1

The core process of the company is running on a 4 year old server, and at this point the I.T. Staff got the budget to buy a new server. The core process is a web app with intensive writing/reading from a poorly structured MySQL database, that's why we need a good match between a powerful processor and an excellent drive to keep the process running until the database is remodeled.

We got two offers one from a DELL retailer and another from HP.

  1. DELL PowerEdge T630 Server ($5196 USD):

Intel Xeon E5-2650 v4 x1 2.2GHz 30M Cache (12 Cores) - RAM 16x2 GB RDIMM - 800GB Solid State Drive SATA Mix Use MLC 6Gbps 2.5in Hot-plug Drive

  1. HPE ProLiant DL180 Gen9 ($4692 USD):

Intel Xeon E5-2620 v4 x2 2.1GHz 20M Cache (8 Cores Each) - RAM 16x2 GB RDIMM - 240GB Solid State Drive SATA RI-2 LFF SCC SSD - 1TB HP HDD 6G SATA 3,5" 7,2K SC MDL

As you can see, the HP brings two processors, but it is relatively low in price, which gives me confusion.

Taking into account that the fact that this database is poorly structured, implies that it is necessary to resolve queries as fast as possible (speed per core) and also many queries simultaneously (number of cores); which one would you recommend?. Thanks in advance.

  • 1
    I would recommend benchmarking your current server to see whether the limiting factor is CPU power or disk-access speed. A poorly-structured database could be having problems with either. – Mark Dec 7 '16 at 21:14
  • Going by what is listed the lower price on the dual CPU HP likely comes from the CPU having lower on die cache as well as a slightly lower MHZ rating, as well as a lower capacity SSD. – NZKshatriya Dec 7 '16 at 22:58
  • So, the dual CPU has a total of 32 combined processing threads, and a combined 40M of on die cache, while the single CPU has 24 processing threads, and 30M of on die cache. From a pure hardware perspective the dual CPU setup sounds better to me, but I am not too familiar with server operations, so I am leaving this as a comment, not an answer. – NZKshatriya Dec 8 '16 at 5:02
  • 1
    @ElberCM Well there is always Passmark, and PCmark8, those are benches I am familiar wish. – NZKshatriya Dec 12 '16 at 18:16
  • 2
    @ElberCM, the best benchmark is the one that is most similar to the task you intend to perform. Ideally, you'd run a simulated load on your actual server software while monitoring CPU load, I/O bandwidth, memory usage, and any other parameters you think might be significant. – Mark Dec 12 '16 at 20:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.