accessing the operandsoperands are usually place

Question

Accessing the Operands:

• operands are usually place in one of two places:

-registers (32 int, 32 fp)

-memory (232locations)

• registers are

-simple to specify

-shut to the processor (fast access)

• The idea that we desire to access registers whenever probable led to load-store architectures.

-normally all arithmetic instructions only access registers

-just access memory with explicit and stores and loads.

Load-store architectures

It can do: add r1=r2+r3 and then load r3, M(address)

⇒forces serious dependence on registers, which is accurately what you desire in today's CPU can't do add r1 = r2 + M(address)

- Other more instructions

+ Fast implementation (for example: easy pipelining)