From: Richard Sandiford <richard.sandiford@arm.com>
To: Richard Biener <rguenther@suse.de>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] gimple-isel: Fall back to using vcond_mask [PR98560]
Date: Thu, 07 Jan 2021 11:28:52 +0000 [thread overview]
Message-ID: <mpt35zdf0jv.fsf@arm.com> (raw)
In-Reply-To: <nycvar.YFH.7.76.2101070853180.17979@zhemvz.fhfr.qr> (Richard Biener's message of "Thu, 7 Jan 2021 08:59:25 +0100 (CET)")
Richard Biener <rguenther@suse.de> writes:
> On Wed, 6 Jan 2021, Richard Sandiford wrote:
>
>> PR98560 is about a case in which the vectoriser initially generates:
>>
>> mask_1 = a < 0;
>> mask_2 = mask_1 & ...;
>> res = VEC_COND_EXPR <mask_2, b, c>;
>>
>> The vectoriser thus expects res to be calculated using vcond_mask.
>> However, we later manage to fold mask_2 to mask_1, leaving:
>>
>> mask_1 = a < 0;
>> res = VEC_COND_EXPR <mask_1, b, c>;
>>
>> gimple-isel then required a combined vcond to exist.
>>
>> On most targets, it's not too onerous to provide all possible
>> (compare x select) combinations. For each data mode, you just
>> need to provide unsigned comparisons, signed comparisons, and
>> floating-point comparisons, with the data mode and type of
>> comparison uniquely determining the mode of the compared values.
>> But for targets like SVE that support “unpacked” vectors,
>> it's not that simple: the level of unpacking adds another
>> degree of freedom.
>>
>> Rather than insist that the combined versions exist, I think
>> we should be prepared to fall back to using separate comparisons
>> and vcond_masks. I think that makes more sense on targets like
>> AArch64 and AArch32 in which compares and selects are fundementally
>> separate operations anyway.
>
> Indeed the mask variants (thus being able to expand the comparison)
> are more fundamental. I guess you're running into this path because
> we did not consider using vcond_mask because of
>
> if (used_vec_cond_exprs >= 2
> && (get_vcond_mask_icode (mode, TYPE_MODE (op0_type))
> != CODE_FOR_nothing)
> && expand_vec_cmp_expr_p (op0a_type, op0_type, tcode))
> {
> /* Keep the SSA name and use vcond_mask. */
> tcode = TREE_CODE (op0);
> }
>
> not triggering? Which also means your patch fails to check/assert
> that we can expand_vec_cmp_expr_p the separate compare?
>
>> Tested on aarch64-linux-gnu and x86_64-linux-gnu. OK to install?
>
> It does feel like the function could need some refactoring ...
>
> But OK - preferably with the assertion that we can actually
> expand the compare (I suggest to do the expand_vec_cmp_expr_p
> above unconditionally and have a 'global' cannot_expand_mask
> flag defaulted to false and checked in the new path).
OK, how does this look? It's not quite what you said because
I wanted to avoid the double negation in !cannot.
Thanks,
Richard
gcc/
PR tree-optimization/98560
* gimple-isel.cc (gimple_expand_vec_cond_expr): If we fail to use
IFN_VCOND{,U,EQ}, fall back on IFN_VCOND_MASK.
gcc/testsuite/
PR tree-optimization/98560
* gcc.dg/vect/pr98560-1.c: New test.
---
gcc/gimple-isel.cc | 26 +++++++++++++++++++-------
gcc/testsuite/gcc.dg/vect/pr98560-1.c | 17 +++++++++++++++++
2 files changed, 36 insertions(+), 7 deletions(-)
create mode 100644 gcc/testsuite/gcc.dg/vect/pr98560-1.c
diff --git a/gcc/gimple-isel.cc b/gcc/gimple-isel.cc
index d40338ce4a2..0f3d6bba229 100644
--- a/gcc/gimple-isel.cc
+++ b/gcc/gimple-isel.cc
@@ -154,6 +154,7 @@ gimple_expand_vec_cond_expr (gimple_stmt_iterator *gsi,
return gimple_build_assign (lhs, tem3);
}
+ bool can_compute_op0 = true;
gcc_assert (!COMPARISON_CLASS_P (op0));
if (TREE_CODE (op0) == SSA_NAME)
{
@@ -184,13 +185,16 @@ gimple_expand_vec_cond_expr (gimple_stmt_iterator *gsi,
tree op0_type = TREE_TYPE (op0);
tree op0a_type = TREE_TYPE (op0a);
+ if (TREE_CODE_CLASS (tcode) == tcc_comparison)
+ can_compute_op0 = expand_vec_cmp_expr_p (op0a_type, op0_type,
+ tcode);
/* Try to fold x CMP y ? -1 : 0 to x CMP y. */
- if (integer_minus_onep (op1)
+ if (can_compute_op0
+ && integer_minus_onep (op1)
&& integer_zerop (op2)
- && TYPE_MODE (TREE_TYPE (lhs)) == TYPE_MODE (TREE_TYPE (op0))
- && expand_vec_cmp_expr_p (op0a_type, op0_type, tcode))
+ && TYPE_MODE (TREE_TYPE (lhs)) == TYPE_MODE (TREE_TYPE (op0)))
{
tree conv_op = build1 (VIEW_CONVERT_EXPR, TREE_TYPE (lhs), op0);
gassign *new_stmt = gimple_build_assign (lhs, conv_op);
@@ -198,10 +202,10 @@ gimple_expand_vec_cond_expr (gimple_stmt_iterator *gsi,
return new_stmt;
}
- if (used_vec_cond_exprs >= 2
+ if (can_compute_op0
+ && used_vec_cond_exprs >= 2
&& (get_vcond_mask_icode (mode, TYPE_MODE (op0_type))
- != CODE_FOR_nothing)
- && expand_vec_cmp_expr_p (op0a_type, op0_type, tcode))
+ != CODE_FOR_nothing))
{
/* Keep the SSA name and use vcond_mask. */
tcode = TREE_CODE (op0);
@@ -254,7 +258,15 @@ gimple_expand_vec_cond_expr (gimple_stmt_iterator *gsi,
}
}
- gcc_assert (icode != CODE_FOR_nothing);
+ if (icode == CODE_FOR_nothing)
+ {
+ gcc_assert (VECTOR_BOOLEAN_TYPE_P (TREE_TYPE (op0))
+ && can_compute_op0
+ && (get_vcond_mask_icode (mode, TYPE_MODE (TREE_TYPE (op0)))
+ != CODE_FOR_nothing));
+ return gimple_build_call_internal (IFN_VCOND_MASK, 3, op0, op1, op2);
+ }
+
tree tcode_tree = build_int_cst (integer_type_node, tcode);
return gimple_build_call_internal (unsignedp ? IFN_VCONDU : IFN_VCOND,
5, op0a, op0b, op1, op2, tcode_tree);
diff --git a/gcc/testsuite/gcc.dg/vect/pr98560-1.c b/gcc/testsuite/gcc.dg/vect/pr98560-1.c
new file mode 100644
index 00000000000..2583fc48f8a
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr98560-1.c
@@ -0,0 +1,17 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-O3 -fno-tree-vrp -fno-tree-fre -fno-tree-pre -fno-code-hoisting -fvect-cost-model=dynamic" } */
+/* { dg-additional-options "-msve-vector-bits=128" { target aarch64_sve } } */
+
+#include <stdint.h>
+
+void
+f (uint16_t *restrict dst, uint32_t *restrict src1, float *restrict src2)
+{
+ int i = 0;
+ for (int j = 0; j < 4; ++j)
+ {
+ uint16_t tmp = src1[i] >> 1;
+ dst[i] = (uint16_t) (src2[i] < 0 && i < 4 ? tmp : 1);
+ i += 1;
+ }
+}
next prev parent reply other threads:[~2021-01-07 11:28 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-06 15:50 Richard Sandiford
2021-01-07 7:59 ` Richard Biener
2021-01-07 11:28 ` Richard Sandiford [this message]
2021-01-07 13:13 ` Richard Biener
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=mpt35zdf0jv.fsf@arm.com \
--to=richard.sandiford@arm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=rguenther@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).