Re: [PERFORM] Benchmarking postgres on Solaris/Linux

2004-03-24 Thread Mark Kirkwood
Josh Berkus wrote: Mark, It might be worth considering Apple if you want a 64-bit chip that has a clock speed comparable to Intel's - the Xserv is similarly priced to Sun V210 (both dual cpu 1U's). Personally I'd stay *far* away from the XServs until Apple learns to build some real ser

Re: [PERFORM] Benchmarking postgres on Solaris/Linux

2004-03-23 Thread Josh Berkus
Mark, > It might be worth considering Apple if you want a 64-bit chip that has a > clock speed comparable to Intel's - the Xserv is similarly priced to Sun > V210 (both dual cpu 1U's). Personally I'd stay *far* away from the XServs until Apple learns to build some real server harware.The cur

Re: [PERFORM] Benchmarking postgres on Solaris/Linux

2004-03-22 Thread Mark Kirkwood
The hardware platform to deploy onto may well influence your choice : Intel is usually the most cost effective , which means using Linux makes sense in that case (anybody measured Pg performance on Solaris/Intel?). If however, you are going to run a very "big in some sense" database, then 6

Re: [PERFORM] Benchmarking postgres on Solaris/Linux

2004-03-22 Thread Josh Berkus
Stalin, > As anyone done benchmarking tests with postgres running on solaris and linux > (redhat) assuming both environment has similar hardware, memory, processing > speed etc. By reading few posts here, i can see linux would outperform > solaris cause linux being very good at kernel caching than