Question about allocateUnifiedGPUPage() allocation strategy in device.go

[Taoger-Xu]

To Reproduce
MGPUSim version of commit ID: 0a6e944

Hello,

I have a question regarding the allocateUnifiedGPUPage() method in the device.go file. This function appears to allocate a physical page on a GPU, using devSelected to keep track of the last selected device.

From the code, it looks like a round-robin allocation strategy might be intended. However, I noticed that the method does not break out of the loop upon finding a suitable ActualGPU device with available pages, which results in traversing the entire list of ActualGPUs.

Thank you for your time and any insights you can provide!