| 查看: 827 | 回复: 0 | ||
| 【悬赏金币】回答本帖问题，作者lilliankung将赠送您 10 个金币 | ||
lilliankung银虫 (小有名气)
|
[求助]
几何优化后的结构计算能带报错
|
|
|
结构是纳米线，共68原子，几何优化成功以后，task选择energy，properties勾选能带和态密度,计算时报错（）。内存是32G，按理说没有超过所需内存容量，为什么还报错呀？有没有好心人来讨论一下 +---------------- MEMORY AND SCRATCH DISK ESTIMATES PER PROCESS --------------+ | Memory Disk | | ----------------------------- | | Approx. total storage required per process 3685.5 MB 868.7 MB | | | | Requirements will fluctuate during execution and may exceed these estimates | +-----------------------------------------------------------------------------+ ===================================================================== + + + Calculation of Spectral Properties + + + + Calculation re-parallelised over 18 processes. + + Data distributed by G-vector(3-way), k-point(6-way), band(1-way) + + Data distributed by G-vector(3-way), k-point(6-way) + + + ===================================================================== Starting k-point: 1 of 1 on this process |<-- SPEC ************************************************************************ The memory required for this calculation exceeds machine capacity. Unable to continue. Run will be ABORTED. Suggestions: Check the run setup is reasonable; in particular cell dimensions, number of atoms, plane wave cutoff all affect memory consumption. Investigate the parameters "opt_strategy"/"opt_strategy_bias" and "page_wvfns" which may be used to reduce memory requirements. If the calculation is being run on a distributed memory computer consider increasing the number of processors. It may be necessary to run this calculation on another computer with more RAM or upgrade the physical RAM on the current system. ************************************************************************ Information for developers: ALLOCATE request failed for array "tmp_data1" in routine "local_dot_all_many_many" Current trace stack: local_dot_all_many_many coeffs_dot_all_many_many wave_dot_all_slice_slice wave_diagonalise_H_slice hamiltonian_diagonalise_slice spectral_calculate_dft castep ************************************************************************ |
回复此楼» 猜你喜欢
拟解决的关键科学问题还要不要写
已经有9人回复
-
为什么一个当量的苯甲醛反应不完呢
已经有5人回复
-
招博士
已经有3人回复
-
存款400万可以在学校里躺平吗
已经有35人回复
-
为什么nbs上溴 没有产物点出现呢
已经有8人回复
-
最失望的一年
已经有18人回复
-
求推荐博导
已经有4人回复
-
求助一下有机合成大神
已经有4人回复
-
求推荐英文EI期刊
已经有5人回复
-
基金委咋了？2026年的指南还没有出来？
已经有10人回复
» 本主题相关商家推荐: (我也要在这里推广)
