« Reply #12 on: September 23, 2010, 11:18:33 AM »
No, it doesn't.
Note: I answered the question.
I think you might have missed my second edit - I was referring to file systems in general (in the context Amiga file systems).
IIRC the Amiga has PIO IDE rather than DMA so you'd expect performance benefits if the block size is optimal for clock cycles, i.e. the minimum number of cycles required to request across the IDE bus and CRC the block of a given size would be optimal. Having a faster CPU I expect would improve this as it would allow more CPU cycles to be dedicated to IO, assuming a modern drive with fast IO.
Later 68K processors (IMHO) benefit IO from having a data cache (i.e. not the 68000/68020). 68030's data cache is 256B, 4K for '040 and 8K for '060 so IMHO an 040 upwards using PIO and a decent FS should (assuming not a lot of small files) see benefits of using a 1-4K block size.
Does this sound right (I'm not a hardware engineer)?
« Last Edit: September 23, 2010, 11:24:13 AM by brownb2 »
Logged
A600 (3.48/4.46 MIPS) - 2MB Chip, 4MB PCMCIA, 11MB Fast, ACA-620 OC 680EC20@16.67/25MHz, RTC, 512MB CF HD, WB 2.1
A1200 (7.36 MIPS) - 2MB Chip 8MB Fast, MTEC Viper 68030@42MHz MMU, 68882 FPU, RTC, 1GB CF HD, T.Turbo 2, 2 Ext. Floppies, WB 3.0.
Atari 520 STFM - 1MB, Multiface ST.
Commodore 16 - 64K Mod, SD2IEC Drive
Commodore 64C - 64K, SD2IEC Drive
ZX Spectrum 48K Rubber Key - 48K, Composite Mod
ZX Spectrum +2B - 128KB
ZX Spectrum +3 - 128KB, DivIDE+ 64MB CF HD, Multiface 3.