[PATCH] D16110: [Power9] Implement new vsx instructions: quad-precision move, fp-arithmetic
Kit Barton via llvm-commits
llvm-commits at lists.llvm.org
Wed Jan 27 13:14:15 PST 2016
kbarton added inline comments.
Comment at: lib/Target/PowerPC/PPC.td:129
@@ -127,1 +128,3 @@
+ "Enable POWER9 instructions",
+ [FeatureP8Crypto, FeatureP8Vector]>;
def FeatureDirectMove :
Why are we creating a P9 feature?
For other processors we have defined a processor model, which is a composition of other features (see P7Model and P8Model).
We need to think very carefully about how we want these models to be composed, and what can be enabled/disabled as a result of that composition (and whether that is useful). I'm not immediately convinced this is the best way to handle it.
Comment at: lib/Target/PowerPC/PPCInstrVSX.td:1776
@@ +1775,3 @@
+ // X_VT5_XO5_VB5 represents [PO VRT XO VRB XO /], usually means VRT = op VRB.
+ // TODO: Add intrinsic support, like "class VX1_Int_Ty" in PPCInstrAltivec.td
+ class X_VT5_XO5_VB5<bits<6> opcode, bits<5> xo2, bits<10> xo, string opc,
I would prefer to track this in a separate file.
Comment at: lib/Target/PowerPC/PPCInstrVSX.td:1792
@@ +1791,3 @@
+ // means VRT = VRA op VRB.
+ // TODO: Add intrinsic support, or map to llvm instruction
+ class X1_VT5_VA5_VB5<bits<6> opcode, bits<10> xo, string opc,
Same comment about tracking TODOs
More information about the llvm-commits