This changes the semantics of `HLFHE.dot_eint_int` from memref-based
reference semantics to tensor-based value semantics. The former:
"HLFHE.dot_eint_int"(%arg0, %arg1, %arg2) :
(memref<Nx!HLFHE.eint<0>>, memref<Nxi32>, memref<!HLFHE.eint<0>>) -> ()
becomes:
"HLFHE.dot_eint_int"(%arg0, %arg1) :
(tensor<Nx!HLFHE.eint<0>>, tensor<Nxi32>) -> !HLFHE.eint<0>
As a side effect, data-flow analyses become much easier. With the
previous memref type of the plaintext argument it is difficult to
check whether the plaintext values are statically defined constants or
originate from a memory region changed at execution time (e.g., for
analyses evaluating the impact on noise). Changing the plaintext type
from `memref` to `vector` makes such analyses significantly easier.
* feat(compiler): low level fhe dialect
* feat(compiler): using generated printer/parser in LowLFHE
* feat(compiler): new types and ops for LowLFHE
* tests(compiler): LowLFHE types and ops
* feat(compiler): fill ops
* cleanup
* summary + description
* tests(compiler): use new CLI args
* formatting
This adds a new command line option
`--convert-hlfhe-tensor-ops-to-linalg` that invokes a conversion pass
replacing any HLFHE tensor operation with an appropriate instance of
`linalg.generic`.
The option `--verify-diagnostics` allows the user to interleave
invalid source code with diagnostic messages that should be produced
during parsing and checks if these messages are actually produced. The
expected diagnostic messages are specified using `// expected-error`
markers.
The new option `--split-input-file` allows the user to add `// -----`
markers to a source file, causing the file to be split into multiple
chunks that are treated as if they were specified in different files.
Use a more idiomatic approach for I/O using `llvm::SourceManager` to
manage input files and `mlir::openOutputFile` to write to the
output. This commit prepares management of diagnostics in subsequent
changes.
LLVM already provides a command line parser, covering the use of tclap
in the sources. This patch uses the LLVM command line parser instead
of tclap and thus reduces the number of external dependencies.
