Release Notes

v1.3.1 (2023/04/17)

mpiQulacs v1.3.1 Release
- Added error handling when insufficient memory is encountered during sampling process
- Supported QuantumState.get_entropy function with multi-cpu
- Improved performance of Swap Gate processing

Update of system usage with system configuration change
- Number of compute nodes changed from 64 to 512
- Introduction of the Job Scheduler Slurm
- Change the login destination to the login server instead of the representative compute node
- Some IP addresses/host names of various servers, such as a jump server, were changed.
- Add Job Scheduler Slurm Description
- User account name at login, HostName on jump server, port number changed
- Job execution method, Jupyter, and debugging (Visual Studio Code) steps changed
- Adding v2.1.1 to PySCF in Python Package
mpiQulacs v1.3.0 Release
- Capture Correction in Qulacs v1.3.1
  - Corrects memory leak failures
- The QuantumCircuitOptimizer function now supports optimization on a gate-by-gate basis and optimization with FusedSWAP at once.
- Improved state vector update processing performance for some gates.

Changed the contents of job.sh、job_debug.sh、ipcluster_config.py. Update the contents of each file you are using.
- job.sh: We added the LD_PRELOAD environment variable only when using qiskit-qulacs, but changed it to always set the LD_PRELOAD environment variable whether or not you use qiskit-qulacs.
- job_debug.sh: This file is used for debugging with VSCode. Added setting of environment variables (OMP_NUM_THREADS, QULACS_NUM_THREADS) to control the number of threads and setting of the LD_PRELOAD environment variable.
- ipcluster_config.py: Configuration file for Jupyter. Added the LD_PRELOAD environment variable setting to reflect the changes made to job.sh.
Added version numbers available in the Qiskit applications library when using qiskit-qulacs.
- If you use Qiskit Nature, Machine Learning, Optimization, Finance, install the version described on the Installation page.
Changed the way you log in to the document server to use ssh config.

mpiQulacs v1.2.2 Release
- Fixed a bug that could cause Observable.get_expectation_value(state) to return an incorrect value.
  - Occurrence condition: Can occur if an operator contains one or more pauli-X/Y and acts only on pauli-Z for global_qubit.

mpiQulacs v1.2.1 Release
- Corrects a bug where a QuantumState.load(vector) on a distributed vector does not load correctly for processes with rank numbers greater than 1.
Documentation Updates
- Added a description of the QuantumState load method’s behavior during state vector distribution.
- Added a way to load/get an entire distributed state vector (example code).

Fix class method support on mpiQulacs API documentation
- RandomUnitary supported, get_transition_amplitude not supported
Added supplementary information about mpiexec execution parameters and how to determine the value of the hostfile slot.
Add the following description to how Jupyter is run
- How to start Jupyter outside node01
- How to use JupyterLab
- How to execute ssh without ssh config
- How to set environment variables
The name of the configuration file when using Jupyter is added to the installation method of PySCF.
Change how to deal with cannot allocate memory in static TLS block in Troubleshooting

mpiQulacs v1.1.0 Release
- Support Observable, PauliOperator
- Change update_quantum_state(…) argument to allow random seed
  - In probabilistic gates such as Measurement, random numbers are used to achieve probabilistic effects. You can specify the seed of the random number.
qiskit-qulacs v1.1.0 Release
- Update version of mpiQulacs used from v1.0.0 to v1.1.0
Add instructions for installing PySCF
Add OpenMP threading control method to job.sh sample code

Initial release