完善mps的vidal机制,多节点并行;补充tn搜索时dask集群搜索的方式

README.md

@@ -12,7 +12,7 @@ Tensor Network Types:
 Tensor Network contractions to:
 
 - dense vectors
-- expecation values of given Pauli string
+- expectation values of given Pauli strings or Pauli-sum observables
 
 The supported HPC configurations are:
 
@@ -26,6 +26,18 @@ Currently, the supported tensor network libraries are:
 - [cuQuantum](https://github.com/NVIDIA/cuQuantum), an NVIDIA SDK of optimized libraries and tools for accelerating quantum computing workflows.
 - [quimb](https://quimb.readthedocs.io/en/latest/), an easy but fast python library for ‘quantum information many-body’ calculations, focusing primarily on tensor networks.
 
+## CPU expectation benchmarks
+
+The current CPU expectation entrypoint is:
+
+```sh
+python -u benchmark_cpu_expectation.py --ansatz mps --nqubits 40 --nlayers 10 --bond 2048 --circuits brickwall_cnot --observables ring_xz
+```
+
+Use `--ansatz tn` for the generic TN path and `--mpi` under `mpiexec` for MPI runs.
+Reusable circuit and observable builders live in `src/qibotn/benchmark_cases.py`; execution logic lives in `src/qibotn/expectation_runner.py`.
+For Vidal/MPS 1D-chain scale tests, use `run_vidal_mps_cases.sh`.
+
 ## Installation
 
 To get started: