遇见数据集

Datasets for the aTTN Cookbook

收藏
Zenodo2025-07-30 更新2026-05-26 收录
官方服务:

资源简介:

The files contain the datasets with the results of the manuscript The Augmented Tree Tensor Network Cookbook. All the datasets have been produced using the Quantum TEA software, with the user guide repository for running the simulations linked in the URL. The dataset contains 7 folders: raw_data_ising4numDE/ - square lattice Ising model benchmarks, energy and runtime for different number of disentanglers (DE)Subfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/gpu_symm5_sweeps{number-of-dmrg-sweeps}_numde{number-of-DE}/out/ raw_data_ising4extfields/ - square lattice Ising model benchmarks, energy and runtime for different external fieldsSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ raw_data_ising4bonddims/ - square lattice Ising model benchmarks, energy and runtime for different system sizes and bond dimensionsSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/cpu+gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ raw_data_ising_memories/ - square lattice Ising model benchmarks, max allocated GPU memory for different system sizes and bond dimensionsSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/cpu+gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ raw_data_ising_bestresult/ - square lattice Ising model benchmarks, energy and runtime for highest possible bond dimension over different external fieldsSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/cpu+gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ raw_data_ising4deoptsweeps/ - square lattice Ising model benchmarks, energy and runtime for different number of sweeps without disentanglersSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_g{external-field}/gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ raw_data_heis4bonddims/ - triangular lattice Heisenberg model benchmarks, energy, runtime, and max allocated GPU memory for different system sizes and bond dimensionsSubfolder structure: {TN-ansatz}/Lx{size-along-x}_Ly{size-along-y}_chi{bond-dimension}_D1.0000/cpu+gpu_symm5_sweeps{number-of-dmrg-sweeps}_deopt{sweeps-without-DE}/out/ Remark: symm5 in all folder names marks that the grond state search was ran without enforced symmetries, i.e. conserved quantities., gpu/cpu+gpu marks whether the simulation was run fully on a GPU or with the mixed device mode

提供机构:
Zenodo
创建时间:
2025-07-30
二维码
社区交流群
二维码
科研交流群
商业服务