The Quadra 700 and 900 introduced the 68040 in 1991. In great part due to a much larger L1 cache (4 KB for data and 4 KB for instructions vs. 256 bytes in the 68030) and parts of the CPU running at twice clock speed, the 68040 provides 2.5-3 times the performance of the 68030 at the same clock speed.
The regular 68040 includes its own FPU (floating point unit or math coprocessor), which had been a separate chip with the 68020 and 68030. The FPU is missing in the less expensive 68LC040 used in some low-end Macs.
Because it has such a large L1 cache, none of the 68040-based Macs shipped with with a L2 cache, although some 128 KB cards were offered by other manufacturers. As with L2 cards on the Mac IIci, these provide a 20-30% performance boost.
The next chip in the 68000 family was the 68060, a wicked fast CPU with architectural features similar to Intel’s Pentium. However, Apple decided to move in a different direction, and the ‘060 only made its way into one Amiga computer (the A4000T), some Amiga accelerators, and some Alpha computers. With such a small market, the ‘060 marked the end of the 68000 family.
Motorola 680×0 Family Overview
CPU speed* L1 cache FPU** notes 68000 8-16 MHz none none 16-bit data bus, 24-bit addressing 68020 16 MHz 256 bytes 68881 68030 16-40 MHz 2x256 bytes 68882 internal PMMU, supports L2 cache 68LC040 20-25 MHz 2x4096 bytes none can be replaced with 68040 68040 25-40 MHz 2x4096 bytes internal 68060 50-75 MHz 2x8192 bytes internal __________ * as used in Apple computers ** FPU typically used with this CPU
- Great CPUs, past and present, John Bayko. See especially sections on 8080/85, Z-80, 6502, 6809, 680×0, 80×86, ARM (used in Newton), PA-RISC, Sparc, Alpha, PowerPC, and Merced.
- MacTips, RISC, CISC, and Your Mac
- PC Magazine, Motorola and PowerPC (also covers 680×0 series)
- Pipelines, MHz, latency, caches, and more, MacKiDo
Keywords: #motorola68040 #motorolacpus
Short link: http://goo.gl/BN7YU7