#instruction sets

[feathershy]

the failure of Itanium is interesting since it is a more extreme version of "what if we push more responsibilities onto the compiler?" that RISC does. Namely, the compiler is responsible for what would be a superscalar architecture. (or so is my naive understanding). maybe there are trade-offs between pushing work onto the compiler and letting code run on different CPUs (and designing new CPUs which run the code faster)

[destinationisontheroadonly]

Twisted Transistor!

Computers are integer machines and are capable of representing real numbers only by using complex codes. The most popular code for representing real numbers is called the IEEE Floating-Point Standard. The term floating point is derived from the fact that there is no fixed number of digits before and after the decimal point; that is, the decimal point can float. There are also representations in which the number of digits before and after the decimal point is set, called fixed-point representations. In general, floating-point representations are slower and less accurate than fixed-point representations, but they can handle a larger range of numbers. Note that most floating-point numbers a computer can represent are just approximations. One of the difficulties in programming with floating-point values is ensuring that the approximations lead to reasonable results. Because mathematics with floating-point numbers requires a great deal of computing power, many microprocessors come with a chip, called a floating point unit (FPU), specialized for performing floating-point arithmetic. FPUs are also called math coprocessors and numeric coprocessors.

Since the earlier microprocessors didn't actually have any floating-point capabilities, they only dealt with integers. Floating-point calculations were done on separate, dedicated hardware, usually in the form of a math coprocessor. But when the integrated circuit chip technology began to sore, the computing industry found that the size of the transistor could be reduced - therefore more transistors could be etched onto the semi-conductor material inside of the physical chip. The decrease in transistor size permitted a floating-point unit directly on the main CPU die. Adding these units to the main CPU die (physical core that functions at the central processing unit) added hardware and floating point instructions into the mix. It would eventually introduce a set of extended (opcode) instructions called SIMD.

Today, modern microprocessors can execute the same instruction on multiple data. This is called Single Instruction Multiple Data (SIMD). SIMD instructions handle floating-point real numbers and also provide important speedups in algorithms. Because the execution units for SIMD instructions usually belong to a physical core, it is possible to run as many SIMD instructions in parallel as available physical cores. As mentioned, the usage of these vector-processing capabilities in parallel can provide important speedups in certain algorithms. The addition of SIMD instructions and hardware to a modern, multi-core CPU is a bit more drastic than the addition of floating point capability. Since their inception, a microprocessor is an SISD device (Single Instruction stream, Single Data stream).