Documentation for NVQC #1412

Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…quantum into experimental/python

Updating experimental/python with updates from main

Updating branch with latest from main

This PR also removes UnitaryOp and init_state from the experimental branch

…ith-mainline Align feature branch with mainline

Align feature branch with the latest updates on mainline

…cpp'

Align feature branch with mainline

* Removed left-over tests pertaining to 'initialize_state' functionality * Adjusted calls for assertion * Addressing the following warning - "DeprecationWarning: expectation_z() is deprecated, use expectation() with the same argument structure." * Renamed tests under 'eager' directory so top-level 'tests' can be run with 'pytest' * Apply Python formatting to the new examples in 'python_mlir' and 'python_runtime' * Spelling check updates

-- Resolved conflicts on 'runtime/cudaq/platform/default/rest/RemoteRESTQPU.cpp' and 'runtime/cudaq/qis/execution_manager.h'

Align feature branch with mainline

…tal/python Align with mainline

[experimental/python] Feature branch align with mainline

* * Move files from 'experimental' directory to top-level 'python' * Install the new Python stack * Mark failing python/unittests as 'skipped' with the error. * Adjusting the build directory structure - removed 'python' subdirectory * Annotate skipped tests with '[SKIP_TEST]' comment for easy tracking. * Changed the PYTHONPATH in test commands since 'python_packages/cudaq' was dropped in python/CMakeLists.txt * Python backend tests updated to explicitly override QPU to 'py_remote_rest' * Skipping 'py-cudaq-tests-quake' since they are timing out. * * Edits to document generation (also, allowing warnings) * Updates to contianer validation script. * * Fix to the 'rpath' issue. * Clean-up CMakeLists.txt file * * Edits to Python wheels workflow -- invoke 'eager' tests separately * Run Python tests per directory in Python wheels validation step * Skipping vqe tests in 'Validate Python examples' stage of Python wheels workflow. * * Remove legacy Python kernel builder code. * Experiment with Python wheels CI stage * Debugging arm64 Python wheels * Other clean up * * Skip new test on 'test_exp_pauli' in kernel builder tests. * Updated Python wheels CI stage to skip 'vqe' tests. * [TEMPORARY FIX] Skip CI stages to build Python wheels on arm64 platform. * * Fix for Python tests skipped due to following reason - "AssertionError on 'assert isinstance(arg, _cext.ir.Value)'" * Unmark corresponding tests as skipped. Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fix for 'cswap' test which seemed to hang. * Unmark test as skipped. Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Merged Python tests under 'compiler' and 'mlir' since they were duplicates and removed 'python/tests/compiler/' directory. * Applied code formatting to all changed files * * Clean up and docs fixes by exposing Python APIs * Restructuring 'python/tests' directory - restored 'compiler' directory and fixed 'PYTHONPATH' in the run command. * * Addressing review comment - retain the 'python' subdirectory in 'build' directory (No changes in install path). * * Fix for LLVM static initialization problem arising due to linking of 'LLVMSupport' library in 'mqpu' (Additional work required to use mqpu from Python, but bindings are working again). Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Updated tests to match the qubit reordering work. * Added the 'remote' tests to Python wheels CI check. * * Attempt to get working docker image minus the remote sim functionality in Python * Still skipping Python wheel on arm64 platform * * Previous attempt to only skip validation of Python wheel on arm64 didn't work; hence, skipping the Python wheel stage(s) altogether on arm64 platform. * * Install target for 'cudaq-remote-simulator-qpu' * * Addressing review comments in PR#1077 * Removed logic for 'override_rest_qpu' since it was not required. Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fix to PyRemoteRESTQPU.cpp -- removed 'py_remote_rest' * * Fix for 'test_observeAsync_QNN' Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fix for 'exp_pauli' test. Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Removed a bad test in test_observe_multi_param Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fixes as per Thien's suggestions - restored RestRemoteServer.cpp' * Changes to RemoteSimulatorQPU * * Fix for test_combine_sweep() Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fix for test_QuakeValueDivOp Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Docs fixes * * Implementation for cudaq.parallel * Enabled corresponding tests, previously skipped Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Implementation of cudaq.vqe * Enabled corresponding tests Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Fix for cudaq.chemistry Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> * * Spellcheck fixes * * Removing commented out lines of code from 'double_excitation_gate_opt()' to satisfy spellchecker requirements. --------- Co-authored-by: Alex McCaskey <amccaskey@nvidia.com>

Co-authored-by: Thien Nguyen <thiennguyen@nvidia.com>

Enabling remote qpu w/o multi-qpu

[experimental/python] Align with mainline

…VIDIA#1171) * Improve handling of quantum library calls in kernel body Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * add the extra ast test Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * spell / format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * try and fix issues with CI finding the right python package, fix bug with same kernel name Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * fixes for build errors, put back the ctest target Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Add python path to pycudaq tests Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * wip Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Unify the ast/eager tests. Fixed bug with lambda being popped off the value stack to early Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * ci Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * rename test files, remove print stmt Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix bug in py_alt_launch_kernel where MLIR Context was stored beyond its lifetime Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * address pr comments Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * pr feedback Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix issue with observe_async with eager mode kernels Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Implement sample_async for eager mode Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…VIDIA#1197) * Fix mismatch between new[] and delete (NVIDIA#1180) * [rt] Minor clean up of `BasicExecutionManager` (NVIDIA#1176) * Add SECURITY.md (NVIDIA#1179) * Build and install guide update (NVIDIA#1188) * MPI related docs and commit update (NVIDIA#1191) * Look for a different port if previous is unresponsive. (NVIDIA#1190) Fix for issue# 1117 * Fix tensornet bug with async python functions (NVIDIA#1177) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> Co-authored-by: Ben Howe <141149032+bmhowe23@users.noreply.github.com> Co-authored-by: Bruno Schmitt <bschmitt@nvidia.com> Co-authored-by: Bettina Heim <heimb@outlook.com> Co-authored-by: Alex McCaskey <amccaskey@nvidia.com>

…ernels (NVIDIA#1196) * Build kernel call graph and add all dependent kernels Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * fix spelling Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix issue where inner kernel function was being searched for Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

* * Making JIT execution classes thread safe Co-authored-by: Ben Howe <bhowe@nvidia.com> * * Enable 'test_multi_qpus' --------- Co-authored-by: Ben Howe <bhowe@nvidia.com>

* Expose 'OptimizationResult' in Python API

Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…IA#1200)

…lueStack not cleared (NVIDIA#1204) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

* * Additional backend tests on Quantinuum backend for new Python mode (i.e. 'ast' mode; 'eager' mode is not supported) * Rename existing test with 'builder' suffix * * Re-enable 'observe_async' tests since it has been fixed now.

… defined after that kernel (NVIDIA#1205) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…uda-quantum into align-with-main

[experimental/python] Align with mainline

* enable state, get_state, and get_state_async Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix the docs Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

Co-authored-by: Alex McCaskey <amccaskey@nvidia.com>

…AArch64 for the branch (NVIDIA#1216) * Re-enable the arm64 jobs. * Modify script to build correct wheels. * Update .github/workflows/ci.yml * Update .github/workflows/ci.yml * Update .github/workflows/create_cache_command.yml --------- Co-authored-by: Bettina Heim <heimb@outlook.com>

* * Clean-up : drop 'qlist' * Re-enable working test(s) * Applied code formatting * * Addressing review comment - when passing as function argument, use the non-owing container 'qview'

[experimental/python] Align with mainline

* Fix bug in logic for decrementing range Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * better error message for invalid binary operation Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

* * Make 'MLIR' mode the default , i.e. JIT=True * * No need to explicitly enable JIT in examples since it is the default now.

[experimental/python] Align with mainline

* Fixes to support Python v3.8 * Adding tests * Removed 'id' check since 'Subscript' object has no attribute 'id'

* Added tests for 'sample_async' and 'observe_async' * Consolidate 'remote' tests since using pytest fixture with 'session' scope (Also avoid code duplication) * Rename kernel for avoiding name clash (caching issue) * Uniquely named kernels for testing purposes.

… return types (NVIDIA#1253) * Implement list construction from list comprehension. Implement primitive kernel return types. Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * spell check Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * spell check Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * wip, got uccsd compiling Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * remove library link Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Improved handling of libraries Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * add list() cast support Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * rebased Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * clang format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * python format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

) * Fixes for failing tests

…VIDIA#1274) * Improve error diagnostics. Fix range incrementing bug Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * clean up Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * spell fixes Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix py38 issues Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * apply ngc link fix Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * somehow the test I had didn't get committed Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * pr comments Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…ode (NVIDIA#1279) * Enabling "cx", "cy"..., and "crx", "cr1"... etc. in MLIR mode * Added tests * Addressing review comment - verify the operands to the control operations (match builder behavior)

…A#1256) * Step#1 in eliminating library mode AKA eager mode from Python workflow * Step#2: Updating existing tests ('jit' is always 'True') * Step#3: Adding more tests * Step#4: Clean-up -- 'globalImportedKernels' no longer used / needed. -- Eager mode QIS clean-up

[experimental/python] Align with 'main'

…#1287) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…validation (NVIDIA#1276) Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

…al/python (NVIDIA#1290) The experimental/python branch has an additional JIT point. Apply the same optimization from NVIDIA#1261 to that spot.

Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

* Move AST-to-MLIR compilation to the call operator. Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * python format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * forgot to compile for get_qir Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * fix bug with validation of observe Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

…VIDIA#1293) * Fix bug where args were not validated and processed for sample_async Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * add a test Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * code format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

…uantum into python_docs

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

* Better error checking for erroneous broadcast arguments Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Spellcheck and formatting * Make the new test work on Python 3.8 by using `List` instead of `list` --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> Co-authored-by: Pradnya Khalate <pkhalate@nvidia.com>

Remove 'enable_jit()' call

* implement float augassign Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Fix issue with print_i64 and print_f64 in same kernel Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * clang format Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * update with aug-assign mult for uccsd for test Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> Co-authored-by: Pradnya Khalate <148914294+khalatepradnya@users.noreply.github.com>

'list[float]' --> 'List[float]'

…builder (NVIDIA#1301) * Implement improved error diagnostics for kernel_decorator and kernel_builder Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * Disallow capture variables Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * fix spelling Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com>

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

Description What this PR does do: Refactors the entire docs index structure "Tutorials" -> "Applications" New sections that contain existing docs: Backends Installation Integration Extending Creates a new Quick Start Guide and Basics section Refactors the cuda-quantum/docs folder to be more similar to index structure Reproduces examples in C++ and Python, and moves them to alternate tabs What this PR doesn't do: Add any documentation for new python support Make any large modifications to examples/applications content Make any large modifications to existing content in the bottom half of the index structure * first pass of docs changes for quick start and section renames Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pr comments and more docs reshuffling Signed-off-by: A.M. Santana <anthonys@nvidia.com> * heavily refactor docs folder and reorganize new sections Signed-off-by: A.M. Santana <anthonys@nvidia.com> * store changes Signed-off-by: A.M. Santana <anthonys@nvidia.com> * more basics section Signed-off-by: A.M. Santana <anthonys@nvidia.com> * create separate backends section Signed-off-by: A.M. Santana <anthonys@nvidia.com> * first pass of pr comments Signed-off-by: A.M. Santana <anthonys@nvidia.com> * move new code snippets to their own files Signed-off-by: A.M. Santana <anthonys@nvidia.com> * more cleanup Signed-off-by: A.M. Santana <anthonys@nvidia.com> * changes from discussion Signed-off-by: A.M. Santana <anthonys@nvidia.com> * fix install guide index and clang format Signed-off-by: A.M. Santana <anthonys@nvidia.com> * update default ops Signed-off-by: A.M. Santana <anthonys@nvidia.com> * big pass of updates to python and basics Signed-off-by: A.M. Santana <anthonys@nvidia.com> * overhaul of examples section Signed-off-by: A.M. Santana <anthonys@nvidia.com> * finish observe section of basics Signed-off-by: A.M. Santana <anthonys@nvidia.com> * clean up and ci fixes Signed-off-by: A.M. Santana <anthonys@nvidia.com> * more spelling and formatting fixes Signed-off-by: A.M. Santana <anthonys@nvidia.com> * more spellcheck fixes Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pay the spellchecker toll Signed-off-by: A.M. Santana <anthonys@nvidia.com> * remove broken cpp file Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pay the formatter toll Signed-off-by: A.M. Santana <anthonys@nvidia.com> * first pass of pr comments Signed-off-by: A.M. Santana <anthonys@nvidia.com> * fix merge conflicts Signed-off-by: A.M. Santana <anthonys@nvidia.com> * fix issue with versions not generating Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pr comments Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pr comments Signed-off-by: A.M. Santana <anthonys@nvidia.com> * backend links Signed-off-by: A.M. Santana <anthonys@nvidia.com> * pr comments Signed-off-by: A.M. Santana <anthonys@nvidia.com> * remove nvcf backend from list Signed-off-by: A.M. Santana <anthonys@nvidia.com> --------- Signed-off-by: A.M. Santana <anthonys@nvidia.com>

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

* Remove exact version specifier in Python install instructions (NVIDIA#1373) Anaconda doesn't have linux-aarch64 packages for Python 3.8.0 and 3.9.0, so without this change, these instructions will fail on ARM. However, when switching to a single "=", conda will allow the user to specify a version up to the number of digits provided, and then use the latest version available after that. For example, if the user provides "3.8", then conda will choose "3.8.18", which is the largest patch version matching "3.8". * Python readme update (NVIDIA#1378) * Fix a bug in sample.h affecting sample_async in library mode (NVIDIA#1379) In library mode, we need to run a tracing pass before actual execution. With multi-qpu, we set_exec_ctx on specific qpu id but didn't reset the context on that particular qpu id (fall back to reset QPU 0 context all the time). * Fix bug - enable list[T] argument specification with kernel builder (NVIDIA#1370) * Enable list[T] specification with kernel builder. Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * small fixes Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> * * Add docstrings for 'draw' (NVIDIA#1377) * Make nvqc_proxy.py multi-threaded (NVIDIA#1387) Full backstory and test details are in PR. * Fix a few Python notebook examples for release (NVIDIA#1386) * Fix NVIDIA#1348 - synthesis of vectors with small integers. (NVIDIA#1395) The constant array op is using at ArrayAttr, which does not naturally support integral sizes less than i32. This patch upcasts the short values to i32 to get around the issue. Another approach would be to have the constant array op take a DenseArrayAttr, which does support the shorter integral types. Might need a custom parser/pretty printer if we go that route? * [test] Fix deployment issue with NVIDIA#1370 (NVIDIA#1394) * Fix NVIDIA#1383: bridge not lowering to correct IR (NVIDIA#1402) The operator[] returns a reference, but the AST doesn't have a cast node. So the value appearing in the conditional is not loaded. This adds an explicit check and an error message if the load is returning an unexpected value. Fix NVIDIA#1383. * Fix NVIDIA#1392: "syntehsis" of callables (NVIDIA#1398) Refactor the code that was to catch struct value passing (which is not yet implemented) from arguments that are callables. Fix NVIDIA#1392 * Path fix in publishing (NVIDIA#1399) --------- Signed-off-by: Alex McCaskey <amccaskey@nvidia.com> Co-authored-by: Ben Howe <141149032+bmhowe23@users.noreply.github.com> Co-authored-by: Bettina Heim <heimb@outlook.com> Co-authored-by: Thien Nguyen <58006629+1tnguyen@users.noreply.github.com> Co-authored-by: Alex McCaskey <amccaskey@nvidia.com> Co-authored-by: Pradnya Khalate <148914294+khalatepradnya@users.noreply.github.com> Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>

Fix commit history from merge

Signed-off-by: A.M. Santana <anthonys@nvidia.com>

Description Replaces all python builder code in examples and tutorials with MLIR mode Removes python_mlir and python_runtime folders from docs examples Adds notebooks from PR Adding new tutorials NVIDIA#1013

Catch up to upstream

Catch up to main

Adds commit: [refactoring] Remove code duplication for nvcf client classes (NVIDIA#1334)

…da-quantum into features/docs_refactor

Documentation for NVQC #1412

Documentation for NVQC #1412

Commits on Sep 26, 2023

Commits on Nov 8, 2023

Commits on Nov 9, 2023

Commits on Nov 24, 2023

Commits on Nov 28, 2023

Commits on Dec 1, 2023

Commits on Dec 5, 2023

Commits on Dec 6, 2023

Commits on Dec 12, 2023

Commits on Dec 13, 2023

Commits on Dec 27, 2023

Commits on Dec 28, 2023

Commits on Dec 29, 2023

Commits on Jan 17, 2024

Commits on Jan 18, 2024

Commits on Jan 25, 2024

Commits on Feb 3, 2024

Commits on Feb 4, 2024

Commits on Feb 5, 2024

Commits on Feb 6, 2024

Commits on Feb 8, 2024

Commits on Feb 9, 2024

Commits on Feb 10, 2024

Commits on Feb 13, 2024

Commits on Feb 14, 2024

Commits on Feb 15, 2024

Commits on Feb 16, 2024

Commits on Feb 20, 2024

Commits on Feb 21, 2024

Commits on Feb 22, 2024

Commits on Feb 23, 2024

Commits on Feb 24, 2024

Commits on Feb 26, 2024

Commits on Feb 27, 2024

Commits on Mar 4, 2024

Commits on Mar 5, 2024

Commits on Mar 13, 2024

Commits on Mar 14, 2024

Commits on Mar 15, 2024

Commits on Mar 17, 2024

Commits on Mar 18, 2024