I'm building a high powered system that is going to be used as a database server (personal academic/research use, I don't need ECC or server cpus or anything), some tables having as many as 300 million rows (and 20+ GB of data). I need to squeeze out all the performance I can get, so I'm going to be getting 64 GB of ram and loading the entire database into memory.
Any idea whether there would be a large difference in performance (in my case that would be query times, even shaving off fractions of a second per query would make a big difference) between these kits?
They are all G.SKILL Ripjaws Z series, DDR3 1600, 2133, and 2400:
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231508 (1600, CAS 10, 10-10-10-30, $315)
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231525 (2133, CAS 11, 11-11-11-31, $700)
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231523 (2400, CAS 10, 10-12-12-31, $700)
Any idea whether there would be a large difference in performance (in my case that would be query times, even shaving off fractions of a second per query would make a big difference) between these kits?
They are all G.SKILL Ripjaws Z series, DDR3 1600, 2133, and 2400:
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231508 (1600, CAS 10, 10-10-10-30, $315)
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231525 (2133, CAS 11, 11-11-11-31, $700)
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231523 (2400, CAS 10, 10-12-12-31, $700)