9 Commits

Author SHA1 Message Date
Jon Chesterfield
d3316bc111 [amdgpu] Delete elide-module-lds attribute
Requires D155190

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D155238
2023-07-14 00:36:33 +01:00
Jon Chesterfield
e75ce77cd7 [amdgpu][lds] Fix missing markUsedByKernel calls and undef lookup table elements
More robust association between the kernels and lds struct.

Use poison instead of value() for lookup table elements introduced by dynamic lds lowering.

Extracted from D154946, new test from there verbatim. Segv fixed.

Fixes issues/63338

Fixes SWDEV-404491

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D154972
2023-07-12 00:37:21 +01:00
Nikita Popov
bdf2fbba9c [AMDGPU] Convert some tests to opaque pointers (NFC) 2022-12-19 12:41:13 +01:00
Jon Chesterfield
d77ae7f251 [amdgpu] Reimplement LDS lowering
Renames the current lowering scheme to "module" and introduces two new
ones, "kernel" and "table", plus a "hybrid" that chooses between those three
on a per-variable basis.

Unit tests are set up to pass with the default lowering of "module" or "hybrid"
with this patch defaulting to "module", which will be a less dramatic codegen
change relative to the current. This reflects the sparsity of test coverage for
the table lowering method. Hybrid is better than module in every respect and
will be default in a subsequent patch.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D139433
2022-12-07 22:02:54 +00:00
Nico Weber
a862d09a92 Revert "[amdgpu] Reimplement LDS lowering"
This reverts commit 982017240d7f25a8a6969b8b73dc51f9ac5b93ed.
Breaks check-llvm, see https://reviews.llvm.org/D139433#3974862
2022-12-06 12:01:36 -05:00
Jon Chesterfield
982017240d [amdgpu] Reimplement LDS lowering
Renames the current lowering scheme to "module" and introduces two new
ones, "kernel" and "table", plus a "hybrid" that chooses between those three
on a per-variable basis.

Unit tests are set up to pass with the default lowering of "module" or "hybrid"
with this patch defaulting to "module", which will be a less dramatic codegen
change relative to the current. This reflects the sparsity of test coverage for
the table lowering method. Hybrid is better than module in every respect and
will be default in a subsequent patch.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D139433
2022-12-06 16:28:15 +00:00
Jon Chesterfield
bc78c09952 [amdgpu] Elide module lds allocation in kernels with no callees
Introduces a string attribute, amdgpu-requires-module-lds, to allow
eliding the module.lds block from kernels. Will allocate the block as before
if the attribute is missing or has its default value of true.

Patch uses the new attribute to detect the simplest possible instance of this,
where a kernel makes no calls and thus cannot call any functions that use LDS.

Tests updated to match, coverage was already good. Interesting cases is in
lower-module-lds-offsets where annotating the kernel allows the backend to pick
a different (in this case better) variable ordering than previously. A later
patch will avoid moving kernel variables into module.lds when the kernel can
have this attribute, allowing optimal ordering and locally unused variable
elimination.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D122091
2022-05-04 22:42:07 +01:00
Stanislav Mekhanoshin
8de4db697f [AMDGPU] Lower kernel LDS into a sorted structure
Differential Revision: https://reviews.llvm.org/D102954
2021-05-25 11:29:29 -07:00
Stanislav Mekhanoshin
748db5bfac [AMDGPU] Fix module LDS selection
Accesses to global module LDS variable start from null,
but kernel also thinks its variables start address is
null. Fixed by not using a null as an address.

Differential Revision: https://reviews.llvm.org/D102882
2021-05-20 15:59:01 -07:00