«

Chris Dollin wrote:
> Sensei wrote:
>
>
>>On 2005-10-10 15:31:40 +0200, Chris Dollin <> said:
>>
>>
>>>Sensei wrote:
>>>
>>>
>>>>So why I see so often something like:
>>>>
>>>>#define CEXCERPT do { \
>>>>some(); \
>>>>C_code(); \
>>>>here(); \
>>>>} while(0)
>>>>
>>>>Is it just some person thinking he would achieve more speed?
>>>
>>>No, it's some person arranging that their macro can be expanded
>>>as CEXCERPT;
>>>
>>>without confusion. (I would have thought, even more likely
>>>with a parameterised macro.)
>>
>>
>>I understood why he does not use a ; at the end, but why someone would
>>have such a big macro... is beyond my knowledge, something like the
>>xge_whatever stuff I found...
>
>
> How big is "big"?
>
> #define NEEDS(mill, n) \
> do { \
> if (vm->myHeap->available < (Size) (n)) \
> (FREEZE(), millRegenerate( mill ), MELT()); \
> } while (0)
>
#define INLINE inline

INLINE void needs(? mill, Size n) {
if (vm -> myHeap -> available < n) {
freeze();
millRegenerate(mill);
melt();
}
}

S.

Skarmander wrote:

> Chris Dollin wrote:
>> Sensei wrote:
>>
>>
>>>On 2005-10-10 15:31:40 +0200, Chris Dollin <> said:
>>>
>>>
>>>>Sensei wrote:
>>>>
>>>>
>>>>>So why I see so often something like:
>>>>>
>>>>>#define CEXCERPT do { \
>>>>>some(); \
>>>>>C_code(); \
>>>>>here(); \
>>>>>} while(0)
>>>>>
>>>>>Is it just some person thinking he would achieve more speed?
>>>>
>>>>No, it's some person arranging that their macro can be expanded
>>>>as CEXCERPT;
>>>>
>>>>without confusion. (I would have thought, even more likely
>>>>with a parameterised macro.)
>>>
>>>
>>>I understood why he does not use a ; at the end, but why someone would
>>>have such a big macro... is beyond my knowledge, something like the
>>>xge_whatever stuff I found...
>>
>>
>> How big is "big"?
>>
>> #define NEEDS(mill, n) \
>> do { \
>> if (vm->myHeap->available < (Size) (n)) \
>> (FREEZE(), millRegenerate( mill ), MELT()); \
>> } while (0)
>>
> #define INLINE inline
>
> INLINE void needs(? mill, Size n) {
> if (vm -> myHeap -> available < n) {
> freeze();
> millRegenerate(mill);
> melt();
> }
> }

Won't work.

FREEZE() is a macro, MELT is a macro, both of them use variables
available where NEEDS is called - such as, for example, `vm`.

And `inline` wasn't routinely available when this code was written;
furthermore, failure to inline is likely to make the code run
significantly more slowly. Since this is part of the system core,
it would matter.

--
Chris "electric hedgehog" Dollin
"I know three kinds: hot, cool, and what-time-does-the-tune-start?"

Chris Dollin wrote:
> Skarmander wrote:
>
>
>>Chris Dollin wrote:
>>
>>>Sensei wrote:
>>>
<snip>
>>>>I understood why he does not use a ; at the end, but why someone would
>>>>have such a big macro... is beyond my knowledge, something like the
>>>>xge_whatever stuff I found...
>>>
>>>
>>>How big is "big"?
>>>
>>>#define NEEDS(mill, n) \
>>> do { \
>>> if (vm->myHeap->available < (Size) (n)) \
>>> (FREEZE(), millRegenerate( mill ), MELT()); \
>>> } while (0)
>>>
>>
>>#define INLINE inline
>>
>>INLINE void needs(? mill, Size n) {
>>if (vm -> myHeap -> available < n) {
>>freeze();
>>millRegenerate(mill);
>>melt();
>>}
>>}
>
>
> Won't work.
>
> FREEZE() is a macro, MELT is a macro, both of them use variables
> available where NEEDS is called - such as, for example, `vm`.

Pass as parameters or promote to global status. Probably the latter.
"Global" with respect to the smallest unit appropriate, of course.

> And `inline` wasn't routinely available when this code was written;
> furthermore, failure to inline is likely to make the code run
> significantly more slowly. Since this is part of the system core,
> it would matter.

Sure. If inlining really won't work and profiling shows you you do need
that performance, take the macro route.

S.

Skarmander wrote:

> Chris Dollin wrote:
>> Skarmander wrote:
>>
>>
>>>Chris Dollin wrote:
>>>
>>>>Sensei wrote:
>>>>
> <snip>
>>>>>I understood why he does not use a ; at the end, but why someone would
>>>>>have such a big macro... is beyond my knowledge, something like the
>>>>>xge_whatever stuff I found...
>>>>
>>>>
>>>>How big is "big"?
>>>>
>>>>#define NEEDS(mill, n) \
>>>> do { \
>>>> if (vm->myHeap->available < (Size) (n)) \
>>>> (FREEZE(), millRegenerate( mill ), MELT()); \
>>>> } while (0)
>>>>
>>>
>>>#define INLINE inline
>>>
>>>INLINE void needs(? mill, Size n) {
>>>if (vm -> myHeap -> available < n) {
>>>freeze();
>>>millRegenerate(mill);
>>>melt();
>>>}
>>>}
>>
>>
>> Won't work.
>>
>> FREEZE() is a macro, MELT is a macro, both of them use variables
>> available where NEEDS is called - such as, for example, `vm`.
>
> Pass as parameters or promote to global status. Probably the latter.
> "Global" with respect to the smallest unit appropriate, of course.
>
>> And `inline` wasn't routinely available when this code was written;
>> furthermore, failure to inline is likely to make the code run
>> significantly more slowly. Since this is part of the system core,
>> it would matter.
>
> Sure. If inlining really won't work and profiling shows you you do need
> that performance, take the macro route.

Inlining really didn't work.

--
Chris "electric hedgehog" Dollin
Essare - to be. Essen - to be at Messe Essen for SPIEL 05.