From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1652) id 2851B3857800; Mon, 23 Nov 2020 10:42:52 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 2851B3857800 Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: Christophe Lyon To: gcc-cvs@gcc.gnu.org Subject: [gcc(refs/users/clyon/heads/mve-autovec)] arm: Auto-vectorization for MVE: veor X-Act-Checkin: gcc X-Git-Author: Christophe Lyon X-Git-Refname: refs/users/clyon/heads/mve-autovec X-Git-Oldrev: ef044ad53da1b55e3b480e656a8d4a8dcbe143d4 X-Git-Newrev: 8e5c88e0e2b6503c85e8291c2def2498223c408a Message-Id: <20201123104252.2851B3857800@sourceware.org> Date: Mon, 23 Nov 2020 10:42:52 +0000 (GMT) X-BeenThere: gcc-cvs@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-cvs mailing list List-Unsubscribe: , List-Archive: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 23 Nov 2020 10:42:52 -0000 https://gcc.gnu.org/g:8e5c88e0e2b6503c85e8291c2def2498223c408a commit 8e5c88e0e2b6503c85e8291c2def2498223c408a Author: Christophe Lyon Date: Fri Nov 13 13:05:43 2020 +0000 arm: Auto-vectorization for MVE: veor This patch enables MVE veorq instructions for auto-vectorization. MVE veorq insns in mve.md are modified to use xor instead of unspec expression to support xor3. The xor3 expander is added to vec-common.md 2020-11-12 Christophe Lyon gcc/ * config/arm/iterators.md (supf): Remove VEORQ_S and VEORQ_U. (VEORQ): Remove. * config/arm/mve.md (mve_veorq_u): New entry for veor instruction using expression xor. (mve_veorq_s): New expander. * config/arm/neon.md (xor3): Renamed into xor3_neon. * config/arm/unspscs.md (VEORQ_S, VEORQ_U): Remove. * config/arm/vec-common.md (xor3): New expander. gcc/testsuite/ * gcc.target/arm/simd/mve-veor.c: Add tests for veor. Diff: --- gcc/config/arm/iterators.md | 3 +- gcc/config/arm/mve.md | 17 +++-- gcc/config/arm/neon.md | 2 +- gcc/config/arm/unspecs.md | 2 - gcc/config/arm/vec-common.md | 9 ++- gcc/testsuite/gcc.target/arm/simd/mve-veor.c | 96 ++++++++++++++++++++++++++++ 6 files changed, 118 insertions(+), 11 deletions(-) diff --git a/gcc/config/arm/iterators.md b/gcc/config/arm/iterators.md index 5fcb7afe565..01952756e08 100644 --- a/gcc/config/arm/iterators.md +++ b/gcc/config/arm/iterators.md @@ -1237,7 +1237,7 @@ (VCADDQ_ROT270_U "u") (VCADDQ_ROT90_S "s") (VCMPEQQ_S "s") (VCMPEQQ_U "u") (VCADDQ_ROT90_U "u") (VCMPEQQ_N_S "s") (VCMPEQQ_N_U "u") (VCMPNEQ_N_S "s") - (VCMPNEQ_N_U "u") (VEORQ_S "s") (VEORQ_U "u") + (VCMPNEQ_N_U "u") (VHADDQ_N_S "s") (VHADDQ_N_U "u") (VHADDQ_S "s") (VHADDQ_U "u") (VHSUBQ_N_S "s") (VHSUBQ_N_U "u") (VHSUBQ_S "s") (VMAXQ_S "s") (VMAXQ_U "u") (VHSUBQ_U "u") @@ -1507,7 +1507,6 @@ (define_int_iterator VCMPEQQ [VCMPEQQ_U VCMPEQQ_S]) (define_int_iterator VCMPEQQ_N [VCMPEQQ_N_S VCMPEQQ_N_U]) (define_int_iterator VCMPNEQ_N [VCMPNEQ_N_U VCMPNEQ_N_S]) -(define_int_iterator VEORQ [VEORQ_U VEORQ_S]) (define_int_iterator VHADDQ [VHADDQ_S VHADDQ_U]) (define_int_iterator VHADDQ_N [VHADDQ_N_U VHADDQ_N_S]) (define_int_iterator VHSUBQ [VHSUBQ_S VHSUBQ_U]) diff --git a/gcc/config/arm/mve.md b/gcc/config/arm/mve.md index 0f040444861..a5f5d7527f4 100644 --- a/gcc/config/arm/mve.md +++ b/gcc/config/arm/mve.md @@ -1204,17 +1204,24 @@ ;; ;; [veorq_u, veorq_s]) ;; -(define_insn "mve_veorq_" +(define_insn "mve_veorq_u" [ (set (match_operand:MVE_2 0 "s_register_operand" "=w") - (unspec:MVE_2 [(match_operand:MVE_2 1 "s_register_operand" "w") - (match_operand:MVE_2 2 "s_register_operand" "w")] - VEORQ)) + (xor:MVE_2 (match_operand:MVE_2 1 "s_register_operand" "w") + (match_operand:MVE_2 2 "s_register_operand" "w"))) ] "TARGET_HAVE_MVE" - "veor %q0, %q1, %q2" + "veor\t%q0, %q1, %q2" [(set_attr "type" "mve_move") ]) +(define_expand "mve_veorq_s" + [ + (set (match_operand:MVE_2 0 "s_register_operand") + (xor:MVE_2 (match_operand:MVE_2 1 "s_register_operand") + (match_operand:MVE_2 2 "s_register_operand"))) + ] + "TARGET_HAVE_MVE" +) ;; ;; [vhaddq_n_u, vhaddq_n_s]) diff --git a/gcc/config/arm/neon.md b/gcc/config/arm/neon.md index 669c34da4e0..e1263b00b39 100644 --- a/gcc/config/arm/neon.md +++ b/gcc/config/arm/neon.md @@ -747,7 +747,7 @@ [(set_attr "type" "neon_logic")] ) -(define_insn "xor3" +(define_insn "xor3_neon" [(set (match_operand:VDQ 0 "s_register_operand" "=w") (xor:VDQ (match_operand:VDQ 1 "s_register_operand" "w") (match_operand:VDQ 2 "s_register_operand" "w")))] diff --git a/gcc/config/arm/unspecs.md b/gcc/config/arm/unspecs.md index f111ad80ead..78313ea87bc 100644 --- a/gcc/config/arm/unspecs.md +++ b/gcc/config/arm/unspecs.md @@ -608,7 +608,6 @@ VCMPEQQ_S VCMPEQQ_N_S VCMPNEQ_N_S - VEORQ_S VHADDQ_S VHADDQ_N_S VHSUBQ_S @@ -653,7 +652,6 @@ VCMPEQQ_U VCMPEQQ_N_U VCMPNEQ_N_U - VEORQ_U VHADDQ_U VHADDQ_N_U VHSUBQ_U diff --git a/gcc/config/arm/vec-common.md b/gcc/config/arm/vec-common.md index 3e8387c54cb..99ea1cc0b06 100644 --- a/gcc/config/arm/vec-common.md +++ b/gcc/config/arm/vec-common.md @@ -185,4 +185,11 @@ (ior:VDQ (match_operand:VDQ 1 "s_register_operand" "") (match_operand:VDQ 2 "neon_logic_op2" "")))] "TARGET_NEON || TARGET_HAVE_MVE" -) \ No newline at end of file +) + +(define_expand "xor3" + [(set (match_operand:VDQ 0 "s_register_operand" "") + (xor:VDQ (match_operand:VDQ 1 "s_register_operand" "") + (match_operand:VDQ 2 "s_register_operand" "")))] + "TARGET_NEON || TARGET_HAVE_MVE" +) diff --git a/gcc/testsuite/gcc.target/arm/simd/mve-veor.c b/gcc/testsuite/gcc.target/arm/simd/mve-veor.c new file mode 100644 index 00000000000..7584028a645 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/simd/mve-veor.c @@ -0,0 +1,96 @@ +/* { dg-do assemble } */ +/* { dg-require-effective-target arm_v8_1m_mve_ok } */ +/* { dg-add-options arm_v8_1m_mve } */ +/* { dg-additional-options "-O3" } */ + +#include + +/* 64-bit vectors. */ +void test_veor_i32x2 (int32_t * __restrict__ dest, int32_t * a, int32_t * b) { + int i; + for (i=0; i<2; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u32x2 (uint32_t * __restrict__ dest, uint32_t * a, uint32_t * b) { + int i; + for (i=0; i<2; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_i16x4 (int16_t * __restrict__ dest, int16_t * a, int16_t * b) { + int i; + for (i=0; i<4; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u16x4 (uint16_t * __restrict__ dest, uint16_t * a, uint16_t * b) { + int i; + for (i=0; i<4; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_i8x8 (int8_t * __restrict__ dest, int8_t * a, int8_t * b) { + int i; + for (i=0; i<8; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u8x8 (uint8_t * __restrict__ dest, uint8_t * a, uint8_t * b) { + int i; + for (i=0; i<8; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +/* 128-bit vectors. */ +void test_veor_i32x4 (int32_t * __restrict__ dest, int32_t * a, int32_t * b) { + int i; + for (i=0; i<4; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u32x4 (uint32_t * __restrict__ dest, uint32_t * a, uint32_t * b) { + int i; + for (i=0; i<4; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_i16x8 (int16_t * __restrict__ dest, int16_t * a, int16_t * b) { + int i; + for (i=0; i<8; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u16x8 (uint16_t * __restrict__ dest, uint16_t * a, uint16_t * b) { + int i; + for (i=0; i<8; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_i8x16 (int8_t * __restrict__ dest, int8_t * a, int8_t * b) { + int i; + for (i=0; i<16; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +void test_veor_u8x16 (uint8_t * __restrict__ dest, uint8_t * a, uint8_t * b) { + int i; + for (i=0; i<16; i++) { + dest[i] = a[i] ^ b[i]; + } +} + +/* MVE has only 128-bit vectors, so we can vectorize only half of the + functions above. */ +/* { dg-final { scan-assembler-times {veor\tq[0-9]+, q[0-9]+, q[0-9]+} 6 } } */