impr!(so, dws.csw, dpar.csw, st,  lp)
+                        end
+                end
+            end               
+
+        end
+    end 
+    
+    return nothing
+end
+
 
 function flw(U, psi, int::FlowIntr{NI,T}, ns::Int64, eps, gp::GaugeParm, dpar::DiracParam, lp::SpaceParm, ymws::YMworkspace, dws::DiracWorkspace) where {NI,T}    
     @timeit "Integrating flow equations" begin

src/YM/YMio.jl

@@ -161,19 +161,6 @@ function save_cnfg(fname::String, U, lp::SpaceParm{4,M,B,D}, gp::GaugeParm; run:
         end
     end
     
-    if B == BC_SF_AFWB || B == BC_SF_ORBI
-        for i3 in 1:lp.iL[3]
-            for id in 1:lp.ndim-1
-                assign!(id, V, i3, Ubnd[id])
-            end
-            assign!(4, V, i3, zero(M3x3{Float64}))
-        end
-        for i1 in 1:lp.iL[1]
-            for i2 in 1:lp.iL[2]
-                BDIO_write!(fb, V)
-            end
-        end
-    end        
     BDIO_write_hash!(fb)
     BDIO_close!(fb)