构建大的超晶胞模拟，磁矩随温度的演化是离散的，没有演化的规律。

[pengxb1228]

您好，老师:
首先很感谢您写了这么好用的软件模拟居里温度。在使用mcsolver的过程中，我遇到了一个新问题。当构建30x30x1超晶胞的时候，模拟出来的结果是有规律的，能获得想要的距离温度。然而，同样的参数，当采用150x150x1这么大的超晶胞的时候，模拟出来的居里温度是离散。 不知道哪里设置有问题？麻烦老师有时间帮忙看看。祝老师身体健康，一切顺利！
以下是我的输入文件
This is mcsolver's save file, version: 3.0
Lattice:
1.000000000 0.000000000 0.000000000
0.000000000 2.000000000 0.000000000
0.000000000 0.000000000 1.000000000
Supercell used in MC simulations:
30 30 1
Orbitals in cell:
2
Positions, initial spin states and onsite-anisotropy of every orbital:
orb 0: type 0 spin 2.000000000 pos [0.500000000 0.166666000 0.000000000] Dx 0.000000000 Dy 0.000000000 Dz -0.012000000 h 0.000000000
orb 1: type 0 spin 2.000000000 pos [0.000000000 0.666666667 0.000000000] Dx 0.000000000 Dy 0.000000000 Dz -0.012000000 h 0.000000000
Bonds:
12
id, source, target, overLat, exchange matrix elements of each bond:
bond 0: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 1 to orb 1 over [1 0 0]
bond 1: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 0 over [1 0 0]
bond 2: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 1 to orb 0 over [0 1 0]
bond 3: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [1 0 0]
bond 4: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [0 0 0]
bond 5: Jx -159.210000000 Jy -159.210000000 Jz -160.620000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 1 to orb 0 over [-1 1 0]
bond 6: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 0 over [0 1 0]
bond 7: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [2 -1 0]
bond 8: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [-1 0 0]
bond 9: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [-1 -1 0]
bond 10: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 0 to orb 1 over [2 0 0]
bond 11: Jx 10.140000000 Jy 10.140000000 Jz 10.610000000 Jxy 0.000000000 Jxz 0.000000000 Jyz 0.000000000 Jyx 0.000000000 Jzx 0.000000000 Jzy 0.000000000 orb 1 to orb 1 over [0 2 0]
Temperature scanning region:
Tmin 400.000000000 Tmax 800.000000000 nT 80
Field scanning region (in unit 1.48872 T, only if Kelvin and uB is used for energy and spin):
Hmin 0.000000000 Hmax 0.100000000 nH 1
Dipole long-range coupling:
alpha 0.000000
Measurement:
measure the correlation function between orb0 and orb0 over [0 0 0]
Supergroup
OrbGroup:1
Supergroup
group0 orb0-orb0

  Topological section      <<<

LocalCircuit per cell: 0 (set to 0 to skip the calc. for topo. Q)

End of Topological section <<<
Distribution output frame: 0
Sweeps for thermalization and statistics, and relaxiation step for each sweep:
80000 640000 0
XAxis type:
T
Model type:
Heisenberg
Algorithm:
Metropolis
Ncores:
40

[golddoushi]

您好，目前发布的软件吃不下150x150x1的模型，建议尺寸设置在100x100x1以下也就是总共磁矩不要过万。支持大模型的代码已经基本写好，还需测试一小段时间就能发布了

[pengxb1228]

谢谢！非常期待老师您的mcsolver支持大模型。 2024-04-25 08:43:53 "Liang Liu" ***@***.***> 写道： 您好，目前发布的软件吃不下150x150x1的模型，建议尺寸设置在100x100x1以下也就是总共磁矩不要过万。支持大模型的代码已经基本写好，还需测试一小段时间就能发布了 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>