[m-rev.] [reuse] diff: cell cache statistics

[Fergus Henderson]

On 19-Mar-2001, Peter Ross <peter.ross at miscrit.be> wrote:
> Hi,
> --- compiler/mlds_to_c.m	2001/03/18 16:56:51	1.55.2.9
> +++ compiler/mlds_to_c.m	2001/03/19 06:19:12
> @@ -2390,15 +2390,13 @@
>  	%
>  	% heap management
>  	%
> -mlds_output_atomic_stmt(_Indent, _FuncInfo, delete_object(_Lval, _Size), _) -->
> -	[].
> -	% Commented out until MR_compile_time_gc implemented.
> -	% mlds_indent(Indent),
> -	% io__write_string("MR_compile_time_gc(MR_strip_tag("),
> -	% mlds_output_lval(Lval),
> -	% io__write_string("), "),
> -	% io__write_int(Size),
> -	% io__write_string(");\n").
> +mlds_output_atomic_stmt(Indent, _FuncInfo, delete_object(Lval, Size), _) -->
> +	mlds_indent(Indent),
> +	io__write_string("MR_compile_time_gc(MR_strip_tag("),
> +	mlds_output_lval(Lval),
> +	io__write_string("), "),
> +	io__write_int(Size),
> +	io__write_string(");\n").

It would be better to put the strip_tag in the lval during
code generation, if it is needed, rather than always adding it here.

> Index: library/benchmarking.m
> ===================================================================
> RCS file: /home/mercury1/repository/mercury/library/benchmarking.m,v
> retrieving revision 1.34.4.3
> diff -u -r1.34.4.3 benchmarking.m
> --- library/benchmarking.m	2001/03/18 17:45:08	1.34.4.3
> +++ library/benchmarking.m	2001/03/19 06:19:27
> @@ -259,6 +259,11 @@
>  
>  #endif /* PROFILE_MEMORY */
>  
> +	/*
> +	** Cell cache statistics.
> +	*/
> +	MR_output_cell_cache_stats();
> +
>  	fprintf(stderr, ""]\\n"");
>  }

That should not be done unconditionally.
It should be inside some #ifdef.

> Index: runtime/mercury_memory.c
> +void MR_compile_time_gc(MR_Word cell, size_t size)
> +{
> +	if (size <= MR_MAX_CACHED_CELL_SIZE) {
> +		record_stats(size);
> +	}
> +}
> +
> +void MR_update_cell_cache_statistics(size_t size)
> +{
> +	if (size <= MR_MAX_CACHED_CELL_SIZE) {
> +		stats[size].sum_length += stats[size].length;
> +
> +		if (stats[size].length == 0) {
> +			stats[size].misses++;
> +		} else {
> +			stats[size].length--;
> +			stats[size].hits++;
> +		}
> +	}
> +}

It would be much nicer to put all the stats which
don't fit in the last bucket, rather than just
discarding them.  i.e.

	int bucket = max(size, MR_MAX_CACHED_CELL_SIZE);
	record_stats(bucket);

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
                                    |  of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh>  |     -- the last words of T. S. Garp.
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------