Greets:
So, I'm building a new machine that is designed specifically to store and crunch large amounts of data for statistically analysis. My current setup is really chugging due to the memory limitations where "out of memory" among other issues are cropping up.
Now, I need something with more memory capacity that can really store a good bulk of data in memory while it's being processed. Thus, the idea of building a machine with 64 GB or even 32 GB of memory came to mind.
With a machine with that much memory, someone mentioned to me that going dual socket would help with that much memory. But, from my understanding, that only means processing the data would be faster and not that my memory capacity would be limited right?
I plan on purchasing a fast processor, but the speed in which the data is processed is not a critical factor to me. The real problem I'm trying to solve is how much data I can load into the memory before it's processed. 8 to 16 GB is just not cutting it anymore.
Any insight would be great, especially on building such a memory intensive machine.
So, I'm building a new machine that is designed specifically to store and crunch large amounts of data for statistically analysis. My current setup is really chugging due to the memory limitations where "out of memory" among other issues are cropping up.
Now, I need something with more memory capacity that can really store a good bulk of data in memory while it's being processed. Thus, the idea of building a machine with 64 GB or even 32 GB of memory came to mind.
With a machine with that much memory, someone mentioned to me that going dual socket would help with that much memory. But, from my understanding, that only means processing the data would be faster and not that my memory capacity would be limited right?
I plan on purchasing a fast processor, but the speed in which the data is processed is not a critical factor to me. The real problem I'm trying to solve is how much data I can load into the memory before it's processed. 8 to 16 GB is just not cutting it anymore.
Any insight would be great, especially on building such a memory intensive machine.