If it is a laptop manufactured in 2022, I would not expect any issues running VAMPIRE for a simulation needing a small amount of computing resources. It is hard to say precisely what system requirements are needed as it can vary from one simulation to another as it depends on the setup of the simulation. For example, more atoms, larger geometries, smaller cell size, and higher number of
temperature-increment to name a few can require more computing resources (e.g., RAM and hard drive or SSD space).
Under "System Requirements" in section "2 Installation" of the VAMPIRE 5.0 manual [1], it states that the more memory (RAM and free storage space), faster the clock speed, and higher the number of processor cores the better for being able to perform VAMPIRE simulations.
Usually, it is best to use VAMPIRE in Linux or Mac OSX, where either of those operating systems are directly installed on the hardware of the system (e.g., laptop).
Ubuntu seems to be a fairly popular distribution for Linux. It has minimum system requirements such as 2 GHz dual core processor, 4GiB RAM, and 25 GB hard drive space according to [2]. It can be a good choice as the
VAMPIRE
download page has a Ubuntu binary package for it [3]. Though, there are other Linux distributions
[4] you could chose to try to use.
For Windows, you could try using VAMPIRE in Windows Subsystem for Linux (WSL) [5] or use Linux in a virtual machine such as VirtualBox or QEMU [6]. Those environment I believe would be okay for the learning how to use VAMPIRE. In other words, many of the tutorials [7] and workshop examples [8] would likely run using that. It would also likely make a good sandbox for playing around with VAMPIRE or for preparing input files. However, with Linux running alongside or on top of Windows there will be less free RAM and other computing resources for performing a VAMPIRE simulation compared to having the computer system run only Linux on it. Thus, when not playing around and doing a serious simulation it typically is better to do it on a system where Linux is install directly on the hardware.
Whether you would need to use a GPU or not would again depend on the setup of your simulation. For example, simulating a magnetic single layer might not need it, but a multilayer [9] simulation on the other hand might need it. It is known that some micromagnetic simulations can be
speed-up using a GPU on the order of 100x [10] compared to that of a CPU. According to the VAMPIRE 5.0 release notes [11], doing GPU simulations with CUDA are beta, such that simulations using that might not be entirely stable. CUDA is for NVIDIA [12], such that your laptop would likely need to have a NVIDIA graphics card for that to work.
Parallel simulations using mpi can be used to speed up calculations but
non-orthogonal cells are not supported yet [13,14].
To get the most benefit from mpi, I would try to use a high performance computing (HPC) cluster [15-17].
When I was a graduate research assistant, my personal laptop wasn't good enough for some of my simulations. However, I worked with my advising professor at the university, and they were able to get me login access to either an on-campus or off-campus HPC. Next, I was able to remotely connect my laptop to the HPC, then the HPC allowed me to perform my simulations.
Kind Regards,
Gavin
VAMPIRE user