Re: C versus Ada (once again :-)), was: Re: F-22 ADA Programming

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Mon, 03 Nov 2014 14:58:06 -0700, Shark8 wrote:

> On 03-Nov-14 05:40, Florian Weimer wrote:
>> * Simon Clubley:
>>
>>> As I keep saying in one way or another:
>>>
>>> 	Ada: nice language, lousy compiler situation.
>>
>> If you are trying to parallelize your programs for performance
>> reasons, Ada (as standardized) is not such a nice language anymore
>> because it lacks a modern memory model (which reflects the variations
>> of existing hardware) and access to various forms of atomic
>> instructions with well-defined properties.
> 
> Isn't the atomic instruction thing properly recognized as platform 
> dependent?

Instruction is a wrong term, should be "atomic primitive" instead.

> If that's the case, then should it be part of Ada, or would 
> it be acceptable to have a machine-language (e.g. System.Machine_Code) 
> insert for handling it in the implementation part of some package?

No, that would be practically useless.

Atomic primitives must be machine-independent. That is, the compiler would
select the most appropriate implementation for.

For example, consider the atomic increment primitive (of some package
Interfaces type, e.g. Unsigned_64). When there is a machine instruction the
compiler uses it, otherwise it falls back to a protected operation or OS
calls.
 
> (Would such atomic instructions need to be incorporated at the Run-time 
> level?)

Only if the implementation is based on OS calls. E.g. using Windows API
function InterlockedIncrement.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de