|
|
My view is that the workstation configurations above are not at all well suited for your desired application.
Projects that require lots of calculations in Gaussian and ORCA benefit greatly from CPUs with high core count, as computations can be parallelised quite efficiently. 24 cores (albeit fast ones) will be quite limiting in this regard.
The memory configuration is also not sufficient. 128 GB will be very limiting for this type of work, especially if you require analytical Hessian calculations.
Gaussian and ORCA do not make use of GPU hardware accelerators, which are furthermore quite expensive. The RTX 4000 Ada is not particularly powerful, not useful for this application, and a very expensive component. These professional workstation cards command a high price, and the benefits they confer over consumer cards are quite niche (e.g. MIG, certified drivers, higher VRAM, and capability with certain floating point formats). |
|
发表于 Post on 2025-8-6 18:36:53
|
只看该作者 Only view this author
|只看大图 View big picture
|正序浏览
|阅读模式 Reading model
收藏 Add to favorites
能给这样配置的采购也是心够黑的，直接拉黑就行，摆明想要狠狠坑一笔钱。你们老师也够心大的，对硬件不了解的情况下就开始想着买机器，经费这么宽裕？
楼主 Author