I've read through several posts and am cognizant of the RAM/cpu core guidelines on performance.
With that being said, is there a limit on the priority that should be placed on RAM/core when specifying out a system?
For instance, consider the following 2 systems:
System 1) 3 gHz 4-core cpu, 128 Gb Ram 8 dimms 2133 Hz
System 2) 3 gHz 8-core cpu, 128 Gb Ram 8 dimms 2133 Hz
Both have the same amount of same speed memory. Only difference is that system 1 has less processor cores to share the memory.
Would system 1 be capable of handling larger models due to more RAM/core?
If I'm using iterative solvers, I'm guessing they would be very close to being equivalent.
I mostly model fluid acoustic fields (with occasional structural mechanics), if that is pertinent information.
I would appreciate anyone's opinion on the matter.
Thanks,
Seth
With that being said, is there a limit on the priority that should be placed on RAM/core when specifying out a system?
For instance, consider the following 2 systems:
System 1) 3 gHz 4-core cpu, 128 Gb Ram 8 dimms 2133 Hz
System 2) 3 gHz 8-core cpu, 128 Gb Ram 8 dimms 2133 Hz
Both have the same amount of same speed memory. Only difference is that system 1 has less processor cores to share the memory.
Would system 1 be capable of handling larger models due to more RAM/core?
If I'm using iterative solvers, I'm guessing they would be very close to being equivalent.
I mostly model fluid acoustic fields (with occasional structural mechanics), if that is pertinent information.
I would appreciate anyone's opinion on the matter.
Thanks,
Seth