From: Xianwei Zhang <xianwei.zhang@amd.com>
Date: Thu, 24 May 2018 17:49:43 +0000 (-0400)
Subject: arch-gcn3: Implement instruction v_div_fmas_f32
X-Git-Tag: v20.1.0.0~557
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=fb7796933eb23c7730519e0f445439084b926831;p=gem5.git

arch-gcn3: Implement instruction v_div_fmas_f32

Instruction v_div_fmas_f32 was unimplemented. The
implementation was added by mimicking v_div_fmas_f64.

Change-Id: I262820a7a66877d140eb99b538715c3cae4d1860
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/29923
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Xianwei Zhang <xianwei.zhang@amd.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
Tested-by: kokoro <noreply+kokoro@google.com>
---

diff --git a/src/arch/gcn3/insts/instructions.cc b/src/arch/gcn3/insts/instructions.cc
index 2789f3e7f..308fd5dae 100644
--- a/src/arch/gcn3/insts/instructions.cc
+++ b/src/arch/gcn3/insts/instructions.cc
@@ -28879,8 +28879,49 @@ namespace Gcn3ISA
     void
     Inst_VOP3__V_DIV_FMAS_F32::execute(GPUDynInstPtr gpuDynInst)
     {
-        panicUnimplemented();
-    }
+        Wavefront *wf = gpuDynInst->wavefront();
+        ConstVecOperandF32 src0(gpuDynInst, extData.SRC0);
+        ConstVecOperandF32 src1(gpuDynInst, extData.SRC1);
+        ConstVecOperandF32 src2(gpuDynInst, extData.SRC2);
+        VecOperandF64 vdst(gpuDynInst, instData.VDST);
+
+        src0.readSrc();
+        src1.readSrc();
+        src2.readSrc();
+
+        if (instData.ABS & 0x1) {
+            src0.absModifier();
+        }
+
+        if (instData.ABS & 0x2) {
+            src1.absModifier();
+        }
+
+        if (instData.ABS & 0x4) {
+            src2.absModifier();
+        }
+
+        if (extData.NEG & 0x1) {
+            src0.negModifier();
+        }
+
+        if (extData.NEG & 0x2) {
+            src1.negModifier();
+        }
+
+        if (extData.NEG & 0x4) {
+            src2.negModifier();
+        }
+
+        for (int lane = 0; lane < NumVecElemPerVecReg; ++lane) {
+            if (wf->execMask(lane)) {
+                vdst[lane] = std::fma(src0[lane], src1[lane], src2[lane]);
+            }
+        }
+
+        //vdst.write();
+    } // execute
+    // --- Inst_VOP3__V_DIV_FMAS_F64 class methods ---
 
     Inst_VOP3__V_DIV_FMAS_F64::Inst_VOP3__V_DIV_FMAS_F64(InFmt_VOP3 *iFmt)
         : Inst_VOP3(iFmt, "v_div_fmas_f64", false)