On Tue, 7 Nov 2023, Tamar Christina wrote:
> > -----Original Message-----
> > From: Richard Biener <rguent...@suse.de>
> > Sent: Tuesday, November 7, 2023 9:43 AM
> > To: Tamar Christina <tamar.christ...@arm.com>
> > Cc: gcc-patches@gcc.gnu.org; nd <n...@arm.com>
> > Subject: RE: [PATCH v6 0/21]middle-end: Support early break/return auto-
> > vectorization
> >
> > On Mon, 6 Nov 2023, Tamar Christina wrote:
> >
> > > > -----Original Message-----
> > > > From: Richard Biener <rguent...@suse.de>
> > > > Sent: Monday, November 6, 2023 2:25 PM
> > > > To: Tamar Christina <tamar.christ...@arm.com>
> > > > Cc: gcc-patches@gcc.gnu.org; nd <n...@arm.com>
> > > > Subject: Re: [PATCH v6 0/21]middle-end: Support early break/return
> > > > auto- vectorization
> > > >
> > > > On Mon, 6 Nov 2023, Tamar Christina wrote:
> > > >
> > > > > Hi All,
> > > > >
> > > > > This patch adds initial support for early break vectorization in GCC.
> > > > > The support is added for any target that implements a vector
> > > > > cbranch optab, this includes both fully masked and non-masked targets.
> > > > >
> > > > > Depending on the operation, the vectorizer may also require
> > > > > support for boolean mask reductions using Inclusive OR. This is
> > > > > however only checked then the comparison would produce multiple
> > statements.
> > > > >
> > > > > Note: I am currently struggling to get patch 7 correct in all
> > > > > cases and could
> > > > use
> > > > > some feedback there.
> > > > >
> > > > > Concretely the kind of loops supported are of the forms:
> > > > >
> > > > > for (int i = 0; i < N; i++)
> > > > > {
> > > > > <statements1>
> > > > > if (<condition>)
> > > > > {
> > > > > ...
> > > > > <action>;
> > > > > }
> > > > > <statements2>
> > > > > }
> > > > >
> > > > > where <action> can be:
> > > > > - break
> > > > > - return
> > > > > - goto
> > > > >
> > > > > Any number of statements can be used before the <action> occurs.
> > > > >
> > > > > Since this is an initial version for GCC 14 it has the following
> > > > > limitations and
> > > > > features:
> > > > >
> > > > > - Only fixed sized iterations and buffers are supported. That is to
> > > > > say any
> > > > > vectors loaded or stored must be to statically allocated arrays with
> > known
> > > > > sizes. N must also be known. This limitation is because our primary
> > target
> > > > > for this optimization is SVE. For VLA SVE we can't easily do cross
> > > > > page
> > > > > iteraion checks. The result is likely to also not be beneficial.
> > > > > For that
> > > > > reason we punt support for variable buffers till we have
> > > > > First-Faulting
> > > > > support in GCC.
> >
> > Btw, for this I wonder if you thought about marking memory accesses required
> > for the early break condition as required to be vector-size aligned, thus
> > peeling
> > or versioning them for alignment? That should ensure they do not fault.
> >
> > OTOH I somehow remember prologue peeling isn't supported for early break
> > vectorization? ..
> >
> > > > > - any stores in <statements1> should not be to the same objects as in
> > > > > <condition>. Loads are fine as long as they don't have the
> > > > > possibility to
> > > > > alias. More concretely, we block RAW dependencies when the
> > > > > intermediate
> > > > value
> > > > > can't be separated fromt the store, or the store itself can't be
> > > > > moved.
> > > > > - Prologue peeling, alignment peelinig and loop versioning are
> > > > > supported.
> >
> > .. but here you say it is. Not sure if peeling for alignment works for VLA
> > vectors
> > though. Just to say x86 doesn't support first-faulting loads.
>
> For VLA we support it through masking. i.e. if you need to peel N
> iterations, we
> generate a masked copy of the loop vectorized which masks off the first N
> bits.
>
> This is not typically needed, but we do support it. But the problem with this
> scheme and early break is obviously that the peeled loop needs to be
> vectorized
> so you kinda end up with the same issue again. So Atm it rejects it for VLA.
Hmm, I see. I thought peeling by masking is an optimization. Anyhow,
I think it should still work here - since all accesses are aligned and
we know that there's at least one original scalar iteration in the
first masked and the following "unmasked" vector iterations there
should never be faults for any of the aligned accesses.
I think going via alignment is a way easier method to guarantee this
than handwaving about "declared" arrays and niter. One can try that
in addition of course - it's not always possible to align all
vector loads we are going to speculate (for VLA one could also
find common runtime (mis-)alignment and restrict the vector length based
on that, for RISC-V it seems to be efficient, not sure whether altering
that for SVE is though).
Richard.
> Regards,
> Tamar
>
> >
> > > > > - Fully masked loops, unmasked loops and partially masked loops
> > > > > are supported
> > > > > - Any number of loop early exits are supported.
> > > > > - No support for epilogue vectorization. The only epilogue supported
> > > > > is
> > the
> > > > > scalar final one. Peeling code supports it but the code motion code
> > cannot
> > > > > find instructions to make the move in the epilog.
> > > > > - Early breaks are only supported for inner loop vectorization.
> > > > >
> > > > > I have pushed a branch to
> > > > > refs/users/tnfchris/heads/gcc-14-early-break
> > > > >
> > > > > With the help of IPA and LTO this still gets hit quite often.
> > > > > During bootstrap it hit rather frequently. Additionally TSVC
> > > > > s332, s481 and
> > > > > s482 all pass now since these are tests for support for early exit
> > > > vectorization.
> > > > >
> > > > > This implementation does not support completely handling the early
> > > > > break inside the vector loop itself but instead supports adding
> > > > > checks such that if we know that we have to exit in the current
> > > > > iteration then we branch to scalar code to actually do the final
> > > > > VF iterations which
> > > > handles all the code in <action>.
> > > > >
> > > > > For the scalar loop we know that whatever exit you take you have
> > > > > to perform at most VF iterations. For vector code we only case
> > > > > about the state of fully performed iteration and reset the scalar
> > > > > code to the (partially)
> > > > remaining loop.
> > > > >
> > > > > That is to say, the first vector loop executes so long as the
> > > > > early exit isn't needed. Once the exit is taken, the scalar code
> > > > > will perform at most VF extra iterations. The exact number
> > > > > depending on peeling
> > > > and iteration start and which
> > > > > exit was taken (natural or early). For this scalar loop, all early
> > > > > exits are
> > > > > treated the same.
> > > > >
> > > > > When we vectorize we move any statement not related to the early
> > > > > break itself and that would be incorrect to execute before the break
> > > > > (i.e.
> > > > > has side effects) to after the break. If this is not possible we
> > > > > decline to
> > > > vectorize.
> > > > >
> > > > > This means that we check at the start of iterations whether we are
> > > > > going to exit or not. During the analyis phase we check whether
> > > > > we are allowed to do this moving of statements. Also note that we
> > > > > only move the scalar statements, but only do so after peeling but
> > > > > just before we
> > > > start transforming statements.
> > > > >
> > > > > Codegen:
> > > > >
> > > > > for e.g.
> > > > >
> > > > > #define N 803
> > > > > unsigned vect_a[N];
> > > > > unsigned vect_b[N];
> > > > >
> > > > > unsigned test4(unsigned x)
> > > > > {
> > > > > unsigned ret = 0;
> > > > > for (int i = 0; i < N; i++)
> > > > > {
> > > > > vect_b[i] = x + i;
> > > > > if (vect_a[i] > x)
> > > > > break;
> > > > > vect_a[i] = x;
> > > > >
> > > > > }
> > > > > return ret;
> > > > > }
> > > > >
> > > > > We generate for Adv. SIMD:
> > > > >
> > > > > test4:
> > > > > adrp x2, .LC0
> > > > > adrp x3, .LANCHOR0
> > > > > dup v2.4s, w0
> > > > > add x3, x3, :lo12:.LANCHOR0
> > > > > movi v4.4s, 0x4
> > > > > add x4, x3, 3216
> > > > > ldr q1, [x2, #:lo12:.LC0]
> > > > > mov x1, 0
> > > > > mov w2, 0
> > > > > .p2align 3,,7
> > > > > .L3:
> > > > > ldr q0, [x3, x1]
> > > > > add v3.4s, v1.4s, v2.4s
> > > > > add v1.4s, v1.4s, v4.4s
> > > > > cmhi v0.4s, v0.4s, v2.4s
> > > > > umaxp v0.4s, v0.4s, v0.4s
> > > > > fmov x5, d0
> > > > > cbnz x5, .L6
> > > > > add w2, w2, 1
> > > > > str q3, [x1, x4]
> > > > > str q2, [x3, x1]
> > > > > add x1, x1, 16
> > > > > cmp w2, 200
> > > > > bne .L3
> > > > > mov w7, 3
> > > > > .L2:
> > > > > lsl w2, w2, 2
> > > > > add x5, x3, 3216
> > > > > add w6, w2, w0
> > > > > sxtw x4, w2
> > > > > ldr w1, [x3, x4, lsl 2]
> > > > > str w6, [x5, x4, lsl 2]
> > > > > cmp w0, w1
> > > > > bcc .L4
> > > > > add w1, w2, 1
> > > > > str w0, [x3, x4, lsl 2]
> > > > > add w6, w1, w0
> > > > > sxtw x1, w1
> > > > > ldr w4, [x3, x1, lsl 2]
> > > > > str w6, [x5, x1, lsl 2]
> > > > > cmp w0, w4
> > > > > bcc .L4
> > > > > add w4, w2, 2
> > > > > str w0, [x3, x1, lsl 2]
> > > > > sxtw x1, w4
> > > > > add w6, w1, w0
> > > > > ldr w4, [x3, x1, lsl 2]
> > > > > str w6, [x5, x1, lsl 2]
> > > > > cmp w0, w4
> > > > > bcc .L4
> > > > > str w0, [x3, x1, lsl 2]
> > > > > add w2, w2, 3
> > > > > cmp w7, 3
> > > > > beq .L4
> > > > > sxtw x1, w2
> > > > > add w2, w2, w0
> > > > > ldr w4, [x3, x1, lsl 2]
> > > > > str w2, [x5, x1, lsl 2]
> > > > > cmp w0, w4
> > > > > bcc .L4
> > > > > str w0, [x3, x1, lsl 2]
> > > > > .L4:
> > > > > mov w0, 0
> > > > > ret
> > > > > .p2align 2,,3
> > > > > .L6:
> > > > > mov w7, 4
> > > > > b .L2
> > > > >
> > > > > and for SVE:
> > > > >
> > > > > test4:
> > > > > adrp x2, .LANCHOR0
> > > > > add x2, x2, :lo12:.LANCHOR0
> > > > > add x5, x2, 3216
> > > > > mov x3, 0
> > > > > mov w1, 0
> > > > > cntw x4
> > > > > mov z1.s, w0
> > > > > index z0.s, #0, #1
> > > > > ptrue p1.b, all
> > > > > ptrue p0.s, all
> > > > > .p2align 3,,7
> > > > > .L3:
> > > > > ld1w z2.s, p1/z, [x2, x3, lsl 2]
> > > > > add z3.s, z0.s, z1.s
> > > > > cmplo p2.s, p0/z, z1.s, z2.s
> > > > > b.any .L2
> > > > > st1w z3.s, p1, [x5, x3, lsl 2]
> > > > > add w1, w1, 1
> > > > > st1w z1.s, p1, [x2, x3, lsl 2]
> > > > > add x3, x3, x4
> > > > > incw z0.s
> > > > > cmp w3, 803
> > > > > bls .L3
> > > > > .L5:
> > > > > mov w0, 0
> > > > > ret
> > > > > .p2align 2,,3
> > > > > .L2:
> > > > > cntw x5
> > > > > mul w1, w1, w5
> > > > > cbz w5, .L5
> > > > > sxtw x1, w1
> > > > > sub w5, w5, #1
> > > > > add x5, x5, x1
> > > > > add x6, x2, 3216
> > > > > b .L6
> > > > > .p2align 2,,3
> > > > > .L14:
> > > > > str w0, [x2, x1, lsl 2]
> > > > > cmp x1, x5
> > > > > beq .L5
> > > > > mov x1, x4
> > > > > .L6:
> > > > > ldr w3, [x2, x1, lsl 2]
> > > > > add w4, w0, w1
> > > > > str w4, [x6, x1, lsl 2]
> > > > > add x4, x1, 1
> > > > > cmp w0, w3
> > > > > bcs .L14
> > > > > mov w0, 0
> > > > > ret
> > > > >
> > > > > On the workloads this work is based on we see between 2-3x
> > > > > performance uplift using this patch.
> > > > >
> > > > > Follow up plan:
> > > > > - Boolean vectorization has several shortcomings. I've filed
> > > > > PR110223 with
> > > > the
> > > > > bigger ones that cause vectorization to fail with this patch.
> > > > > - SLP support. This is planned for GCC 15 as for majority of the
> > > > > cases
> > build
> > > > > SLP itself fails.
> > > >
> > > > It would be nice to get at least single-lane SLP support working. I
> > > > think you need to treat the gcond as SLP root stmt and basically do
> > > > discovery on the condition as to as if it were a mask generating
> > > > condition.
> > >
> > > Hmm ok, will give it a try.
> > >
> > > >
> > > > Code generation would then simply schedule the gcond root instances
> > > > first (that would get you the code motion automagically).
> > >
> > > Right, so you're saying treat the gcond's as the seed, and stores as a
> > > sink.
> > > And then schedule only the instances without a gcond around such that
> > > we can still vectorize in place to get the branches. Ok, makes sense.
> > >
> > > >
> > > > So, add a new slp_instance_kind, for example
> > > > slp_inst_kind_early_break, and record the gcond as root stmt.
> > > > Possibly "pattern" recognizing
> > > >
> > > > gcond <_1 != _2>
> > > >
> > > > as
> > > >
> > > > _mask = _1 != _2;
> > > > gcond <_mask != 0>
> > > >
> > > > makes the SLP discovery less fiddly (but in theory you can of course
> > > > handle gconds directly).
> > > >
> > > > Is there any part of the series that can be pushed independelty? If
> > > > so I'll try to look at those parts first.
> > > >
> > >
> > > Aside from:
> > >
> > > [PATCH 4/21]middle-end: update loop peeling code to maintain LCSSA
> > > form for early breaks [PATCH 7/21]middle-end: update IV update code to
> > > support early breaks and arbitrary exits
> > >
> > > The rest lie dormant and don't do anything or disrupt the tree until those
> > two are in.
> > > The rest all just touch up different parts piecewise.
> > >
> > > They do rely on the new field introduced in:
> > >
> > > [PATCH 3/21]middle-end: Implement code motion and dependency analysis
> > > for early breaks
> > >
> > > But can split them out.
> > >
> > > I'll start respinning no #4 and #7 with your latest changes now.
> >
> > OK, I'll simply go 1-n then.
> >
> > Richard.
> >
> > > Thanks,
> > > Tamar
> > >
> > > > Thanks,
> > > > Richard.
>
--
Richard Biener <rguent...@suse.de>
SUSE Software Solutions Germany GmbH,
Frankenstrasse 146, 90461 Nuernberg, Germany;
GF: Ivo Totev, Andrew McDonald, Werner Knoblich; (HRB 36809, AG Nuernberg)
