From patchwork Thu Dec 16 15:43:45 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Richard Sandiford X-Patchwork-Id: 49015 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id B340F3858413 for ; Thu, 16 Dec 2021 15:44:17 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B340F3858413 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1639669457; bh=Ar+vICoQfdtLtiu2WoWmwz5UQfvb1Ek2L+uEaRyA2Q0=; h=To:Subject:Date:List-Id:List-Unsubscribe:List-Archive:List-Post: List-Help:List-Subscribe:From:Reply-To:Cc:From; b=f6RlGd2csqn1QyagasJBYxC5Mo0OYPFIj0+quA7BAhbKL+UYhynO/SCu4I6HVm4mB brYWHjuGs7cQ7V6J+nS+Wv6sij0gZfPbNTbtBPBK8pYdZJwD5a4pC+6rlUxXVIw6kb Gv8kf3/McEGPj4Q3mhoCRDT5KhTr/qXOzgtURWB8= X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by sourceware.org (Postfix) with ESMTP id 833F83858C27 for ; Thu, 16 Dec 2021 15:43:47 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 833F83858C27 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id EED981435; Thu, 16 Dec 2021 07:43:46 -0800 (PST) Received: from localhost (e121540-lin.manchester.arm.com [10.32.98.88]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 78DA13F73B; Thu, 16 Dec 2021 07:43:46 -0800 (PST) To: gcc-patches@gcc.gnu.org Mail-Followup-To: gcc-patches@gcc.gnu.org, rguenther@suse.de, richard.sandiford@arm.com Subject: [PATCH] vect: Fix multi-vector SLP gather loads [PR103744] Date: Thu, 16 Dec 2021 15:43:45 +0000 Message-ID: User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.3 (gnu/linux) MIME-Version: 1.0 X-Spam-Status: No, score=-12.4 required=5.0 tests=BAYES_00, GIT_PATCH_0, KAM_DMARC_STATUS, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Richard Sandiford via Gcc-patches From: Richard Sandiford Reply-To: Richard Sandiford Cc: rguenther@suse.de Errors-To: gcc-patches-bounces+patchwork=sourceware.org@gcc.gnu.org Sender: "Gcc-patches" [Repost with address typo fixed] This PR shows that I didn't properly test the multi-vector case when adding support for SLP gather loads. The patch fixes that case using the same approach as we do for non-SLP cases: keep the scalar base the same, but iterate through the (also multi-vector) vector offsets. “vec_num * j + i” is already used elsewhere as a way of handling both the multi-vector SLP case and the multi-vector non-SLP case. Tested on aarch64-linux-gnu. It probably falls just short of being obvious, so: OK to install? Thanks, Richard gcc/ PR tree-optimization/103744 * tree-vect-stmts.c (vectorizable_load): Handle multi-vector SLP gather loads. gcc/testsuite/ PR tree-optimization/103744 * gcc.dg/vect/pr103744-1.c: New test. * gcc.dg/vect/pr103744-2.c: Likewise. --- gcc/testsuite/gcc.dg/vect/pr103744-1.c | 20 +++++++++++++++++ gcc/testsuite/gcc.dg/vect/pr103744-2.c | 31 ++++++++++++++++++++++++++ gcc/tree-vect-stmts.c | 6 +++-- 3 files changed, 55 insertions(+), 2 deletions(-) create mode 100644 gcc/testsuite/gcc.dg/vect/pr103744-1.c create mode 100644 gcc/testsuite/gcc.dg/vect/pr103744-2.c diff --git a/gcc/testsuite/gcc.dg/vect/pr103744-1.c b/gcc/testsuite/gcc.dg/vect/pr103744-1.c new file mode 100644 index 00000000000..1bc81e26fe4 --- /dev/null +++ b/gcc/testsuite/gcc.dg/vect/pr103744-1.c @@ -0,0 +1,20 @@ +/* { dg-do compile } */ + +int r; + +void +foo (short int *s, short int *d1, short int *d2, int z) +{ + int *a; + + while (z < 1) + { + int i; + + i = *s++ - (*d1++ + *d2++); + r += a[i]; + i = *s++ - (*d1++ + *d2++); + r += a[i]; + ++z; + } +} diff --git a/gcc/testsuite/gcc.dg/vect/pr103744-2.c b/gcc/testsuite/gcc.dg/vect/pr103744-2.c new file mode 100644 index 00000000000..52307abf246 --- /dev/null +++ b/gcc/testsuite/gcc.dg/vect/pr103744-2.c @@ -0,0 +1,31 @@ +/* { dg-do compile } */ + +int +f1 (int *restrict x, unsigned short *restrict y) +{ + int res = 0; + for (int i = 0; i < 100; i += 2) + { + unsigned short i1 = y[i + 0] + 1; + unsigned short i2 = y[i + 1] + 2; + res += x[i1]; + res += x[i2]; + } + return res; +} + +void +f2 (int *restrict x, unsigned short *restrict y) +{ + int res1 = 0; + int res2 = 0; + for (int i = 0; i < 100; i += 2) + { + unsigned short i1 = y[i + 0] + 1; + unsigned short i2 = y[i + 1] + 2; + res1 += x[i1]; + res2 += x[i2]; + } + x[0] = res1; + x[1] = res2; +} diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c index 8c427174b37..720cf36caaf 100644 --- a/gcc/tree-vect-stmts.c +++ b/gcc/tree-vect-stmts.c @@ -9247,6 +9247,8 @@ vectorizable_load (vec_info *vinfo, group_size = vec_num = 1; group_gap_adj = 0; ref_type = reference_alias_ptr_type (DR_REF (first_dr_info->dr)); + if (slp) + vec_num = SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node); } gcc_assert (alignment_support_scheme); @@ -9592,7 +9594,7 @@ vectorizable_load (vec_info *vinfo, final_mask = prepare_vec_mask (loop_vinfo, mask_vectype, final_mask, vec_mask, gsi); - if (i > 0) + if (i > 0 && !STMT_VINFO_GATHER_SCATTER_P (stmt_info)) dataref_ptr = bump_vector_ptr (vinfo, dataref_ptr, ptr_incr, gsi, stmt_info, bump); @@ -9609,7 +9611,7 @@ vectorizable_load (vec_info *vinfo, && gs_info.ifn != IFN_LAST) { if (STMT_VINFO_GATHER_SCATTER_P (stmt_info)) - vec_offset = vec_offsets[j]; + vec_offset = vec_offsets[vec_num * j + i]; tree zero = build_zero_cst (vectype); tree scale = size_int (gs_info.scale); gcall *call;