add DCT outer butterfly svshape setup

[openpower-isa.git] / openpower / isa / simplev.mdwn

diff --git a/openpower/isa/simplev.mdwn b/openpower/isa/simplev.mdwn
index 667d1dad6c0254616f046013e10de6e4c44dbaa6..7549a19297a9f6221671b7f174662f56a9dabbe2 100644 (file)
--- a/openpower/isa/simplev.mdwn
+++ b/openpower/isa/simplev.mdwn
@@ -12,7 +12,7 @@ SVL-Form
 Pseudo-code:
 
     if (vf & (¬vs) & ¬(ms)) = 1 then
-        SVSTATE_NEXT
+        SVSTATE_NEXT(SVi)
     else
         VLimm <- SVi + 1
         if vs = 1 then
@@ -31,8 +31,8 @@ Pseudo-code:
         SVSTATE[0:6] <- MVL
         SVSTATE[7:13] <- VL
         if _RT != 0b00000 then
-            GPR(_RT) <- [0]*57 || VL
-        MSR[6] <- vf
+           GPR(_RT) <- [0]*57 || VL
+        SVSTATE[63] <- vf
 
 Special Registers Altered:
 
@@ -42,17 +42,20 @@ Special Registers Altered:
 
 SVRM-Form
 
-* svremap SVme, mi0, mi1, mi2, mo0, mo1
+* svremap SVme, mi0, mi1, mi2, mo0, mo1, pst
 
 Pseudo-code:
 
-    # place into SVREMAP SPR
-    SVREMAP[10:14] <- SVme
-    SVREMAP[0:1] <- mi0
-    SVREMAP[2:3] <- mi1
-    SVREMAP[4:5] <- mi2
-    SVREMAP[6:7] <- mo0
-    SVREMAP[8:9] <- mo1
+    # registers RA RB RC RT EA/FRS SVSHAPE0-3 indices
+    SVSTATE[32:33] <- mi0
+    SVSTATE[34:35] <- mi1
+    SVSTATE[36:37] <- mi2
+    SVSTATE[38:39] <- mo0
+    SVSTATE[40:41] <- mo1
+    # enable bit for RA RB RC RT EA/FRS
+    SVSTATE[42:46] <- SVme
+    # persistence bit (applies to more than one instruction)
+    SVSTATE[62] <- pst
 
 Special Registers Altered:
 
@@ -62,17 +65,23 @@ Special Registers Altered:
 
 SVM-Form
 
-* svshape SVxd, SVyd, SVzd, SVRM
+* svshape SVxd, SVyd, SVzd, SVRM, vf
 
 Pseudo-code:
 
+    # for convenience, VL to be calculated and stored in SVSTATE
+    vlen <- [0] * 7
+    SVSTATE[0:63] <- [0] * 64
     # clear out all SVSHAPEs
     SVSHAPE0[0:31] <- [0] * 32
     SVSHAPE1[0:31] <- [0] * 32
     SVSHAPE2[0:31] <- [0] * 32
     SVSHAPE3[0:31] <- [0] * 32
     # set schedule up for multiply
-    if (SVRM = 0b00000) then
+    if (SVRM = 0b0000) then
+        # VL in Matrix Multiply is xd*yd*zd
+        n <- (0b00 || SVxd) * (0b00 || SVyd) * (0b00 || SVzd)
+        vlen[0:6] <- n[14:20]
         # set up template in SVSHAPE0, then copy to 1-3
         SVSHAPE0[0:5] <- (0b0 || SVxd)   # xdim
         SVSHAPE0[6:11] <- (0b0 || SVyd)   # ydim
@@ -88,8 +97,16 @@ Pseudo-code:
         # FRC
         SVSHAPE2[18:20] <- 0b001          # permute x,z,y
         SVSHAPE2[28:29] <- 0b11           # skip y
-    # set schedule up for butterfly
-    if (SVRM = 0b00001) then
+    # set schedule up for FFT butterfly
+    if (SVRM = 0b0001) then
+        # calculate O(N log2 N)
+        n <- [0] * 3
+        do while n < 5
+           if SVxd[4-n] = 0 then
+               leave
+           n <- n + 1
+        n <- ((0b0 || SVxd) + 1) * n
+        vlen[0:6] <- n[1:7]
         # set up template in SVSHAPE0, then copy to 1-3
         # for FRA and FRT
         SVSHAPE0[0:5] <- (0b0 || SVxd)   # xdim
@@ -101,6 +118,55 @@ Pseudo-code:
         SVSHAPE1[28:29] <- 0b01           # j+halfstep schedule
         # FRC (coefficients)
         SVSHAPE2[28:29] <- 0b10           # k schedule
+    # set schedule up for DCT Inner butterfly
+    if (SVRM = 0b0010) then
+        # calculate O(N log2 N)
+        n <- [0] * 3
+        do while n < 5
+           if SVxd[4-n] = 0 then
+               leave
+           n <- n + 1
+        n <- ((0b0 || SVxd) + 1) * n
+        vlen[0:6] <- n[1:7]
+        # set up template in SVSHAPE0, then copy to 1-3
+        # set up FRB and FRS
+        SVSHAPE0[0:5] <- (0b0 || SVxd)   # xdim
+        SVSHAPE0[30:31] <- 0b01          # Butterfly mode
+        SVSHAPE0[18:20] <- 0b001         # DCT Inner Butterfly sub-mode
+        SVSHAPE0[21:23] <- 0b001         # "inverse" on outer loop
+        # copy
+        SVSHAPE1[0:31] <- SVSHAPE0[0:31]
+        # for FRA and FRT
+        SVSHAPE0[28:29] <- 0b01           # j+halfstep schedule
+    # set schedule up for DCT Outer butterfly
+    if (SVRM = 0b0011) then
+        # calculate O(N log2 N) number of outer butterfly overlapping adds
+        vlen[0:6] <- [0] * 7
+        n <- 0b001
+        size <- 0b0000001
+        itercount[0:6] <- (0b00 || SVxd) + 0b0000001
+        itercount[0:6] <- (0b0 || itercount[0:5])
+        do while n < 5
+           if SVxd[4-n] = 0 then
+               leave
+           n <- n + 1
+           count <- (itercount - 0b0000001) * size
+           vlen[0:6] <- vlen + count[7:13]
+           size[0:6] <- (size[1:6] || 0b0)
+           itercount[0:6] <- (0b0 || itercount[0:5])
+        # set up template in SVSHAPE0, then copy to 1-3
+        # set up FRB and FRS
+        SVSHAPE0[0:5] <- (0b0 || SVxd)   # xdim
+        SVSHAPE0[30:31] <- 0b01          # Butterfly mode
+        SVSHAPE0[18:20] <- 0b011         # DCT Outer Butterfly sub-mode
+        # copy
+        SVSHAPE1[0:31] <- SVSHAPE0[0:31]
+        # for FRA and FRT
+        SVSHAPE0[28:29] <- 0b01           # j+halfstep schedule
+    # set VL, MVL and Vertical-First
+    SVSTATE[0:6] <- vlen
+    SVSTATE[7:13] <- vlen
+    SVSTATE[63] <- vf
 
 Special Registers Altered: