[Vulkan][Intel] boolean matrix `and()` intrinsic in compute shader is producing garbage results.

[farzonl]

Summary

Intel Vulkan matrix_bool_and_operator produces incorrect results. This is different than #703 because the same pipeline completes successfully (no validation-layer errors), but the results are wrong.

offloader.exe fails

Test case

d:\repos\llvm-build\bin\offloader.exe -debug-layer -adapter-regex="intel" D:\repos\llvm-build\tools\OffloadTest\test\clang-vk\Basic\Matrix\Output\matrix_bool_and_operator.test.tmp/pipeline.yaml D:\repos\llvm-build\tools\OffloadTest\test\clang-vk\Basic\Matrix\Output\matrix_bool_and_operator.test.tmp.o
# executed command: 'd:\repos\llvm-build\bin\offloader.exe' -debug-layer -adapter-regex=intel 'D:\repos\llvm-build\tools\OffloadTest\test\clang-vk\Basic\Matrix\Output\matrix_bool_and_operator.test.tmp/pipeline.yaml' 'D:\repos\llvm-build\tools\OffloadTest\test\clang-vk\Basic\Matrix\Output\matrix_bool_and_operator.test.tmp.o'
# .---command stdout------------
# | Using Vulkan API
# | Physical device created.
# | Shader module created.
# | Copy command buffer created.
# | Memory buffers created.
# | Executed copy command buffer.
# | Execute command buffer created.
# | Descriptors: { type = 5, count = 5 }
# | Descriptor pool created.
# | Binding 5 descriptors.
# | Num Descriptor sets: 1
# | Updating Descriptor [0] { 0, 0 }
# | Updating Descriptor [1] { 0, 1 }
# | Updating Descriptor [2] { 0, 2 }
# | Updating Descriptor [3] { 0, 3 }
# | Updating Descriptor [4] { 0, 4 }
# | WriteDescriptors: 5
# | Descriptor sets created.
# | Compute pipeline created.
# | Dispatched compute shader: { 1, 1, 1 }
# | Commands created.
# | Executed compute command buffer.
# | Compute pipeline created.
# | Cleanup complete.
# `-----------------------------
# .---command stderr------------
# | Test failed: AndTrueTrueOut
# | Comparison Rule: BufferExact
# | Expected:
# | ---
# | Name:            AndTrueTrueExpectedOut
# | Format:          Int32
# | Data:            [ 1, 1, 1, 1, 1, 1 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Got:
# | ---
# | Name:            AndTrueTrueOut
# | Format:          Int32
# | Data:            [ 1, 1, 0, 0, 0, 0 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Full Hex 64bit representation of Expected Buffer Values:
# | [ 0x1, 0x1, 0x1, 0x1, 0x1, 0x1 ]
# | Full Hex 64bit representation of Actual Buffer Values:
# | [ 0x1, 0x1, 0x0, 0x0, 0x0, 0x0 ]
# |
# | Test failed: AndTrueMixOut
# | Comparison Rule: BufferExact
# | Expected:
# | ---
# | Name:            AndTrueMixExpectedOut
# | Format:          Int32
# | Data:            [ 0, 0, 0, 1, 1, 1 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Got:
# | ---
# | Name:            AndTrueMixOut
# | Format:          Int32
# | Data:            [ 0, 0, 0, 0, 0, 0 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Full Hex 64bit representation of Expected Buffer Values:
# | [ 0x0, 0x0, 0x0, 0x1, 0x1, 0x1 ]
# | Full Hex 64bit representation of Actual Buffer Values:
# | [ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 ]
# |
# | Test failed: AndTrueMix2Out
# | Comparison Rule: BufferExact
# | Expected:
# | ---
# | Name:            AndTrueMix2ExpectedOut
# | Format:          Int32
# | Data:            [ 1, 1, 1, 0, 0, 0 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Got:
# | ---
# | Name:            AndTrueMix2Out
# | Format:          Int32
# | Data:            [ 1, 1, 0, 0, 0, 0 ]
# | OutputProps:
# |   Height:          0
# |   Width:           0
# |   Depth:           0
# | ...
# | Full Hex 64bit representation of Expected Buffer Values:
# | [ 0x1, 0x1, 0x1, 0x0, 0x0, 0x0 ]
# | Full Hex 64bit representation of Actual Buffer Values:
# | [ 0x1, 0x1, 0x0, 0x0, 0x0, 0x0 ]
# |
# `-----------------------------
# error: command failed with exit status: 1

--