<feed xmlns='http://www.w3.org/2005/Atom'>
<title>llvm-project.git/llvm/test/CodeGen/AMDGPU/amdgpu-alias-analysis.ll, branch users/nico/python-2</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>AMDGPU: Replace undef global initializers in tests with poison (#131051)</title>
<updated>2025-03-13T06:41:03+00:00</updated>
<author>
<name>Matt Arsenault</name>
<email>Matthew.Arsenault@amd.com</email>
</author>
<published>2025-03-13T06:41:03+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=26ae98c4d73d13f973eeb8450e0f5c7d1cd25993'/>
<id>26ae98c4d73d13f973eeb8450e0f5c7d1cd25993</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
<entry>
<title>AMDGPU: Replace ptr undef in tests with poison (#131050)</title>
<updated>2025-03-13T06:34:52+00:00</updated>
<author>
<name>Matt Arsenault</name>
<email>Matthew.Arsenault@amd.com</email>
</author>
<published>2025-03-13T06:34:52+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=7811075b6f667d9a92038fce48262c55327c10d0'/>
<id>7811075b6f667d9a92038fce48262c55327c10d0</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
<entry>
<title>[AMDGPU] Fix spurious NoAlias results (#122309)</title>
<updated>2025-01-20T14:19:30+00:00</updated>
<author>
<name>Fraser Cormack</name>
<email>fraser@codeplay.com</email>
</author>
<published>2025-01-20T14:19:30+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=9cf24652e783147b5063925ce025b9f8a6e8fe25'/>
<id>9cf24652e783147b5063925ce025b9f8a6e8fe25</id>
<content type='text'>
After a30e50fc, AMDGPUAAResult is being called in more situations where
BasicAA isn't sure. This exposed some regressions where NoAlias is being
incorrectly returned for two identical pointers.

The fix is to check the underlying objects for equality before returning
NoAlias.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
After a30e50fc, AMDGPUAAResult is being called in more situations where
BasicAA isn't sure. This exposed some regressions where NoAlias is being
incorrectly returned for two identical pointers.

The fix is to check the underlying objects for equality before returning
NoAlias.</pre>
</div>
</content>
</entry>
<entry>
<title>[AMDGPU] - Add address space for strided buffers (#74471)</title>
<updated>2023-12-15T14:49:25+00:00</updated>
<author>
<name>Jessica Del</name>
<email>50999226+OutOfCache@users.noreply.github.com</email>
</author>
<published>2023-12-15T14:49:25+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=32f9983c064557883223b585810eac3c6797d500'/>
<id>32f9983c064557883223b585810eac3c6797d500</id>
<content type='text'>
This is an experimental address space for strided buffers. These buffers
can have structs as elements and
a stride &gt; 1.
These pointers allow the indexed access in units of stride, i.e., they
point at `buffer[index * stride]`.
Thus, we can use the `idxen` modifier for buffer loads.

We assign address space 9 to 192-bit buffer pointers which contain a
128-bit descriptor, a 32-bit offset and a 32-bit index. Essentially,
they are fat buffer pointers with an additional 32-bit index.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This is an experimental address space for strided buffers. These buffers
can have structs as elements and
a stride &gt; 1.
These pointers allow the indexed access in units of stride, i.e., they
point at `buffer[index * stride]`.
Thus, we can use the `idxen` modifier for buffer loads.

We assign address space 9 to 192-bit buffer pointers which contain a
128-bit descriptor, a 32-bit offset and a 32-bit index. Essentially,
they are fat buffer pointers with an additional 32-bit index.</pre>
</div>
</content>
</entry>
<entry>
<title>AMDGPU: Convert amdgpu-alias-analysis.ll to opaque pointers</title>
<updated>2022-11-29T23:08:53+00:00</updated>
<author>
<name>Matt Arsenault</name>
<email>Matthew.Arsenault@amd.com</email>
</author>
<published>2022-11-29T22:54:34+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=a1ac8902e3db3e354948aeb920f1dfc372089c19'/>
<id>a1ac8902e3db3e354948aeb920f1dfc372089c19</id>
<content type='text'>
This one was slightly tricky. The AA debug printing usually, but not
always, uses the old pointer syntax. Also, we need to stop folding out
0 index GEPs in a few of these cases.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This one was slightly tricky. The AA debug printing usually, but not
always, uses the old pointer syntax. Also, we need to stop folding out
0 index GEPs in a few of these cases.
</pre>
</div>
</content>
</entry>
<entry>
<title>[AAEval] Make compatible with opaque pointers</title>
<updated>2022-03-16T09:02:11+00:00</updated>
<author>
<name>Nikita Popov</name>
<email>npopov@redhat.com</email>
</author>
<published>2022-03-15T09:18:19+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=57d57b1afd87d714d622b4287f891d17a474ecb6'/>
<id>57d57b1afd87d714d622b4287f891d17a474ecb6</id>
<content type='text'>
With opaque pointers, we cannot use the pointer element type to
determine the LocationSize for the AA query. Instead, -aa-eval
tests are now required to have an explicit load or store for any
pointer they want to compute alias results for, and the load/store
types are used to determine the location size.

This may affect ordering of results, and sorting within one result,
as the type is not considered part of the sorted string anymore.

To somewhat minimize the churn, printing still uses faux typed
pointer notation.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
With opaque pointers, we cannot use the pointer element type to
determine the LocationSize for the AA query. Instead, -aa-eval
tests are now required to have an explicit load or store for any
pointer they want to compute alias results for, and the load/store
types are used to determine the location size.

This may affect ordering of results, and sorting within one result,
as the type is not considered part of the sorted string anymore.

To somewhat minimize the churn, printing still uses faux typed
pointer notation.
</pre>
</div>
</content>
</entry>
<entry>
<title>[test] Use new PM for -aa-eval tests</title>
<updated>2022-03-08T22:15:53+00:00</updated>
<author>
<name>Arthur Eubanks</name>
<email>aeubanks@google.com</email>
</author>
<published>2022-03-08T22:00:08+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=b81d5baa0fb06b17e646e703c9771478ca190249'/>
<id>b81d5baa0fb06b17e646e703c9771478ca190249</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
<entry>
<title>[NewPM][AMDGPU] Make amdgpu-aa work with NewPM</title>
<updated>2021-01-04T20:36:27+00:00</updated>
<author>
<name>Arthur Eubanks</name>
<email>aeubanks@google.com</email>
</author>
<published>2020-12-29T20:30:25+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=191552344bba04c428de4a34b83d6f7537a4a596'/>
<id>191552344bba04c428de4a34b83d6f7537a4a596</id>
<content type='text'>
An AMDGPUAA class already existed that was supposed to work with the new
PM, but it wasn't tested and was a bit broken.

Fix up the existing classes to have the right keys/parameters.
Wire up AMDGPUAA inside AMDGPUTargetMachine.

Add it to the list of alias analyses for the "default" AAManager since
in adjustPassManager() amdgpu-aa is added into the pipeline at the
beginning.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93914
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
An AMDGPUAA class already existed that was supposed to work with the new
PM, but it wasn't tested and was a bit broken.

Fix up the existing classes to have the right keys/parameters.
Wire up AMDGPUAA inside AMDGPUTargetMachine.

Add it to the list of alias analyses for the "default" AAManager since
in adjustPassManager() amdgpu-aa is added into the pipeline at the
beginning.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D93914
</pre>
</div>
</content>
</entry>
<entry>
<title>[amdgpu] Enhance AMDGPU AA.</title>
<updated>2020-10-20T13:54:12+00:00</updated>
<author>
<name>Michael Liao</name>
<email>michael.hliao@gmail.com</email>
</author>
<published>2020-10-08T22:46:00+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=2a0e4d1c01c93bc326957c0ebf2a3203dfb145d9'/>
<id>2a0e4d1c01c93bc326957c0ebf2a3203dfb145d9</id>
<content type='text'>
- In general, a generic point may alias to pointers in all other address
  spaces. However, for certain cases enforced by the programming model,
  we may found a generic point won't alias to pointers to local objects.
  * When a generic pointer is loaded from the constant address space, it
    could only be a pointer to the GLOBAL or CONSTANT address space.
    Thus, it won't alias to pointers to the PRIVATE or LOCAL address
    space.
  * When a generic pointer is passed as a kernel argument, it also could
    only be a pointer to the GLOBAL or CONSTANT address space. Thus, it
    also won't alias to pointers to the PRIVATE or LOCAL address space.

Differential Revision: https://reviews.llvm.org/D89525
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
- In general, a generic point may alias to pointers in all other address
  spaces. However, for certain cases enforced by the programming model,
  we may found a generic point won't alias to pointers to local objects.
  * When a generic pointer is loaded from the constant address space, it
    could only be a pointer to the GLOBAL or CONSTANT address space.
    Thus, it won't alias to pointers to the PRIVATE or LOCAL address
    space.
  * When a generic pointer is passed as a kernel argument, it also could
    only be a pointer to the GLOBAL or CONSTANT address space. Thus, it
    also won't alias to pointers to the PRIVATE or LOCAL address space.

Differential Revision: https://reviews.llvm.org/D89525
</pre>
</div>
</content>
</entry>
<entry>
<title>[BasicAA] Rename -disable-basicaa to -disable-basic-aa to be consistent with the canonical name "basic-aa"</title>
<updated>2020-06-27T03:55:44+00:00</updated>
<author>
<name>Fangrui Song</name>
<email>maskray@google.com</email>
</author>
<published>2020-06-27T03:55:44+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=4cd19a6e15120cb21c6178f38f5c09fd7123c391'/>
<id>4cd19a6e15120cb21c6178f38f5c09fd7123c391</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
</feed>
