Commit Graph

567 Commits

Author SHA1 Message Date
David Neto
c843ef8ab5 validator: OpModuleProcessed allowed in layout section 7c
Recent spec fix from SPIR Working group:
  Allow OpModuleProcessed after debug names, but before any
  annotation instructions.
2017-09-07 17:45:51 -04:00
Andrey Tuganov
b36acbec0e Update MARK-V to version 1.01
Includes:
- Multi-sequence move-to-front
- Coding by id descriptor
- Statistical coding of non-id words
- Joint coding of opcode and num_operands

Removed explicit form Huffman codec constructor
- The standard use case for it is to be constructed from initializer list.

Using serialization for Huffman codecs
2017-09-06 16:03:16 -04:00
David Neto
25ddfec08e Inliner: Fix LoopMerge when inline into loop header of multi block loop
This adapts the fix for the single-block loop.  Split the loop like
before.  But when we move the OpLoopMerge back to the loop header,
redirect the continue target only when the original loop was a single
block loop.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/800
2017-09-05 19:46:24 -04:00
Andrey Tuganov
82df4bbd68 Add validation pass for arithmetic operations
The pass checks if arithmetic operations (such as OpFMul) receive
correct operands.
2017-09-05 12:21:53 -04:00
Andrey Tuganov
32cf85dd5a Fix mingw build (source/print.cpp)
source/print.cpp doesn't compile due to integer conversion.

Tested by @dneto0 on a Windows machine.
2017-09-01 16:07:18 -04:00
David Neto
860c4197b0 Inliner: Remap callee entry block id to single-trip loop header
Otherwise cloned phis can be invalid.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/790
2017-09-01 15:56:14 -04:00
David Neto
efff5fabfa Inline: Fix single-block loop caller cases
If the caller block is a single-block loop and inlining will
replace the caller block by several blocks, then:
- The original OpLoopMerge instruction will end up in the *last*
  such block.  That's the wrong place to put it.
- Move it back to the end of the first block.
- Update its Continue Target ID to point to the last block

We also have to take care of cases where the inlined code
begins with a structured header block.  In this case
we need to ensure the restored OpLoopMerge does not appear
in the same block as the merge instruction from the callee's
first block.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/787
2017-09-01 15:47:17 -04:00
David Neto
cff2cd3343 BasicBlock: add ctail, GetMergeInst, GetLoopMergeInst 2017-09-01 11:01:36 -04:00
Andrey Tuganov
725284c2ef Extension allows multiple same OpTypePointer types
SPV_KHR_variable_pointers allows OpTypePointer to declare multiple
pointer identical types.

https://github.com/KhronosGroup/SPIRV-Tools/issues/781
2017-09-01 10:14:15 -04:00
GregF
7c3de19ce7 DeadBranchElim: Fix dead block detection to ignore backedges
- DeadBranchElim: Make sure to mark orphan'd merge blocks and continue
targets as live.
- Add test with loop in dead branch
- Add test that orphan'd merge block is handled.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/776
2017-08-30 13:37:46 -04:00
GregF
a699d1ade7 Inline: Fix remapping of non-label forward references in callee phi 2017-08-29 18:35:05 -06:00
Andrey Tuganov
d41a52415a Fix encode zero bits on word boundary bug
Bit stream writer was manifesting incorrect behaviour when the following
two conditions were met:
- writer was on 64-bit word boundary
- WriteBits was invoked with num_bits=0 (can happen when a Huffman codec has only one
value)

The bug was causing very rare sporadic corruption which was detected by
tests after a random experimental change in MARK-V model.
2017-08-28 13:36:39 -04:00
David Neto
63e1e348b0 Show result id for CompositeInsert validation failure 2017-08-25 15:13:31 -04:00
David Neto
0167758727 Windows: Increase intensity of blue text 2017-08-24 10:40:17 -04:00
Lukas Hermanns
4fe8e389a7 Fix: background color was erroneously reset on Win32 platform.
Fix: background color was erroneously reset on Win32 platform.
2017-08-24 10:40:17 -04:00
GregF
429ca05b3f Opt: Create InlineOpaquePass
Only inline calls to functions with opaque params or return

TODO: Handle parameter type or return type where the opqaue
type is buried within an array.
2017-08-18 18:04:30 -04:00
GregF
c8c86a0d36 Opt: Have "size" passes process full entry point call tree.
Includes code to deal correctly with OpFunctionParameter. This
is needed by opaque propagation which may not exhaustively inline
entry point functions.

Adds ProcessEntryPointCallTree: a method to do work on the
functions in the entry point call trees in a deterministic order.
2017-08-18 10:16:01 -04:00
Andrey Tuganov
17d941af4f Huffman codec can serialize to text
Refactored the Huffman codec implementation and added ability to
serialize to C++-like text format. This would reduce the time-complexity
if loading hard-coded codecs.
2017-08-15 23:57:21 -04:00
Andrey Tuganov
78cf86150e Add id descriptor feature to SPIR-V
Id descriptors are computed as a recursive hash of all instructions used
to define an id. Descriptors are invarint of actual id values and
the similar code in different files would produce the same descriptors.

Multiple ids can have the same descriptor. For example
%1 = OpConstant %u32 1
%2 = OpConstant %u32 1
would produce two ids with the same descriptor. But
%3 = OpConstant %s32 1
%4 = OpConstant %u32 2
would have descriptors different from %1 and %2.

Descriptors will be used as handles of move-to-front sequences in SPIR-V
compression.
2017-08-10 18:44:52 -04:00
GregF
b0310a4156 ADCE: Add support for function calls
ADCE will now generate correct code in the presence of function calls.
This is needed for opaque type optimization needed by glslang. Currently
all function calls are marked as live. TODO: mark calls live only if they
write a non-local.
2017-08-10 17:30:05 -04:00
David Neto
2a1014be9c Inliner: callee can have early return that isn't multi-return
Avoid generating an invalid OpLabel.
Create the continue target for the single-trip loop only if
you actually created the header for the single-trip loop.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/755
2017-08-10 11:43:44 -04:00
GregF
f0fe601dc8 AccessChainConvert: Add HasOnlySupportedRefs()
This avoids conversion on variables which will not ultimately be optimized.
Also removed an obsolete restriction from FindTargetVars(). Also added
decorates to supported refs (eg. RelaxedPrecision). Also fixed name to
IsNonTypeDecorate().
2017-08-04 18:11:44 -04:00
GregF
e28bd39997 Inline: Split out InlineExhaustivePass from InlinePass 2017-08-04 17:56:46 -04:00
GregF
d9a450121e Mem2Reg: Allow Image and Sampler types as base target types. 2017-08-04 17:52:32 -04:00
GregF
f4b29f3bf7 Add CommonUniformElim pass
- UniformElim: Only process reachable blocks

- UniformElim: Don't reuse loads of samplers and images across blocks.
  Added a second phase which only reuses loads within a block for samplers
  and images.

- UniformElim: Upgrade CopyObject skipping in GetPtr

- UniformElim: Add extensions whitelist
  Currently disallowing SPV_KHR_variable_pointers because it doesn't
  handle extended pointer forms.

- UniformElim: Do not process shaders with GroupDecorate

- UniformElim: Bail on shaders with non-32-bit ints.

- UniformElim: Document support for only single index and add TODO.
2017-08-03 11:34:58 -04:00
GregF
c1b46eedbd Add MemPass, move all shared functions to it. 2017-08-02 14:24:02 -04:00
Andrey Tuganov
30bee67439 Add multi-sequence move-to-front implementation
Add MultiMoveToFront class which supports multiple move-to-front
sequences and allows to promote value in all sequences at once.

Added caching for last accessed sequence handle and last accessed value
in each sequence.
2017-08-02 14:07:24 -04:00
Andrey Tuganov
55b73a0365 Added C++ code generation to spirv-stats
The tool can now generate C++ code returning some of the historgrams and
Huffman codecs generated from those historgrams.
2017-08-01 15:41:42 -04:00
GregF
7954740d54 Opt: Delete names and decorations of dead instructions 2017-07-26 18:36:41 -04:00
Lei Zhang
9f6efc76c8 Opt: HasOnlySupportedRefs should consider OpCopyObject
This fixes test failure after merging the previous pull request.
2017-07-25 23:22:09 -04:00
Lei Zhang
4a539d77ef Revert "Revert "Opt: LocalBlockElim: Add HasOnlySupportedRefs""
This reverts commit df96e243c6.
2017-07-25 23:22:09 -04:00
GregF
1182415581 Add extension whitelists to size-reduction passes.
Currently only SPV_KHR_variable_pointers is disallowed in passes which
do pointer analysis. Positive and negative tests of the general extensions
mechanism were added to aggressive_dce but cover all passes.
2017-07-25 19:14:02 -04:00
Lei Zhang
df96e243c6 Revert "Opt: LocalBlockElim: Add HasOnlySupportedRefs"
This reverts commit 2d0f7fbc11.
2017-07-22 10:48:56 -04:00
greg-lunarg
2d0f7fbc11 Opt: LocalBlockElim: Add HasOnlySupportedRefs
Verifies that targeted variables have only access chain and direct
loads and stores as references.
2017-07-22 10:32:19 -04:00
GregF
adb237f3bd Fix handling of CopyObject in GetPtr and its call sites 2017-07-21 18:08:01 -04:00
Lenny Komow
e9e4393b1c Fix Visual Studio size_t cast compiler warning
Visual Studio was complaining about possible loss of data on 64-bit
builds, due to an implicit cast from size_t to int. This changes the
data to use an int with no cast.
2017-07-13 13:02:43 -06:00
Greg Fischer
fe24e0316f LocalMultiStore: Always put varId for backedge on loop phi function.
And always patch the backedge operand when patching phi functions. This
approach is more correct and cleaner. The previous code was generating
incorrect phis when the backedge block had no predecessors.
2017-07-12 16:42:07 -04:00
GregF
e2544ddc90 DeadBranchElim: Improve algorithm to only remove blocks with no predecessors
Must be careful not to remove blocks pointed at by unreachable blocks
2017-07-12 15:58:42 -04:00
David Neto
06d4fd52c2 Minor code review feedback on AggressiveDCE 2017-07-10 11:45:59 -04:00
GregF
9de4e69856 Add AggressiveDCEPass
Create aggressive dead code elimination pass
This pass eliminates unused code from functions. In addition,
it detects and eliminates code which may have spurious uses but which do
not contribute to the output of the function. The most common cause of
such code sequences is summations in loops whose result is no longer used
due to dead code elimination. This optimization has additional compile
time cost over standard dead code elimination.

This pass only processes entry point functions. It also only processes
shaders with logical addressing. It currently will not process functions
with function calls. It currently only supports the GLSL.std.450 extended
instruction set. It currently does not support any extensions.

This pass will be made more effective by first running passes that remove
dead control flow and inlines function calls.

This pass can be especially useful after running Local Access Chain
Conversion, which tends to cause cycles of dead code to be left after
Store/Load elimination passes are completed. These cycles cannot be
eliminated with standard dead code elimination.

Additionally: This transform uses a whitelist of instructions that it
knows do have side effects, (a.k.a. combinators).  It assumes other
instructions have side effects: it will not remove them, and assumes
they have side effects via their ID operands.
2017-07-10 11:30:25 -04:00
GregF
cc8bad3a5b Add LocalMultiStoreElim pass
A SSA local variable load/store elimination pass.
For every entry point function, eliminate all loads and stores of function
scope variables only referenced with non-access-chain loads and stores.
Eliminate the variables as well.

The presence of access chain references and function calls can inhibit
the above optimization.

Only shader modules with logical addressing are currently processed.
Currently modules with any extensions enabled are not processed. This
is left for future work.

This pass is most effective if preceeded by Inlining and
LocalAccessChainConvert. LocalSingleStoreElim and LocalSingleBlockElim
will reduce the work that this pass has to do.
2017-07-07 17:54:21 -04:00
GregF
52e247f221 DeadBranchElim: Add DeadBranchElimPass 2017-07-07 15:16:25 -04:00
David Neto
35a0695844 Include memory and semantics IDs when iterating over inbound IDs
Fixes Instruction::ForEachInId so it covers
SPV_OPERAND_TYPE_MEMORY_SEMANTICS_ID and SPV_OPERAND_TYPE_SCOPE_ID.
Future proof a bit by using the common spvIsIdType routine.

Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/697
2017-07-05 10:36:57 -04:00
Andrey Tuganov
abc6f5a672 MARK-V decoder supports extended instructions 2017-07-04 16:31:19 -04:00
d3x0r
fd70a1d7a0 Define variable to skip installation
If this is used as a static library in another project, this does not
need to be installed, and otherwise will just clutter the application's install.
To use, define SKIP_SPIRV_TOOLS_INSTALL which internally defines
ENABLE_SPIRV_TOOLS_INSTALL to control installation.

Also include GNUInstallDirs to get standard output 'lib' directory which is sometimes 'lib64' and not 'lib'
2017-07-04 12:24:44 -04:00
Chris Forbes
78338d5ba9 Convert pattern stack from deque to vector, and share it
Also move various vector::reserve calls to State ctor
Negligible perf benefit, but more tidy.
2017-07-04 12:02:26 -04:00
Andrey Tuganov
e842c17eb5 Added fixed width encoding to bit_stream
Fixed width encoding is intended to be used for small unsigned integers
when the upper bound is known both to the encoder and the decoder
(for example move-to-front rank).
2017-07-04 11:57:13 -04:00
Andrey Tuganov
73e8dac5b9 Added compression tool tools/spirv-markv. Work in progress.
Command line application is located at tools/spirv-markv
API at include/spirv-tools/markv.h

At the moment only very basic compression is implemented, mostly varint.
Scope of supported SPIR-V opcodes is also limited.

Using a simple move-to-front implementation instead of encoding mapped
ids.

Work in progress:
- Does not cover all of SPIR-V
- Does not promise compatibility of compression/decompression across
  different versions of the code.
2017-06-30 12:22:48 -04:00
Andrey Tuganov
8d3882a408 Added log(n) move-to-front implementation
The implementation is based on AVL and order statistic tree.

It accepts all kinds of values and the implementation
doesn't expect the behaviour to be consistent with id coding.

Intended by SPIR-V compression algorithms.
2017-06-29 16:16:18 -04:00
Andrey Tuganov
40a2829611 Added Huffman codec to utils
Attached ids to Huffman nodes for deterministic internal node
comparison.
2017-06-29 14:51:01 -04:00