<feed xmlns='http://www.w3.org/2005/Atom'>
<title>llvm-project.git/llvm/test/Transforms/LoopVectorize/predicate-switch.ll, branch users/mingmingl-llvm/samplefdo-profile-format</title>
<subtitle>Unnamed repository; edit this file 'description' to name the repository.
</subtitle>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/'/>
<entry>
<title>[VPlan] Don't build recipes for unconditional switches (#157323)</title>
<updated>2025-09-08T09:01:43+00:00</updated>
<author>
<name>Luke Lau</name>
<email>luke@igalia.com</email>
</author>
<published>2025-09-08T09:01:43+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=fe6e17840159f7ae9785d8506ff2987a2df7ab08'/>
<id>fe6e17840159f7ae9785d8506ff2987a2df7ab08</id>
<content type='text'>
In #157322 we crash because we try to infer a type for a VPReplicate
switch recipe.

My understanding was that these switches should be removed by
VPlanPredicator, but this switch survived through it because it was
unconditional, i.e. had no cases other than the default case.

This fixes #157322 by not emitting any recipes for unconditional
switches to begin with, similar to how we treat unconditional branches.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
In #157322 we crash because we try to infer a type for a VPReplicate
switch recipe.

My understanding was that these switches should be removed by
VPlanPredicator, but this switch survived through it because it was
unconditional, i.e. had no cases other than the default case.

This fixes #157322 by not emitting any recipes for unconditional
switches to begin with, similar to how we treat unconditional branches.</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Run final VPlan simplifications before codegen.</title>
<updated>2025-08-16T17:54:27+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-08-16T17:54:27+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=351d398a37d8545d8722ba36ad364067e94a69df'/>
<id>351d398a37d8545d8722ba36ad364067e94a69df</id>
<content type='text'>
Dissolving the hierarchical VPlan CFG and converting abstract to
concrete recipes can expose additional simplification opportunities.

Do a final run of simplifyRecipes before executing the VPlan.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Dissolving the hierarchical VPlan CFG and converting abstract to
concrete recipes can expose additional simplification opportunities.

Do a final run of simplifyRecipes before executing the VPlan.
</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Remove VPVectorPointer for part 0 after unrolling. (#149735)</title>
<updated>2025-07-27T12:53:26+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-07-27T12:53:26+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=89ae0858599549ce6f7be1b789b04a5301a626ba'/>
<id>89ae0858599549ce6f7be1b789b04a5301a626ba</id>
<content type='text'>
VPVectorPointer for part 0 is just the pointer operand. Simplify it
after unrolling. This removes a large number of redundant GEPs with
index 0.

PR: https://github.com/llvm/llvm-project/pull/149735</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
VPVectorPointer for part 0 is just the pointer operand. Simplify it
after unrolling. This removes a large number of redundant GEPs with
index 0.

PR: https://github.com/llvm/llvm-project/pull/149735</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Materialize constant vector trip counts before final opts. (#142309)</title>
<updated>2025-07-26T16:16:36+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-07-26T16:16:36+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=fa3ec0c17c48349e6027710d234c83e7bfeaf854'/>
<id>fa3ec0c17c48349e6027710d234c83e7bfeaf854</id>
<content type='text'>
Materialize constant vector trip counts before ::execute, if the trip
count can be computed as Original (TC / (VF * UF)) * (VF * UF). For now
this excludes when the tail is folded or scalar epilogues are required.

This enables removing a number of redundant branches from the middle
block.

For now this is also only done when not vectorizing the epilogue, as the
simplification complicates stitching the 2 plans together.

PR: https://github.com/llvm/llvm-project/pull/142309</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Materialize constant vector trip counts before ::execute, if the trip
count can be computed as Original (TC / (VF * UF)) * (VF * UF). For now
this excludes when the tail is folded or scalar epilogues are required.

This enables removing a number of redundant branches from the middle
block.

For now this is also only done when not vectorizing the epilogue, as the
simplification complicates stitching the 2 plans together.

PR: https://github.com/llvm/llvm-project/pull/142309</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Generalize SCALAR-STEPS removal to any unroll factor.</title>
<updated>2025-03-26T21:03:50+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-03-26T21:03:48+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=2c7d40b2f0c5e68e90297b605ceb70c55d201bca'/>
<id>2c7d40b2f0c5e68e90297b605ceb70c55d201bca</id>
<content type='text'>
Follow-up to dfca6c0d3bf9d1a056 to extend isUnrolled handle any unrolled
VPlan, which means there's a single UF, but it will be &gt; 1 if unrolling
took place.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Follow-up to dfca6c0d3bf9d1a056 to extend isUnrolled handle any unrolled
VPlan, which means there's a single UF, but it will be &gt; 1 if unrolling
took place.
</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Remove no-op SCALAR-STEPS after unrolling. (#123655)</title>
<updated>2025-03-25T12:57:24+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-03-25T12:57:24+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=dfca6c0d3bf9d1a0565b3ff46813ddd66bd493e9'/>
<id>dfca6c0d3bf9d1a0565b3ff46813ddd66bd493e9</id>
<content type='text'>
After unrolling, there may be additional simplifications that can be
applied. One example is removing SCALAR-STEPS for the first part where
only the first lane is demanded.

This removes redundant adds of 0 from a large number of tests (~200),
many which I am still working on updating.

In preparation for removing redundant WideIV steps added in
https://github.com/llvm/llvm-project/pull/119284.

PR: https://github.com/llvm/llvm-project/pull/123655</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
After unrolling, there may be additional simplifications that can be
applied. One example is removing SCALAR-STEPS for the first part where
only the first lane is demanded.

This removes redundant adds of 0 from a large number of tests (~200),
many which I am still working on updating.

In preparation for removing redundant WideIV steps added in
https://github.com/llvm/llvm-project/pull/119284.

PR: https://github.com/llvm/llvm-project/pull/123655</pre>
</div>
</content>
</entry>
<entry>
<title>[LLVM][IR] Use splat syntax when printing Constant[Data]Vector. (#112548)</title>
<updated>2024-11-06T11:53:33+00:00</updated>
<author>
<name>Paul Walker</name>
<email>paul.walker@arm.com</email>
</author>
<published>2024-11-06T11:53:33+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=38fffa630ee80163dc65e759392ad29798905679'/>
<id>38fffa630ee80163dc65e759392ad29798905679</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
<entry>
<title>[VPlan] Run DCE after unrolling.</title>
<updated>2024-09-22T21:08:46+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2024-09-22T20:50:14+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=53266f73f037bd20bcbbd7852fd0c6a7703b4c38'/>
<id>53266f73f037bd20bcbbd7852fd0c6a7703b4c38</id>
<content type='text'>
This cleans up a number of dead recipes after unrolling if only their
first or last parts are used. This simplifies a number of tests.

Fixes https://github.com/llvm/llvm-project/issues/109581.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This cleans up a number of dead recipes after unrolling if only their
first or last parts are used. This simplifies a number of tests.

Fixes https://github.com/llvm/llvm-project/issues/109581.
</pre>
</div>
</content>
</entry>
<entry>
<title>[LV] Only OR unique edges when creating block-in masks.</title>
<updated>2024-08-12T09:17:40+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2024-08-12T09:17:39+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=db0603cb7b8534bffdd0459f5eb5a3b98ea962ef'/>
<id>db0603cb7b8534bffdd0459f5eb5a3b98ea962ef</id>
<content type='text'>
This removes redundant ORs of matching masks.

Follow-up to f0df4fbd0c7b to reduce the number of redundant ORs for
masks.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This removes redundant ORs of matching masks.

Follow-up to f0df4fbd0c7b to reduce the number of redundant ORs for
masks.
</pre>
</div>
</content>
</entry>
<entry>
<title>[LV] Support generating masks for switch terminators. (#99808)</title>
<updated>2024-08-11T18:38:36+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2024-08-11T18:38:36+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=f0df4fbd0c7b6bb369ceaa1fd6f9e0c88d781ae5'/>
<id>f0df4fbd0c7b6bb369ceaa1fd6f9e0c88d781ae5</id>
<content type='text'>
Update createEdgeMask to created masks where the terminator in Src is a
switch. We need to handle 2 separate cases:

1. Dst is not the default desintation. Dst is reached if any of the
cases with destination == Dst are taken. Join the conditions for each
case where destination == Dst using a logical OR.
2. Dst is the default destination. Dst is reached if none of the cases
with destination != Dst are taken. Join the conditions for each case
where the destination is != Dst using a logical OR and negate it.

Edge masks are created for every destination of cases and/or 
default when requesting a mask where the source is a switch.

Fixes https://github.com/llvm/llvm-project/issues/48188.

PR: https://github.com/llvm/llvm-project/pull/99808</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Update createEdgeMask to created masks where the terminator in Src is a
switch. We need to handle 2 separate cases:

1. Dst is not the default desintation. Dst is reached if any of the
cases with destination == Dst are taken. Join the conditions for each
case where destination == Dst using a logical OR.
2. Dst is the default destination. Dst is reached if none of the cases
with destination != Dst are taken. Join the conditions for each case
where the destination is != Dst using a logical OR and negate it.

Edge masks are created for every destination of cases and/or 
default when requesting a mask where the source is a switch.

Fixes https://github.com/llvm/llvm-project/issues/48188.

PR: https://github.com/llvm/llvm-project/pull/99808</pre>
</div>
</content>
</entry>
</feed>
