If you are concerned about two enterprise applications that affect each other's performance, you should move them to two separate virtual machines.
GC is optimized and runs its own stream (s). It is designed for invisibility for the current application. Therefore, on a multiprocessor corporate server, a separate process should not be damaged at all.
On the other hand, the server is still getting some workload from the GC. If you think that the GC somehow slows down your applications, perhaps you can do some memory and CPU profiling to see where the problem is. You can find a way to optimize the code and use less resources.
From J.Richter "CLR via C # v3" p. 585
This mode finely tunes the garbage collector for the server side of the Application. The garbage collector assumes that no other applications (client or server) are running on the machine, and it is assumed that all processors on the machine are available for garbage collection. This GC mode splits the managed heap into several partitions, one per processor. When garbage collection begins, the garbage collector has one thread per processor; each thread collects its own section in parallel with other threads. Parallel collections work well for server applications in which workflows tend to exhibit uniform behavior. This feature requires the application to run on a computer with multiple processors, so that the threads can actually be run simultaneously to achieve improved performance.
source share