IR: Change VFMaxScalarInsert to use IR::OpSize

This commit is contained in:
Ryan Houdek 2024-10-27 18:17:43 -07:00
parent bdfa8ad4f3
commit 92ee071eb2
No known key found for this signature in database

View File

@ -1731,7 +1731,7 @@
"NumElements": "RegisterSize / ElementSize",
"ImplicitFlagClobber": true
},
"FPR = VFMaxScalarInsert OpSize:#RegisterSize, u8:#ElementSize, FPR:$Vector1, FPR:$Vector2, i1:$ZeroUpperBits": {
"FPR = VFMaxScalarInsert OpSize:#RegisterSize, OpSize:#ElementSize, FPR:$Vector1, FPR:$Vector2, i1:$ZeroUpperBits": {
"Desc": ["Does a scalar 'max' between Vector1 and Vector2.",
"Inserting the result in to the lower element of Vector1 and returning the results.",
"If ZeroUpperBits is set then in a 256-bit wide operation it will zero the upper 128-bits of the destination.",