From e548e29a6cee7c30cdb5e66185206599b090131d Mon Sep 17 00:00:00 2001
From: Luke Kenneth Casson Leighton <lkcl@lkcl.net>
Date: Sat, 31 Jul 2021 13:10:22 +0100
Subject: [PATCH] add iDCT modes to interim svshape instruction pseudo-code

---
 openpower/isa/simplev.mdwn | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)

diff --git a/openpower/isa/simplev.mdwn b/openpower/isa/simplev.mdwn
index a48c405e..00410112 100644
--- a/openpower/isa/simplev.mdwn
+++ b/openpower/isa/simplev.mdwn
@@ -148,10 +148,11 @@ Pseudo-code:
         SVSHAPE1[28:29] <- 0b01           # j+halfstep schedule
         # FRC (coefficients)
         SVSHAPE2[28:29] <- 0b10           # k schedule
-    # set schedule up for DCT Inner butterfly
-    # SVRM Mode 2 is for pre-calculated coefficients,
-    # SVRM Mode 4 is for on-the-fly (Vertical-First Mode)
-    if (SVRM = 0b0010) | (SVRM = 0b0100) then
+    # set schedule up for (i)DCT Inner butterfly
+    # SVRM Mode 2 (Mode 6 for iDCT) is for pre-calculated coefficients,
+    # SVRM Mode 4 (Mode 12 for iDCT) is for on-the-fly (Vertical-First Mode)
+    if ((SVRM = 0b0010) | (SVRM = 0b0100) |
+        (SVRM = 0b1010) | (SVRM = 0b1100)) then
         # calculate O(N log2 N)
         n <- [0] * 3
         do while n < 5
@@ -163,11 +164,14 @@ Pseudo-code:
         # 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          # DCT/FFT mode
-        if (SVRM = 0b0100) then
-            SVSHAPE0[6:11] <- 0b000011       # DCT Inner Butterfly mode 4
+        if (SVRM = 0b1010) | (SVRM = 0b1100) then
+            SVSHAPE0[30:31] <- 0b11          # iDCT mode
+        else
+            SVSHAPE0[30:31] <- 0b01          # DCT mode
+        if (SVRM = 0b1100) | (SVRM = 0b0100) then
+            SVSHAPE0[6:11] <- 0b000011       # (i)DCT Inner Butterfly mode 4
         else
-            SVSHAPE0[6:11] <- 0b000001       # DCT Inner Butterfly mode 2
+            SVSHAPE0[6:11] <- 0b000001       # (i)DCT Inner Butterfly mode 2
         SVSHAPE0[18:20] <- 0b001         # DCT Inner Butterfly sub-mode
         SVSHAPE0[21:23] <- 0b001         # "inverse" on outer loop
         # copy
@@ -181,8 +185,8 @@ Pseudo-code:
         SVSHAPE2[28:29] <- 0b10           # ci (k for mode 4) schedule
         if (SVRM != 0b0100) then
             SVSHAPE3[28:29] <- 0b11           # size schedule
-    # set schedule up for DCT Outer butterfly
-    if (SVRM = 0b0011) then
+    # set schedule up for (i)DCT Outer butterfly
+    if (SVRM = 0b0011) | (SVRM = 0b1011) then
         # calculate O(N log2 N) number of outer butterfly overlapping adds
         vlen[0:6] <- [0] * 7
         n <- 0b000
@@ -200,9 +204,13 @@ Pseudo-code:
         # 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          # DCT/FFT mode
+        if (SVRM = 0b1011) then
+            SVSHAPE0[30:31] <- 0b11          # iDCT mode
+            SVSHAPE0[18:20] <- 0b011         # iDCT Outer Butterfly sub-mode
+        else
+            SVSHAPE0[30:31] <- 0b01          # DCT mode
+            SVSHAPE0[18:20] <- 0b100         # DCT Outer Butterfly sub-mode
         SVSHAPE0[6:11] <- 0b000010       # DCT Butterfly mode
-        SVSHAPE0[18:20] <- 0b100         # DCT Outer Butterfly sub-mode
         # copy
         SVSHAPE1[0:31] <- SVSHAPE0[0:31]
         SVSHAPE2[0:31] <- SVSHAPE0[0:31]
-- 
2.30.2