Index: llvm/test/CodeGen/RISCV/addimm-mulimm.ll =================================================================== --- llvm/test/CodeGen/RISCV/addimm-mulimm.ll +++ llvm/test/CodeGen/RISCV/addimm-mulimm.ll @@ -2,97 +2,438 @@ ;; Test that (mul (add x, c1), c2) can be transformed to ;; (add (mul x, c2), c1*c2) if profitable. -; RUN: llc -mtriple=riscv32 -mattr=+m -verify-machineinstrs < %s \ +; RUN: llc -mtriple=riscv32 -mattr=+m,+experimental-zba -verify-machineinstrs < %s \ ; RUN: | FileCheck -check-prefix=RV32IM %s -; RUN: llc -mtriple=riscv64 -mattr=+m -verify-machineinstrs < %s \ +; RUN: llc -mtriple=riscv64 -mattr=+m,+experimental-zba -verify-machineinstrs < %s \ ; RUN: | FileCheck -check-prefix=RV64IM %s -define signext i32 @add_mul_trans_accept_1(i32 %x) { -; RV32IM-LABEL: add_mul_trans_accept_1: +define i32 @add_mul_trans_reject_a1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_reject_a1: ; RV32IM: # %bb.0: -; RV32IM-NEXT: addi a1, zero, 11 +; RV32IM-NEXT: addi a1, zero, 29 ; RV32IM-NEXT: mul a0, a0, a1 -; RV32IM-NEXT: addi a0, a0, 407 +; RV32IM-NEXT: addi a0, a0, 1073 ; RV32IM-NEXT: ret ; -; RV64IM-LABEL: add_mul_trans_accept_1: +; RV64IM-LABEL: add_mul_trans_reject_a1: ; RV64IM: # %bb.0: -; RV64IM-NEXT: addi a1, zero, 11 +; RV64IM-NEXT: addi a1, zero, 29 ; RV64IM-NEXT: mul a0, a0, a1 -; RV64IM-NEXT: addiw a0, a0, 407 +; RV64IM-NEXT: addi a0, a0, 1073 ; RV64IM-NEXT: ret %tmp0 = add i32 %x, 37 - %tmp1 = mul i32 %tmp0, 11 + %tmp1 = mul i32 %tmp0, 29 ret i32 %tmp1 } -define signext i32 @add_mul_trans_accept_2(i32 %x) { -; RV32IM-LABEL: add_mul_trans_accept_2: +define signext i32 @add_mul_trans_reject_a2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_reject_a2: ; RV32IM: # %bb.0: -; RV32IM-NEXT: addi a1, zero, 13 +; RV32IM-NEXT: addi a1, zero, 29 ; RV32IM-NEXT: mul a0, a0, a1 -; RV32IM-NEXT: lui a1, 28 -; RV32IM-NEXT: addi a1, a1, 1701 +; RV32IM-NEXT: addi a0, a0, 1073 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_reject_a2: +; RV64IM: # %bb.0: +; RV64IM-NEXT: addi a1, zero, 29 +; RV64IM-NEXT: mul a0, a0, a1 +; RV64IM-NEXT: addiw a0, a0, 1073 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 37 + %tmp1 = mul i32 %tmp0, 29 + ret i32 %tmp1 +} + +define i64 @add_mul_trans_reject_a3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_reject_a3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 29 +; RV32IM-NEXT: mul a1, a1, a2 +; RV32IM-NEXT: mulhu a3, a0, a2 +; RV32IM-NEXT: add a1, a3, a1 +; RV32IM-NEXT: mul a2, a0, a2 +; RV32IM-NEXT: addi a0, a2, 1073 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_reject_a3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: addi a1, zero, 29 +; RV64IM-NEXT: mul a0, a0, a1 +; RV64IM-NEXT: addi a0, a0, 1073 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 37 + %tmp1 = mul i64 %tmp0, 29 + ret i64 %tmp1 +} + +define i32 @add_mul_trans_reject_reject_b1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_reject_reject_b1: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a1, zero, 23 +; RV32IM-NEXT: mul a0, a0, a1 +; RV32IM-NEXT: lui a1, 50 +; RV32IM-NEXT: addi a1, a1, 1119 ; RV32IM-NEXT: add a0, a0, a1 ; RV32IM-NEXT: ret ; -; RV64IM-LABEL: add_mul_trans_accept_2: +; RV64IM-LABEL: add_mul_trans_reject_reject_b1: ; RV64IM: # %bb.0: -; RV64IM-NEXT: addi a1, zero, 13 +; RV64IM-NEXT: addi a1, zero, 23 ; RV64IM-NEXT: mul a0, a0, a1 -; RV64IM-NEXT: lui a1, 28 -; RV64IM-NEXT: addiw a1, a1, 1701 -; RV64IM-NEXT: addw a0, a0, a1 +; RV64IM-NEXT: lui a1, 50 +; RV64IM-NEXT: addiw a1, a1, 1119 +; RV64IM-NEXT: add a0, a0, a1 ; RV64IM-NEXT: ret %tmp0 = add i32 %x, 8953 - %tmp1 = mul i32 %tmp0, 13 + %tmp1 = mul i32 %tmp0, 23 ret i32 %tmp1 } -define signext i32 @add_mul_trans_reject_1(i32 %x) { -; RV32IM-LABEL: add_mul_trans_reject_1: +define signext i32 @add_mul_trans_reject_reject_b2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_reject_reject_b2: ; RV32IM: # %bb.0: -; RV32IM-NEXT: addi a1, zero, 19 +; RV32IM-NEXT: addi a1, zero, 23 ; RV32IM-NEXT: mul a0, a0, a1 -; RV32IM-NEXT: lui a1, 9 -; RV32IM-NEXT: addi a1, a1, 585 +; RV32IM-NEXT: lui a1, 50 +; RV32IM-NEXT: addi a1, a1, 1119 ; RV32IM-NEXT: add a0, a0, a1 ; RV32IM-NEXT: ret ; -; RV64IM-LABEL: add_mul_trans_reject_1: +; RV64IM-LABEL: add_mul_trans_reject_reject_b2: ; RV64IM: # %bb.0: -; RV64IM-NEXT: addi a1, zero, 19 +; RV64IM-NEXT: addi a1, zero, 23 ; RV64IM-NEXT: mul a0, a0, a1 -; RV64IM-NEXT: lui a1, 9 -; RV64IM-NEXT: addiw a1, a1, 585 +; RV64IM-NEXT: lui a1, 50 +; RV64IM-NEXT: addiw a1, a1, 1119 ; RV64IM-NEXT: addw a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 8953 + %tmp1 = mul i32 %tmp0, 23 + ret i32 %tmp1 +} + +define i64 @add_mul_trans_reject_reject_b3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_reject_reject_b3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 23 +; RV32IM-NEXT: mul a1, a1, a2 +; RV32IM-NEXT: mulhu a3, a0, a2 +; RV32IM-NEXT: add a1, a3, a1 +; RV32IM-NEXT: mul a2, a0, a2 +; RV32IM-NEXT: lui a0, 50 +; RV32IM-NEXT: addi a0, a0, 1119 +; RV32IM-NEXT: add a0, a2, a0 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_reject_reject_b3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: addi a1, zero, 23 +; RV64IM-NEXT: mul a0, a0, a1 +; RV64IM-NEXT: lui a1, 50 +; RV64IM-NEXT: addiw a1, a1, 1119 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 8953 + %tmp1 = mul i64 %tmp0, 23 + ret i64 %tmp1 +} + +define i32 @add_mul_trans_accept_a1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_accept_a1: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a1, zero, 29 +; RV32IM-NEXT: mul a0, a0, a1 +; RV32IM-NEXT: lui a1, 14 +; RV32IM-NEXT: addi a1, a1, -185 +; RV32IM-NEXT: add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_a1: +; RV64IM: # %bb.0: +; RV64IM-NEXT: addi a1, zero, 29 +; RV64IM-NEXT: mul a0, a0, a1 +; RV64IM-NEXT: lui a1, 14 +; RV64IM-NEXT: addiw a1, a1, -185 +; RV64IM-NEXT: add a0, a0, a1 ; RV64IM-NEXT: ret %tmp0 = add i32 %x, 1971 - %tmp1 = mul i32 %tmp0, 19 + %tmp1 = mul i32 %tmp0, 29 ret i32 %tmp1 } -define signext i32 @add_mul_trans_reject_2(i32 %x) { -; RV32IM-LABEL: add_mul_trans_reject_2: +define signext i32 @add_mul_trans_accept_a2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_accept_a2: ; RV32IM: # %bb.0: -; RV32IM-NEXT: lui a1, 792 -; RV32IM-NEXT: addi a1, a1, -1709 +; RV32IM-NEXT: addi a1, zero, 29 ; RV32IM-NEXT: mul a0, a0, a1 -; RV32IM-NEXT: lui a1, 1014660 -; RV32IM-NEXT: addi a1, a1, -1891 +; RV32IM-NEXT: lui a1, 14 +; RV32IM-NEXT: addi a1, a1, -185 ; RV32IM-NEXT: add a0, a0, a1 ; RV32IM-NEXT: ret ; -; RV64IM-LABEL: add_mul_trans_reject_2: +; RV64IM-LABEL: add_mul_trans_accept_a2: ; RV64IM: # %bb.0: -; RV64IM-NEXT: lui a1, 792 -; RV64IM-NEXT: addiw a1, a1, -1709 +; RV64IM-NEXT: addi a1, zero, 29 ; RV64IM-NEXT: mul a0, a0, a1 -; RV64IM-NEXT: lui a1, 1014660 -; RV64IM-NEXT: addiw a1, a1, -1891 +; RV64IM-NEXT: lui a1, 14 +; RV64IM-NEXT: addiw a1, a1, -185 ; RV64IM-NEXT: addw a0, a0, a1 ; RV64IM-NEXT: ret - %tmp0 = add i32 %x, 1841231 - %tmp1 = mul i32 %tmp0, 3242323 + %tmp0 = add i32 %x, 1971 + %tmp1 = mul i32 %tmp0, 29 ret i32 %tmp1 } + +define i64 @add_mul_trans_accept_a3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_accept_a3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 29 +; RV32IM-NEXT: mul a1, a1, a2 +; RV32IM-NEXT: mulhu a3, a0, a2 +; RV32IM-NEXT: add a1, a3, a1 +; RV32IM-NEXT: mul a2, a0, a2 +; RV32IM-NEXT: lui a0, 14 +; RV32IM-NEXT: addi a0, a0, -185 +; RV32IM-NEXT: add a0, a2, a0 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_a3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: addi a1, zero, 29 +; RV64IM-NEXT: mul a0, a0, a1 +; RV64IM-NEXT: lui a1, 14 +; RV64IM-NEXT: addiw a1, a1, -185 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 1971 + %tmp1 = mul i64 %tmp0, 29 + ret i64 %tmp1 +} + +define i32 @add_mul_trans_accept_b1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_accept_b1: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh3add a0, a0, a0 +; RV32IM-NEXT: lui a1, 18 +; RV32IM-NEXT: addi a1, a1, -1728 +; RV32IM-NEXT: sh3add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_b1: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a0, a0, a0 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -1728 +; RV64IM-NEXT: sh3add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 72 + ret i32 %tmp1 +} + +define signext i32 @add_mul_trans_accept_b2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_accept_b2: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh3add a0, a0, a0 +; RV32IM-NEXT: lui a1, 18 +; RV32IM-NEXT: addi a1, a1, -1728 +; RV32IM-NEXT: sh3add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_b2: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a0, a0, a0 +; RV64IM-NEXT: slli a0, a0, 3 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -1728 +; RV64IM-NEXT: addw a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 72 + ret i32 %tmp1 +} + +define i64 @add_mul_trans_accept_b3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_accept_b3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 72 +; RV32IM-NEXT: mulhu a2, a0, a2 +; RV32IM-NEXT: sh3add a1, a1, a1 +; RV32IM-NEXT: sh3add a1, a1, a2 +; RV32IM-NEXT: sh3add a0, a0, a0 +; RV32IM-NEXT: slli a2, a0, 3 +; RV32IM-NEXT: lui a0, 18 +; RV32IM-NEXT: addi a0, a0, -1728 +; RV32IM-NEXT: add a0, a2, a0 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_b3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a0, a0, a0 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -1728 +; RV64IM-NEXT: sh3add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 1000 + %tmp1 = mul i64 %tmp0, 72 + ret i64 %tmp1 +} + +define i32 @add_mul_trans_accept_c1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_accept_c1: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh3add a1, a0, a0 +; RV32IM-NEXT: sh3add a0, a1, a0 +; RV32IM-NEXT: lui a1, 18 +; RV32IM-NEXT: addi a1, a1, -728 +; RV32IM-NEXT: add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_c1: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a1, a0, a0 +; RV64IM-NEXT: sh3add a0, a1, a0 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -728 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 73 + ret i32 %tmp1 +} + +define signext i32 @add_mul_trans_accept_c2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_accept_c2: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh3add a1, a0, a0 +; RV32IM-NEXT: sh3add a0, a1, a0 +; RV32IM-NEXT: lui a1, 18 +; RV32IM-NEXT: addi a1, a1, -728 +; RV32IM-NEXT: add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_c2: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a1, a0, a0 +; RV64IM-NEXT: sh3add a0, a1, a0 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -728 +; RV64IM-NEXT: addw a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 73 + ret i32 %tmp1 +} + +define i64 @add_mul_trans_accept_c3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_accept_c3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 73 +; RV32IM-NEXT: mul a1, a1, a2 +; RV32IM-NEXT: mulhu a3, a0, a2 +; RV32IM-NEXT: add a1, a3, a1 +; RV32IM-NEXT: mul a2, a0, a2 +; RV32IM-NEXT: lui a0, 18 +; RV32IM-NEXT: addi a0, a0, -728 +; RV32IM-NEXT: add a0, a2, a0 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_c3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh3add a1, a0, a0 +; RV64IM-NEXT: sh3add a0, a1, a0 +; RV64IM-NEXT: lui a1, 18 +; RV64IM-NEXT: addiw a1, a1, -728 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 1000 + %tmp1 = mul i64 %tmp0, 73 + ret i64 %tmp1 +} + +define i32 @add_mul_trans_accept_d1(i32 %x) { +; RV32IM-LABEL: add_mul_trans_accept_d1: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh1add a0, a0, a0 +; RV32IM-NEXT: slli a0, a0, 6 +; RV32IM-NEXT: lui a1, 47 +; RV32IM-NEXT: addi a1, a1, -512 +; RV32IM-NEXT: add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_d1: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh1add a0, a0, a0 +; RV64IM-NEXT: slli a0, a0, 6 +; RV64IM-NEXT: lui a1, 47 +; RV64IM-NEXT: addiw a1, a1, -512 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 192 + ret i32 %tmp1 +} + +define signext i32 @add_mul_trans_accept_d2(i32 signext %x) { +; RV32IM-LABEL: add_mul_trans_accept_d2: +; RV32IM: # %bb.0: +; RV32IM-NEXT: sh1add a0, a0, a0 +; RV32IM-NEXT: slli a0, a0, 6 +; RV32IM-NEXT: lui a1, 47 +; RV32IM-NEXT: addi a1, a1, -512 +; RV32IM-NEXT: add a0, a0, a1 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_d2: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh1add a0, a0, a0 +; RV64IM-NEXT: slli a0, a0, 6 +; RV64IM-NEXT: lui a1, 47 +; RV64IM-NEXT: addiw a1, a1, -512 +; RV64IM-NEXT: addw a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i32 %x, 1000 + %tmp1 = mul i32 %tmp0, 192 + ret i32 %tmp1 +} + +define i64 @add_mul_trans_accept_d3(i64 %x) { +; RV32IM-LABEL: add_mul_trans_accept_d3: +; RV32IM: # %bb.0: +; RV32IM-NEXT: addi a2, zero, 192 +; RV32IM-NEXT: mulhu a2, a0, a2 +; RV32IM-NEXT: sh1add a1, a1, a1 +; RV32IM-NEXT: slli a1, a1, 6 +; RV32IM-NEXT: add a1, a2, a1 +; RV32IM-NEXT: sh1add a0, a0, a0 +; RV32IM-NEXT: slli a2, a0, 6 +; RV32IM-NEXT: lui a0, 47 +; RV32IM-NEXT: addi a0, a0, -512 +; RV32IM-NEXT: add a0, a2, a0 +; RV32IM-NEXT: sltu a2, a0, a2 +; RV32IM-NEXT: add a1, a1, a2 +; RV32IM-NEXT: ret +; +; RV64IM-LABEL: add_mul_trans_accept_d3: +; RV64IM: # %bb.0: +; RV64IM-NEXT: sh1add a0, a0, a0 +; RV64IM-NEXT: slli a0, a0, 6 +; RV64IM-NEXT: lui a1, 47 +; RV64IM-NEXT: addiw a1, a1, -512 +; RV64IM-NEXT: add a0, a0, a1 +; RV64IM-NEXT: ret + %tmp0 = add i64 %x, 1000 + %tmp1 = mul i64 %tmp0, 192 + ret i64 %tmp1 +}