I've been attempting to benchmark Drizzle on a Cisco C250 (6 core/12 thread/24 CPU; 346G RAM, FusionIO 320GB MLC card) but Drizzle's InnoDB is old and has a bug (https://bugs.launchpad.net/drizzle/+bug/911948) which causes sysbench to crash it. I've spoken with others about updating InnoDB. Stewart said he had scripts to do this. Once this is fixed, I'll put belief to the test. :-)
Le 4 janv. 2012 à 11:04, Mark Atwood a écrit : > Hi Vijay, > > We (the Drizzle team) don't do Drizzle vs MySQL benchmarks. > > We're of the opinion that benchmarks should be done by 3rd parties, > and if possible, should be the customer's own load on their own > machines. > > That said, we believe that Drizzle is significantly faster than MySQL > when under heavy concurrent load on multicore machines. Often MySQL > is faster, especially when the load is not concurrent and the > underlying machine has fewer cores. > > ..m > > On Wed, Jan 4, 2012 at 9:02 AM, Vijay Samuel <[email protected]> wrote: >> Hi Mark, >> Im going to give a talk about Drizzle next week. I have the slides >> regarding the features of Drizzle ready. I need a few slides containing >> numbers ;) Im going to mention testing using sysbench, sqlbench, >> drizzleslap, randgen and dbt2. I looked at the benchmarks mailing list and >> was able to get some info. Do you have any numbers on MySQL vs Drizzle? If >> im convincing then I can bring Drizzle into the eBay architecture in some >> form or the other. >> >> Cheers, >> -Vijay _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
