Linked by Thom Holwerda on Fri 31st Oct 2008 16:49 UTC, submitted by Michael
Earlier this week Phoronix published a set of comparative benchmarks regarding Ubuntu, testing various recent releases to compare their performance figures on all sorts of different tasks. They've now done the same thing for Fedora, so we can compare Fedora and Ubuntu in terms of performance.
Thread beginning with comment 335980
To read all comments associated with this story, please click here.
To read all comments associated with this story, please click here.
Member since:2005-07-06
Instead of recompiling the kernel, a better baseline for the lowest common denominator would be to run the 64-bit version of the OS (most new machines sold in the last 2-3 years are 64-bit capable). That way, you don't get both the kernel and binaries being dumbed down to an i386 set of opcodes, plus you may see some performance benefit over the 32-bit version. Yes, it'll use more RAM, but that's one thing that's very cheap nowadays.
Recent Original Stories
Recent Comments
Headlines
Random Comments
Random Stories
Random OS Link
Member since:
2006-08-18
I think the Linux kernel and the method Ubuntu and other distributions use to compile it has a big role in the gradual degrade of performance.
My guess is distribution kernels select nearly all the options in Linux's makeconfig utility to provide common denominator support for most generic hardware and select basic performance optimizations for a broad-range of x86-based CPUs.
The situation will only worsen with time; therefore performance conscience users may consider re-compiling the kernel with only the necessary drivers and specific optimizations for their CPU type.