Commit graph

496 commits

Author SHA1 Message Date
Fernando Sahmkow 23cabc98db Shader_IR: Fix regression on TLD4
Originally on the last commit I thought TLD4 acted the same as TLD4S and 
didn't have a mask. It actually does have a component mask. This commit 
corrects that.
2019-10-30 21:14:57 -04:00
Fernando Sahmkow 9293c3a0f2 Shader_IR: Fix TLD4 and add Bindless Variant.
This commit fixes an issue where not all 4 results of tld4 were being
written, the color component was defaulted to red, among other things.
It also implements the bindless variant.
2019-10-30 12:02:03 -04:00
Rodrigo Locatti 26f3e18c5c
Merge pull request #2976 from FernandoS27/cache-fast-brx-rebased
Implement Fast BRX, fix TXQ and addapt the Shader Cache for it
2019-10-26 16:56:13 -03:00
Fernando Sahmkow be856a38d6 Shader_IR: Address Feedback. 2019-10-26 15:38:30 -04:00
Rodrigo Locatti a0d79085c4
Merge pull request #3027 from lioncash/lookup
shader_ir: Use std::array with std::pair instead of std::unordered_map
2019-10-26 05:49:15 -03:00
Rodrigo Locatti d52598173d
Merge pull request #3013 from FernandoS27/tld4s-fix
Shader_Ir: Fix TLD4S from using a component mask.
2019-10-25 20:06:26 -03:00
ReinUsesLisp 78f3e8a757 gl_shader_cache: Implement locker variants invalidation 2019-10-25 09:01:32 -04:00
ReinUsesLisp ec85648af3 gl_shader_disk_cache: Store and load fast BRX 2019-10-25 09:01:31 -04:00
ReinUsesLisp fa2c297f3e const_buffer_locker: Minor style changes 2019-10-25 09:01:31 -04:00
ReinUsesLisp 7b81ba4d8a gl_shader_decompiler: Move entries to a separate function 2019-10-25 09:01:31 -04:00
Fernando Sahmkow 1244f2d368 Shader_IR: Implement Fast BRX and allow multi-branches in the CFG. 2019-10-25 09:01:31 -04:00
Fernando Sahmkow a05120ec0b Shader_IR: Correct typo in Consistent method. 2019-10-25 09:01:30 -04:00
Fernando Sahmkow 33fcec3502 Shader_IR: allow lookup of texture samplers within the shader_ir for instructions that don't provide it 2019-10-25 09:01:30 -04:00
Fernando Sahmkow 8909f52166 Shader_IR: Implement Fast BRX and allow multi-branches in the CFG. 2019-10-25 09:01:30 -04:00
Fernando Sahmkow acd6441134 Shader_Cache: setup connection of ConstBufferLocker 2019-10-25 09:01:29 -04:00
Fernando Sahmkow 1a58f45d76 VideoCore: Unify const buffer accessing along engines and provide ConstBufferLocker class to shaders. 2019-10-25 09:01:29 -04:00
Fernando Sahmkow 2ef696c85a Shader_IR: Implement BRX tracking. 2019-10-25 09:01:29 -04:00
Lioncash 382717172e shader_ir: Use std::array with pair instead of unordered_map
Given the overall size of the maps are very small, we can use arrays of
pairs here instead of always heap allocating a new map every time the
functions are called. Given the small size of the maps, the difference
in container lookups are negligible, especially given the entries are
already sorted.
2019-10-24 00:25:38 -04:00
Lioncash 1f5401c89c video_core/shader: Resolve instances of variable shadowing
Silences a few -Wshadow warnings.
2019-10-23 23:00:31 -04:00
Fernando Sahmkow 1509d2ffbd Shader_Ir: Fix TLD4S from using a component mask.
TLD4S always outputs 4 values, the previous code checked a component 
mask and omitted those values that weren't part of it. This commit 
corrects that and makes sure all 4 values are set.
2019-10-22 10:59:07 -04:00
ReinUsesLisp 1ea07954fb shader_ir/memory: Ignore global memory when tracking fails
Ignore global memory operations instead of invoking undefined behaviour
when constant buffer tracking fails and we are blasting through asserts,
ignore the operation.

In the case of LDG this means filling the destination registers with
zeroes; for STG this means ignore the instruction as a whole.

The default behaviour is still to abort execution on failure.
2019-10-22 02:49:17 -03:00
Lioncash 074b38b7a9 video_core/shader/ast: Make ShowCurrentState() and SanityCheck() const member functions
These can also trivially be made const member functions, with the
addition of a few consts.
2019-10-17 20:59:48 -04:00
Lioncash 222f4b45eb video_core/shader/ast: Make ASTManager::Print a const member function
Given all visiting functions never modify the nodes, we can trivially
make this a const member function.
2019-10-17 20:56:39 -04:00
Lioncash 7831e86c34 video_core/shader/ast: Make ExprPrinter members private
This member already has an accessor, so there's no need for it to be
public.
2019-10-17 20:39:36 -04:00
Lioncash a2eccbf075 video_core/shader/ast: Make Indent() return a string_view
The returned string is simply a substring of our constexpr tabs
string_view, so we can just use a string_view here as well, since the
original string_view is guaranteed to always exist.

Now the function is fully non-allocating.
2019-10-17 20:29:00 -04:00
Lioncash 15d177a6ac video_core/shader/ast: Make Indent() private
It's never used outside of this class, so we can narrow its scope down.
2019-10-17 20:26:13 -04:00
Lioncash 7f6a8a33d4 video_core/shader/ast: Rename Ident() to Indent()
This can be confusing, given "ident" is generally used as a shorthand
for "identifier".
2019-10-17 20:26:13 -04:00
Lioncash 081530686c video_core/shader/ast: Make use of fmt where applicable
Makes a few strings nicer to read and also eliminates a bit of string
churn with operator+.
2019-10-17 20:26:10 -04:00
bunnei 9fe8072c67
Merge pull request #2980 from lioncash/warn
maxwell_3d: Silence truncation warnings
2019-10-17 14:02:16 -04:00
Lioncash 77b4916b33 control_flow: Silence truncation warnings
This can be trivially fixed by making the input size a size_t.
CFGRebuildState's constructor parameter is already a std::size_t, so
this just makes the size type fully conform with it.
2019-10-15 19:10:28 -04:00
Lioncash 67658dd6e8 shader/node: std::move Meta instance within OperationNode constructor
Allows usages of the constructor to avoid an unnecessary copy.
2019-10-15 18:21:59 -04:00
ReinUsesLisp 3d0f357307
shader/half_set_predicate: Fix HSETP2 for constant buffers
HSETP2 when used with a constant buffer parses the second operand type
as F32. This is not configurable.
2019-10-07 14:49:47 -03:00
ReinUsesLisp 632c9e4ee3
shader/half_set_predicate: Reduce DEBUG_ASSERT to LOG_DEBUG 2019-10-07 14:48:58 -03:00
Lioncash f883cd4f0e video_core/control_flow: Eliminate variable shadowing warnings 2019-10-05 09:14:27 -04:00
Lioncash 25702b6256 video_core/control_flow: Eliminate pessimizing moves
These can inhibit the ability of a compiler to perform RVO.
2019-10-05 09:14:27 -04:00
Lioncash d82b181d44 video_core/ast: Unindent most of IsFullyDecompiled() by one level 2019-10-05 09:14:27 -04:00
Lioncash 6c41d1cd7e video_core/ast: Make ShowCurrentState() take a string_view instead of std::string
Allows the function to be non-allocating in terms of the output string.
2019-10-05 09:14:27 -04:00
Lioncash 3c54edae24 video_core/ast: Eliminate variable shadowing warnings 2019-10-05 09:14:26 -04:00
Lioncash 5a0a9c7449 video_core/ast: Replace std::string with a constexpr std::string_view
Same behavior, but without the need to heap allocate
2019-10-05 09:14:26 -04:00
Lioncash 3a20d9734f video_core/ast: Default the move constructor and assignment operator
This is behaviorally equivalent and also fixes a bug where some members
weren't being moved over.
2019-10-05 09:14:26 -04:00
Lioncash 43503a69bf video_core/{ast, expr}: Organize forward declaration
Keeps them alphabetically sorted for readability.
2019-10-05 09:14:26 -04:00
Lioncash 50ad745585 video_core/expr: Supply operator!= along with operator==
Provides logical symmetry to the interface.
2019-10-05 09:14:26 -04:00
Lioncash 8eb1398f8d video_core/{ast, expr}: Use std::move where applicable
Avoids unnecessary atomic reference count increments and decrements.
2019-10-05 09:14:23 -04:00
Lioncash 8e0c80f269 video_core/ast: Supply const accessors for data where applicable
Provides const equivalents of data accessors for use within const
contexts.
2019-10-05 08:22:03 -04:00
Fernando Sahmkow e6eae4b815 Shader_ir: Address feedback 2019-10-04 18:52:57 -04:00
Fernando Sahmkow 3c09d9abe6 Shader_Ir: Address Feedback and clang format. 2019-10-04 18:52:57 -04:00
Fernando Sahmkow 7c756baa77 Shader_IR: clean up AST handling and add documentation. 2019-10-04 18:52:55 -04:00
Fernando Sahmkow 5ea740beb5 Shader_IR: Correct OutwardMoves for Ifs 2019-10-04 18:52:54 -04:00
Fernando Sahmkow b3c46d6948 Shader_IR: corrections and clang-format 2019-10-04 18:52:53 -04:00
Fernando Sahmkow 2e9a810423 Shader_IR: allow else derivation to be optional. 2019-10-04 18:52:52 -04:00