Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
Speed is not strictly tied to data/address path widths: it is basically dependent by the clock frequency applied to the circuitry.
Maybe you are mixing speed with throughput.
As far I know microprocessor speed is not directly related to data/address width: it depends from the internal architecture of the CPU.
A technique commonly used in modern CPU to increase speed is implementation of SIMD (Single Instruction Multiple Data): the same word is used to store the instruction opcode and and two or more data word (the operands) to be processed in a single clock cycle.
Hope this could be helpful to you at least as a hint to start a search on the Internet, or to stimulate the help of people involved with CPU design in this Forum.
Some MCU multiplex the address bus. eg 8051 address bus is only 8 bits wide and requires an external latch. This uses clock cycles and the net result is a mpu that divides the crystal frequency for internal use. Therefore address bus width can affect the speed
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.