Port widening ops to ISLE (AArch64) (#4751)

Ported the existing implementations of the following opcodes for AArch64
to ISLE, and implemented support for 64-bit vectors (per the docs):
- `SwidenLow`
- `SwidenHigh`
- `UwidenLow`
- `UwidenHigh`

Also ported `WideningPairwiseDotProductS` as-is.

Copyright (c) 2022 Arm Limited
This commit is contained in:
Damian Heaton
2022-08-23 17:42:11 +01:00
committed by GitHub
parent da1fb305a3
commit 3b68d76905
10 changed files with 250 additions and 161 deletions

View File

@@ -98,3 +98,23 @@
;;; Rules for `extract_vector` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(rule (lower (extract_vector x 0))
(value_reg (fpu_move_128 (put_in_reg x))))
;;;; Rules for `swiden_low` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(rule (lower (has_type ty (swiden_low x)))
(value_reg (vec_extend (VecExtendOp.Sxtl) x $false (lane_size ty))))
;;;; Rules for `swiden_high` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(rule (lower (has_type ty (swiden_high x)))
(value_reg (vec_extend (VecExtendOp.Sxtl) x $true (lane_size ty))))
;;;; Rules for `uwiden_low` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(rule (lower (has_type ty (uwiden_low x)))
(value_reg (vec_extend (VecExtendOp.Uxtl) x $false (lane_size ty))))
;;;; Rules for `uwiden_high` ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(rule (lower (has_type ty (uwiden_high x)))
(value_reg (vec_extend (VecExtendOp.Uxtl) x $true (lane_size ty))))