It seems from this piece of literature that it can, as you suspected. This is apparently one of the major criticisms of FR-FCFS, that it is unfair. It achieves high throughput, but more complicated design structures sacrifice some throughput to balance out the fairness.
ferozenaina
For a given problem, the programmer can typically change the code, data accesses and structures to ensure locality in cache which gives a large speed boost.
I had asked Prof how can the programmer influence the way memory controller arranges data in physical memory to do something similar. He said that if programmer implements good data locality in cache, it will be retained in the physical memory by the memory controller.
Could FR-FCFS ever lead to starvation? Where we keep servicing the open row, never making it to the rest?
(https://books.google.com/books?id=yJNHICYATYcC&pg=PA176&lpg=PA176&dq=fr+fcfs+starvation&source=bl&ots=Bs8iDumRQn&sig=IQG6XPy_dQLyDRlSatpVtSA-8Nc&hl=en&sa=X&ved=0ahUKEwil1PPP44HQAhUBSSYKHUHdAqgQ6AEISTAH#v=onepage&q=fr fcfs starvation&f=false)
It seems from this piece of literature that it can, as you suspected. This is apparently one of the major criticisms of FR-FCFS, that it is unfair. It achieves high throughput, but more complicated design structures sacrifice some throughput to balance out the fairness.
For a given problem, the programmer can typically change the code, data accesses and structures to ensure locality in cache which gives a large speed boost.
I had asked Prof how can the programmer influence the way memory controller arranges data in physical memory to do something similar. He said that if programmer implements good data locality in cache, it will be retained in the physical memory by the memory controller.