Eduardo wrote:
At 01:38 26/07/2006, you wrote:
Nuno Lucas wrote:
On 7/22/06, Eduardo <[EMAIL PROTECTED]> wrote:
Sorry, but i must disagree. He uses VC6, a compiler from
you use a compiler from pre-altivec era or non altivec optimized
libraries (including libc), your code will be a lot
FYI
ACOVEA (Analysis of Compiler Options via Evolutionary Algorithm)
implements a genetic algorithm to find the "best" options for
compiling programs with the GNU Compiler Collection (GCC) C and C++
compilers. "Best", in this context, is defined as those options that
produce the fastest
At 01:38 26/07/2006, you wrote:
Nuno Lucas wrote:
On 7/22/06, Eduardo <[EMAIL PROTECTED]> wrote:
Sorry, but i must disagree. He uses VC6, a compiler from
you use a compiler from pre-altivec era or non altivec optimized
libraries (including libc), your code will be a lot slower.
While I
Nuno Lucas wrote:
On 7/22/06, Eduardo <[EMAIL PROTECTED]> wrote:
Sorry, but i must disagree. He uses VC6, a compiler from
pre-alot-of-processor-advances. So, the compiler can't, not only
compile for a more modern and different processor (different
pipeline, sse, new processor modes, etc..) but
On 7/22/06, Eduardo <[EMAIL PROTECTED]> wrote:
Sorry, but i must disagree. He uses VC6, a compiler from
pre-alot-of-processor-advances. So, the compiler can't, not only
compile for a more modern and different processor (different
pipeline, sse, new processor modes, etc..) but also use old
At 02:10 22/07/2006, you wrote:
Michael,
The guy who produced Sqlitespy is a member of this forum so he can
confirm or debunk my theory as to why you are getting a big
difference in execution time. I suspect that Sqlitespy might be
storing the SQL in its compiled (from sqlite3_prepare) form
--- michael cuthbertson <[EMAIL PROTECTED]> wrote:
> This qry is used to refresh a scrolling display where any of the constants
> are actually variables.
> The subquery returns about 1000 rows.
> Note that a LIMIT = 1 is only marginally faster than 32.
> Also, as the offset increases, the exec.
Michael,
The guy who produced Sqlitespy is a member of this forum so he can
confirm or debunk my theory as to why you are getting a big difference
in execution time. I suspect that Sqlitespy might be storing the SQL in
its compiled (from sqlite3_prepare) form and when you run it you skip
Thanks to Christian and John for the pointers regarding compilers.
I have not compiled the sqlite sources myself but have used the supplied
binary.
Could either one you give me some tips for compiling the sqlite sources for
either vs 6 or 8?
John, I will follow your advice on inline functions.
It is possible to resolve the issue by using the traditional C profiler.
Compile the SQL library with profiling on the different compilers and
measure where the time is spent during execution.
You can also compile some test programs and look at the assembler output
to get an idea of the
michael cuthbertson uttered:
Brannon:
Thank you for your thoughts.
To be clear, the 'optimize for speed' setting in MY release is actually
slower than MY debug version - I know nothing about Ralf's settings.
That issue is separate from SQLiteSpy - I didn't mean to conflate them.
And the issue
> Daniel:
> Thanks for the suggestion.
> I wasn't aware that the prepare statement gained you that
> much for one-table select queries.
> I use it for multi-100k inserts (along with trans.) and it
> saves quite a bit of time.
> This is my sql for the present problem:
>
> select * from (select
Daniel:
Thanks for the suggestion.
I wasn't aware that the prepare statement gained you that much for one-table
select queries.
I use it for multi-100k inserts (along with trans.) and it saves quite a bit
of time.
This is my sql for the present problem:
select * from (select f1, f2, f3, f4,
f5
Brannon:
Thank you for your thoughts.
To be clear, the 'optimize for speed' setting in MY release is actually
slower than MY debug version - I know nothing about Ralf's settings.
That issue is separate from SQLiteSpy - I didn't mean to conflate them.
And the issue is not which version of VS I'm
The VC6 compiler is from 1998. The VC 7.1 or 8.0 compilers produce better
compilations. I'm certain any Borland or GNU compiler of the past 3 years
would also produce better assembly than VC6. And if somebody has their hands
on a PathScale or Intel compiler, please post some benchmarks!
> I
> I've been using the SQLiteSpy sql browser tool from Ralf
> Junker at The Delphi Inspiration to test and time sql.
> Unfortunately for my assumptions, it appears that SQLiteSpy
> runs queries about 2.5x faster than using the SQLite dll in my code.
> Does anyone know how to speed up SQLite
16 matches
Mail list logo