beachycove
Well-known member
As I understand it:
Regular Nubus operates at 10mhz; Nubus 90 operates at double the speed, 20mhz. The full Nubus 90 implementation allows both for 20mhz communication between slots (block transfers) and the CPU/main memory. The Quadra 650, 700, 800, 900, and 950 implemented partial Nubus 90 features (610 too?), meaning 20mhz between slots, but full Nubus 90 was implemented only in the Quadra AV machines and the early PPC Macs. The difference between full and partial implementation was significant, since it was communication with main memory/CPU that was at stake.
So, the same Nubus card ought to be demonstrably faster in a Quadra 840av than in a Quadra 950 or a IIfx.
Is it really faster? Or is this (like so much that was and is Apple -- e.g., the G5 as a 64bit computer, except that the OS wasn't) mostly smoke and mirrors?
To put this another way, do Nubus cards actually take advantage of Nubus 90 technology? Some Nubus cards or all of them?
Regular Nubus operates at 10mhz; Nubus 90 operates at double the speed, 20mhz. The full Nubus 90 implementation allows both for 20mhz communication between slots (block transfers) and the CPU/main memory. The Quadra 650, 700, 800, 900, and 950 implemented partial Nubus 90 features (610 too?), meaning 20mhz between slots, but full Nubus 90 was implemented only in the Quadra AV machines and the early PPC Macs. The difference between full and partial implementation was significant, since it was communication with main memory/CPU that was at stake.
So, the same Nubus card ought to be demonstrably faster in a Quadra 840av than in a Quadra 950 or a IIfx.
Is it really faster? Or is this (like so much that was and is Apple -- e.g., the G5 as a 64bit computer, except that the OS wasn't) mostly smoke and mirrors?
To put this another way, do Nubus cards actually take advantage of Nubus 90 technology? Some Nubus cards or all of them?