it is really not a PC vs. MAC question ... it is a question of available bandwidth for several components needed to make a computer work. this starts with RAM, continues over the chipset and various controllers ending at the processor(s) and i'm leaving aside now the slowest of all - the harddisk-part.
every single action on a computer has more or less I/O-operations - read from storage, store data to RAM, read from RAM, load into processor cache, load into processor register, write back to memory, send to (audio-)device, ect. now there are three principles to increase the overall number of I/O operations: increase speed of processor, increase number of processors, increase number of clustered computers - each of them has its pros and cons.
a series of adjustments in computer design during the last years resulted in currently more powerfull machines than we ever had - crossbar technology (ICH-chipsets), larger and shared processor chaches, dual-bus RAM, serial communication (sATA, PCIe), but beyond certain limits it becomes very expensive very quickly. everybody who denies or at least palliates that, is - sorry - IMO a hypocrite.
once we have a really working 64bit multi-processor operating system on the desk developers can adjust their applications to make the most possible use of the offered power and i'm predicting: just to find new bottlenecks. considering the amount of data which is needed to be processed we cannot assume one will see a system running eg. a full symphonic cube (and more) plus an application like MIR on a single system.
and let me add another two points regarding platforms and operating systems:
as soon as we have more power available it will be eaten up by some nifty features no one really needs. just as good as i personally don't need aqua, i don't need the playmobile-appearance of XP and VISTA (the first thing to be switched off after a setup), for some reason the very old and very good openGL idea was ignored, sometimes it actually seemed to be dead and it looks like re-inventing the wheel is the favourite job of GUI-designers.
in the PC world i can run windows95 programs still on VISTA (no matter if i'd even need them or not) but try to launch an application built for OS7 and 68K processors on an ourdays MAC. OS8, OS9, OSX (in several flavours) passed by by as well as the PPC processor. i have still in the storeroom my iMAc 233 from 1998 which was never capable to run more than OS 10.1 - on the other side i have an ASUS P55T2P4 (dual PII 266) from 1997 originally running NT which does - tiring but nevertheless - run windows2003.
so whom should one trust if a decision for a platform has to be passed?
christian
every single action on a computer has more or less I/O-operations - read from storage, store data to RAM, read from RAM, load into processor cache, load into processor register, write back to memory, send to (audio-)device, ect. now there are three principles to increase the overall number of I/O operations: increase speed of processor, increase number of processors, increase number of clustered computers - each of them has its pros and cons.
a series of adjustments in computer design during the last years resulted in currently more powerfull machines than we ever had - crossbar technology (ICH-chipsets), larger and shared processor chaches, dual-bus RAM, serial communication (sATA, PCIe), but beyond certain limits it becomes very expensive very quickly. everybody who denies or at least palliates that, is - sorry - IMO a hypocrite.
once we have a really working 64bit multi-processor operating system on the desk developers can adjust their applications to make the most possible use of the offered power and i'm predicting: just to find new bottlenecks. considering the amount of data which is needed to be processed we cannot assume one will see a system running eg. a full symphonic cube (and more) plus an application like MIR on a single system.
and let me add another two points regarding platforms and operating systems:
as soon as we have more power available it will be eaten up by some nifty features no one really needs. just as good as i personally don't need aqua, i don't need the playmobile-appearance of XP and VISTA (the first thing to be switched off after a setup), for some reason the very old and very good openGL idea was ignored, sometimes it actually seemed to be dead and it looks like re-inventing the wheel is the favourite job of GUI-designers.
in the PC world i can run windows95 programs still on VISTA (no matter if i'd even need them or not) but try to launch an application built for OS7 and 68K processors on an ourdays MAC. OS8, OS9, OSX (in several flavours) passed by by as well as the PPC processor. i have still in the storeroom my iMAc 233 from 1998 which was never capable to run more than OS 10.1 - on the other side i have an ASUS P55T2P4 (dual PII 266) from 1997 originally running NT which does - tiring but nevertheless - run windows2003.
so whom should one trust if a decision for a platform has to be passed?
christian
and remember: only a CRAY can run an endless loop in just three seconds.