Tested myself... benh wrote: >> Reasons might be many and various: memory bandwidth being _much_ >> lower on all but the latest Apple machines, compared to equivalent >> i386 boxes, being one of them. RISC vs. CISC code generation might be >> another, also that is pure speculation on my part. Different levels >> of compiler omptimisation might be another.
> Note that all recent Apple machines (the "core99" series) have a very > much higher memory bandwidth (100Mhz on a 64 bits wide bus and better > memory controller) than any previous model. > Also, the page copy optimisations Paulus added to 2.4 seem to give > full benefit on G4 CPUs (better cache management than G3s ?) to reach > about 200Mb/sec on a G4/350 with the hdparm buffer cache test. Same for me on the TiBook/400; between 200 and 220 MB/s And yes, the 2.4 kernel is a _big_, even _huge_ improvement: no less than _double_ the memory bandwidth, according to the hdparm buffer test. Cheers Michel ------------------------------------------------------------------------- Michel Lanners | " Read Philosophy. Study Art. 23, Rue Paul Henkes | Ask Questions. Make Mistakes. L-1710 Luxembourg | email [EMAIL PROTECTED] | http://www.cpu.lu/~mlan | Learn Always. "
