Intel has announced some of the details of its yet to be released Pentium 4 processor. The chip has been built to take over from PIII for high-speed machines. P4 should prove to be faster and should extend the life of Intel’s Pentium brand.

P4 features Intel’s NetBurst micro-architecture, providing advances over PIII. NetBurst incorporates a number of elements that increase the processor’s ability to handle calculations more effectively.

The main aspects that dictate the performance of a processor are the frequency at which it operates (MHz) and the number of software instructions executed per cycle. Each task the computer has to do is broken down into smaller parts by the chip. These parts are executed in sequence, and this sequence is called a pipeline. NetBurst increases the number of stages in the pipeline, which allows the clock speed to be increased without damaging the chip through overheating. P4 utilises 20 stages per pipeline as opposed to PIII’s 10.

This means the P4 can be run faster. The clock speed, measured in MHz, of the P4 can increase beyond the plateau the PIII will eventually reach. At just over 1GHz the PIII in its current state is running out of headroom.

But while expanding the pipeline allows the P4 to run faster, it also makes it less efficient.

If you’re deeply techy, here is the reason why. According to technical staff at PC Advisor Reviews, when a computer program runs, depending on the outcome of a certain operation, execution can jump (or branch) to a different section of the program. When such a jump occurs the processor has to start over again, potentially losing everything in the pipeline. To combat this, processors employ a technique called branch prediction. This allows the processor to be ready when a jump occurs so the handover is reduced. The P4 features Accurate Branch Prediction, eliminating one third of the mispredictions compared to PIII.

NetBurst also features a 400MHz system bus plus additional instructions for multimedia, called Streaming SIMD Extension 2 (SSE2). SSE2 builds on Intel’s previous SSE technology, which in turn was born from MMX, by adding 144 new instructions on the chip, including support for 128bit arithmetic.