Re: GNAT (GCC) Profile Guided Compilation

[Niklas Holsti <niklas.holsti@tidorum.invalid>]

On 12-07-15 02:40 , Keean Schupke wrote:
> After a bit of checking line-by-line to make sure I am using 'and
> then', 'or else' and 'constant' everywhere I can in the code, Ada is
> outperforming C++ when using profile-guided compilation for the first
> time. both C++ and Ada are getting about 40k simulations per second
> with normal compilation, C++ is achieving 56k simulations per second
> profile-guided, and Ada 57k per second.

Good news.

An Ada copmiler is of course free to implement the ordinary 
"long-circuit" operators "and"/"or" using short-circuit code, if the 
evaluation of the operands has no side effects. What are the operands 
typically like in your program? Are they function calls, or simple 
variables?

It seems to me that the general belief, regarding the expected relative 
speeds of the short-circuit code versus the long-circuit code for 
Boolean expressions with simple operands, is that the branch penalties 
on modern processors are so large that the short-circuit form is not 
obviously faster. This may explain why the Ada compiler is not using the 
short-circuit code automatically.

Clearly, if the expression is "(simple operand likely to be True) and 
(longer and longer expression)", at some point the short-circuit code 
(or changing to "and then") will become faster than the long-circuit 
code, whatever the branch penalty. This point will come sooner if 
profile-guidance is used to reduce the branch penalty.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .