llvm-6502/test/CodeGen/Mips/msa/3r_4r_widen.ll
Daniel Sanders ca795b61be [mips][msa] Build all the tests in little and big endian modes and correct an incorrect test.
Summary:
This patch (correctly) breaks some MSA tests by exposing the cases when
SelectionDAG::getConstant() produces illegal types. These have been temporarily
marked XFAIL and the XFAIL flag will be removed when
SelectionDAG::getConstant() is fixed.

There are three categories of failure:
* Immediate instructions are not selected in one endian mode.
* Immediates used in ldi.[bhwd] must be different according to endianness.
  (this only affects cases where the 'wrong' ldi is used to load the correct
   bitpattern. E.g. (bitcast:v2i64 (build_vector:v4i32 ...)))
* Non-immediate instructions that rely on immediates affected by the
  previous two categories as part of their match pattern.
  For example, the bset match pattern is the vector equivalent of
  'ws | (1 << wt)'.

One test needed correcting to expect different output depending on whether big
or little endian was in use. This test was
test/CodeGen/Mips/msa/basic_operations.ll and experiences the second category
of failure shown above. The little endian version of this test is named
basic_operations_little.ll and will be merged back into basic_operations.ll in
a follow up commit now that FileCheck supports multiple check prefixes.

Reviewers: bkramer, jacksprat, dsanders

Reviewed By: dsanders

CC: llvm-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D1972

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@194806 91177308-0d34-0410-b5e6-96231b3b80d8
2013-11-15 11:04:16 +00:00

308 lines
12 KiB
LLVM

; Test the MSA intrinsics that are encoded with the 3R instruction format and
; use the result as a third operand and results in wider elements than the
; operands had.
; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck %s
; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck %s
@llvm_mips_dpadd_s_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_dpadd_s_h_ARG2 = global <16 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 20, i8 21, i8 22, i8 23>, align 16
@llvm_mips_dpadd_s_h_ARG3 = global <16 x i8> <i8 24, i8 25, i8 26, i8 27, i8 28, i8 29, i8 30, i8 31, i8 32, i8 33, i8 34, i8 35, i8 36, i8 37, i8 38, i8 39>, align 16
@llvm_mips_dpadd_s_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_dpadd_s_h_test() nounwind {
entry:
%0 = load <8 x i16>* @llvm_mips_dpadd_s_h_ARG1
%1 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG2
%2 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_s_h_RES
ret void
}
declare <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
; CHECK: llvm_mips_dpadd_s_h_test:
; CHECK: ld.b
; CHECK: ld.b
; CHECK: ld.h
; CHECK: dpadd_s.h
; CHECK: st.h
; CHECK: .size llvm_mips_dpadd_s_h_test
;
@llvm_mips_dpadd_s_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_dpadd_s_w_ARG2 = global <8 x i16> <i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11>, align 16
@llvm_mips_dpadd_s_w_ARG3 = global <8 x i16> <i16 12, i16 13, i16 14, i16 15, i16 16, i16 17, i16 18, i16 19>, align 16
@llvm_mips_dpadd_s_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_dpadd_s_w_test() nounwind {
entry:
%0 = load <4 x i32>* @llvm_mips_dpadd_s_w_ARG1
%1 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG2
%2 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_s_w_RES
ret void
}
declare <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_dpadd_s_w_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.w
; CHECK: dpadd_s.w
; CHECK: st.w
; CHECK: .size llvm_mips_dpadd_s_w_test
;
@llvm_mips_dpadd_s_d_ARG1 = global <2 x i64> <i64 0, i64 1>, align 16
@llvm_mips_dpadd_s_d_ARG2 = global <4 x i32> <i32 2, i32 3, i32 4, i32 5>, align 16
@llvm_mips_dpadd_s_d_ARG3 = global <4 x i32> <i32 6, i32 7, i32 8, i32 9>, align 16
@llvm_mips_dpadd_s_d_RES = global <2 x i64> <i64 0, i64 0>, align 16
define void @llvm_mips_dpadd_s_d_test() nounwind {
entry:
%0 = load <2 x i64>* @llvm_mips_dpadd_s_d_ARG1
%1 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG2
%2 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG3
%3 = tail call <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_s_d_RES
ret void
}
declare <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_dpadd_s_d_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.d
; CHECK: dpadd_s.d
; CHECK: st.d
; CHECK: .size llvm_mips_dpadd_s_d_test
;
@llvm_mips_dpadd_u_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_dpadd_u_h_ARG2 = global <16 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 20, i8 21, i8 22, i8 23>, align 16
@llvm_mips_dpadd_u_h_ARG3 = global <16 x i8> <i8 24, i8 25, i8 26, i8 27, i8 28, i8 29, i8 30, i8 31, i8 32, i8 33, i8 34, i8 35, i8 36, i8 37, i8 38, i8 39>, align 16
@llvm_mips_dpadd_u_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_dpadd_u_h_test() nounwind {
entry:
%0 = load <8 x i16>* @llvm_mips_dpadd_u_h_ARG1
%1 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG2
%2 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_u_h_RES
ret void
}
declare <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
; CHECK: llvm_mips_dpadd_u_h_test:
; CHECK: ld.b
; CHECK: ld.b
; CHECK: ld.h
; CHECK: dpadd_u.h
; CHECK: st.h
; CHECK: .size llvm_mips_dpadd_u_h_test
;
@llvm_mips_dpadd_u_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_dpadd_u_w_ARG2 = global <8 x i16> <i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11>, align 16
@llvm_mips_dpadd_u_w_ARG3 = global <8 x i16> <i16 12, i16 13, i16 14, i16 15, i16 16, i16 17, i16 18, i16 19>, align 16
@llvm_mips_dpadd_u_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_dpadd_u_w_test() nounwind {
entry:
%0 = load <4 x i32>* @llvm_mips_dpadd_u_w_ARG1
%1 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG2
%2 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_u_w_RES
ret void
}
declare <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_dpadd_u_w_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.w
; CHECK: dpadd_u.w
; CHECK: st.w
; CHECK: .size llvm_mips_dpadd_u_w_test
;
@llvm_mips_dpadd_u_d_ARG1 = global <2 x i64> <i64 0, i64 1>, align 16
@llvm_mips_dpadd_u_d_ARG2 = global <4 x i32> <i32 2, i32 3, i32 4, i32 5>, align 16
@llvm_mips_dpadd_u_d_ARG3 = global <4 x i32> <i32 6, i32 7, i32 8, i32 9>, align 16
@llvm_mips_dpadd_u_d_RES = global <2 x i64> <i64 0, i64 0>, align 16
define void @llvm_mips_dpadd_u_d_test() nounwind {
entry:
%0 = load <2 x i64>* @llvm_mips_dpadd_u_d_ARG1
%1 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG2
%2 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG3
%3 = tail call <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_u_d_RES
ret void
}
declare <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_dpadd_u_d_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.d
; CHECK: dpadd_u.d
; CHECK: st.d
; CHECK: .size llvm_mips_dpadd_u_d_test
;
@llvm_mips_dpsub_s_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_dpsub_s_h_ARG2 = global <16 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 20, i8 21, i8 22, i8 23>, align 16
@llvm_mips_dpsub_s_h_ARG3 = global <16 x i8> <i8 24, i8 25, i8 26, i8 27, i8 28, i8 29, i8 30, i8 31, i8 32, i8 33, i8 34, i8 35, i8 36, i8 37, i8 38, i8 39>, align 16
@llvm_mips_dpsub_s_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_dpsub_s_h_test() nounwind {
entry:
%0 = load <8 x i16>* @llvm_mips_dpsub_s_h_ARG1
%1 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG2
%2 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_s_h_RES
ret void
}
declare <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
; CHECK: llvm_mips_dpsub_s_h_test:
; CHECK: ld.b
; CHECK: ld.b
; CHECK: ld.h
; CHECK: dpsub_s.h
; CHECK: st.h
; CHECK: .size llvm_mips_dpsub_s_h_test
;
@llvm_mips_dpsub_s_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_dpsub_s_w_ARG2 = global <8 x i16> <i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11>, align 16
@llvm_mips_dpsub_s_w_ARG3 = global <8 x i16> <i16 12, i16 13, i16 14, i16 15, i16 16, i16 17, i16 18, i16 19>, align 16
@llvm_mips_dpsub_s_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_dpsub_s_w_test() nounwind {
entry:
%0 = load <4 x i32>* @llvm_mips_dpsub_s_w_ARG1
%1 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG2
%2 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_s_w_RES
ret void
}
declare <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_dpsub_s_w_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.w
; CHECK: dpsub_s.w
; CHECK: st.w
; CHECK: .size llvm_mips_dpsub_s_w_test
;
@llvm_mips_dpsub_s_d_ARG1 = global <2 x i64> <i64 0, i64 1>, align 16
@llvm_mips_dpsub_s_d_ARG2 = global <4 x i32> <i32 2, i32 3, i32 4, i32 5>, align 16
@llvm_mips_dpsub_s_d_ARG3 = global <4 x i32> <i32 6, i32 7, i32 8, i32 9>, align 16
@llvm_mips_dpsub_s_d_RES = global <2 x i64> <i64 0, i64 0>, align 16
define void @llvm_mips_dpsub_s_d_test() nounwind {
entry:
%0 = load <2 x i64>* @llvm_mips_dpsub_s_d_ARG1
%1 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG2
%2 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG3
%3 = tail call <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_s_d_RES
ret void
}
declare <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_dpsub_s_d_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.d
; CHECK: dpsub_s.d
; CHECK: st.d
; CHECK: .size llvm_mips_dpsub_s_d_test
;
@llvm_mips_dpsub_u_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_dpsub_u_h_ARG2 = global <16 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 20, i8 21, i8 22, i8 23>, align 16
@llvm_mips_dpsub_u_h_ARG3 = global <16 x i8> <i8 24, i8 25, i8 26, i8 27, i8 28, i8 29, i8 30, i8 31, i8 32, i8 33, i8 34, i8 35, i8 36, i8 37, i8 38, i8 39>, align 16
@llvm_mips_dpsub_u_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_dpsub_u_h_test() nounwind {
entry:
%0 = load <8 x i16>* @llvm_mips_dpsub_u_h_ARG1
%1 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG2
%2 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_u_h_RES
ret void
}
declare <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
; CHECK: llvm_mips_dpsub_u_h_test:
; CHECK: ld.b
; CHECK: ld.b
; CHECK: ld.h
; CHECK: dpsub_u.h
; CHECK: st.h
; CHECK: .size llvm_mips_dpsub_u_h_test
;
@llvm_mips_dpsub_u_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_dpsub_u_w_ARG2 = global <8 x i16> <i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11>, align 16
@llvm_mips_dpsub_u_w_ARG3 = global <8 x i16> <i16 12, i16 13, i16 14, i16 15, i16 16, i16 17, i16 18, i16 19>, align 16
@llvm_mips_dpsub_u_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_dpsub_u_w_test() nounwind {
entry:
%0 = load <4 x i32>* @llvm_mips_dpsub_u_w_ARG1
%1 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG2
%2 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_u_w_RES
ret void
}
declare <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_dpsub_u_w_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.w
; CHECK: dpsub_u.w
; CHECK: st.w
; CHECK: .size llvm_mips_dpsub_u_w_test
;
@llvm_mips_dpsub_u_d_ARG1 = global <2 x i64> <i64 0, i64 1>, align 16
@llvm_mips_dpsub_u_d_ARG2 = global <4 x i32> <i32 2, i32 3, i32 4, i32 5>, align 16
@llvm_mips_dpsub_u_d_ARG3 = global <4 x i32> <i32 6, i32 7, i32 8, i32 9>, align 16
@llvm_mips_dpsub_u_d_RES = global <2 x i64> <i64 0, i64 0>, align 16
define void @llvm_mips_dpsub_u_d_test() nounwind {
entry:
%0 = load <2 x i64>* @llvm_mips_dpsub_u_d_ARG1
%1 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG2
%2 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG3
%3 = tail call <2 x i64> @llvm.mips.dpsub.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_u_d_RES
ret void
}
declare <2 x i64> @llvm.mips.dpsub.u.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_dpsub_u_d_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.d
; CHECK: dpsub_u.d
; CHECK: st.d
; CHECK: .size llvm_mips_dpsub_u_d_test
;