11.2.2 Predication in IA-64
Now that the performance implications of branching have been described, this section overviews predication – the primary IA-64 mechanism used by optimizations described in this section.

Almost all IA-64 instructions can be tagged with a guarding predicate. If the value of the guarding predicate is false at execution time, then the predicated instruction’s architectural updates are suppressed, and the instruction behaves like a nop. If the predicate is true, then the instruction behaves as if it were unpredicated. There are a small number of instructions such as unconditional compares and floating-point square-root and reciprocal approximate instructions whose qualifying predicate do not operate as described above. See Part I: IA-64 Application Architecture Guide for additional information.

2.8 Predication
Predication is the conditional execution of instructions. Conditional execution is implemented through branches in traditional architectures. IA-64 implements this function through the use of predicated instructions. Predication removes branches used for conditional execution resulting in larger basic blocks and the elimination of associated mispredict penalties.

To illustrate, an unpredicated instruction when predicated, would be of the form
In this example p5 is the controlling predicate that decides whether or not the instruction executes and updates state. If the predicate value is true, then the instruction updates state. Otherwise it generally behaves like a nop. Predicates are assigned values by compare instructions.

Predicated execution avoids branches, and simplifies compiler optimizations by converting a control dependence to a data dependence. Consider the original code: The branch at (a>b) can be avoided by converting the code above to the predicated code: The predicate pT is set to 1 if the condition evaluates to true, and to 0 if the condition evaluates to false. The predicate pF is the complement of pT. The control dependence of the instructions c = c + 1 and d = d * e + f on the branch with the condition (a>b) is now converted into a data dependence on compare(a>b) through predicates pT and pF (the branch is eliminated). An added benefit is that the compiler can schedule the instructions under pT and pF to execute in parallel. It is also worth noting that there are several different types of compare instructions that write predicates in different manners including unconditional compares and parallel compares.

Если кого волнует скорость (лишние проверки).
1) ...
2) Не факт что в ущерб. См. след. два сообщения

В заголовке темы ошибка, не "1 ур.вл-ти", а 2.

Вместо 3,

Если я не ошибаюсь, по той же причине из Оберона убрали паскалевский IF ... THEN ...; и заменили на блочный IF THEN ELSE END.

А я с вами не согласен.
...
Любой ветвистый if-овый код может (и должен ) быть сведён к этой канонической форме.