FreeBSD 9.0-RELEASE on PowerMac Dual G5

[Nathan Whitehorn]

Not sure where the the intermediate emails on this thread went, so I'll 
jump in again.

On 02/21/12 03:50, Super Bisquit wrote:
> .
> On 2/19/12, Kevin H. Patterson<kpatterson.home at gmail.com>  wrote:
>> I am interested in raw cpu integer and floating-point performance only, at
>> this point. This is *exactly* what simple synthetic benchmarks (like ubench)
>> are for. There is absolutely NO reason why performance should be 3x better
>> running under macosx than under freebsd, on the exact same hardware.
>
> Things happen that way. I've ran a firefox benchmark on different
> architecture running different systems and received different results.

On G5s that I know are properly configured (a dual-core Powermac and an 
iMac), numerical code runs at exactly equal speed on the two operating 
systems, or a little faster on FreeBSD if it spends a lot of time in the 
math library (Apple appears to have optimized system binaries with -Os 
in 10.5 for PowerPC, which pessimizes performance). If you have purely 
numerical code (check md5 -t, which both OSes have) with large offsets 
then something is clearly not set up correctly.

Did you have a chance to check the reported CPU speed in dmesg? Also, 
did you boot from the OF prompt or via autoboot? Some Apple G5 systems 
will slow the CPU and bus by large factors, in a way FreeBSD can't yet 
correct, if you start from an OF prompt.
-Nathan