PureBasic Forums - English

Restored from previous forum. Originally posted by theogott.

Maybe you assume Fred has a Big "development tower" with 50 people (like MS) ?

I doubt he can test on most CPU's out there and then try optimizations on them.

I guess wheter he has an Intel or an AMD ...

Secondly if you still think integer is "too slow", did you try it on at least the TWO actual platforms ?

Is it as slow on an Intel as well as on an AMD ?

Mostly "slowlyness" has also to do with "error-checking" and I prefer some "error-checking" at least while developement.

Forwhatever 90% of the ppl. do PB is fast enough and if its too slow, noone can beat the builtin Assembler.

My tip-of-the-day: use CMOVE-Mnemonik save BRANCHES ...




*************************
The best time to do things is now !

Restored from previous forum. Originally posted by Amiga5k.

Certainly there are many factors that can skew the results of a speed test, especially under Windows because it is forever doing something in the background (using a swap file, even with 1GB of ram.. What's THAT all about?). But, I would assume that the test of the BB and PB programs were done on the same system, so we can't completely throw away the results. Really, though, because of how bloated and inefficient Windows is, we really need to run this test in a loop, say, 1 million times so that it takes minutes or hours to finish. Then see what the time difference is. If PB is still significantly behind then Fred's got some work to do But I'm guessing that after such a long test the difference will be much smaller (percentage-wise, that is).

Russell

***Commodore 64 - Over one million cycles per second, 16 vibrant colors, 3 incredible audio channels and 38,911 Basic Bytes Free! Who could ask for anything more?***

Restored from previous forum. Originally posted by Thomas.

My post "integer too slow" was a result of a comparison to Blitz Basic 2D. If there is a Basic dialect out there that is faster than PB, Fred can optimize things because someone did it in a better way.

Anyway, I agree that tests have to be done on several CPUs to see if improvements work not only for one specific processor.

I also did not say that PB is a slow compiler, in fact it is one of the fastest. But if there is room for improvement, this forum is the right place to post these results.

The built-in Assembler is a nice feature, but I don't buy a Basic compiler for coding in Assembler. I think with some smart algorithms (see if a variable is used in the next statements and if so do not reload from memory, variable type index which is bound to EDI, i = i + 1 => inc dword[v_i], etc.) you still can enhance the generated machine code.

PureBasic is a very fast compiler but it still can be improved. And the speed and compactness of the generated code is one of the main reasons I bought this compiler.

Restored from previous forum. Originally posted by Amiga5k.

True, there is nothing in existance that can not be improved in some way and PB is no exception. Out of all of the basic languages out there, I would put PB somewhere in the top 20% for running speed and overall 'power'. Someone would have to do an exhaustive side by side comparison of the rest to find out about stability, compatibility (with different system configurations AND with older versions of its own code...), etc to see how it matches up to the competition otherwise.

In the area of price/performance (most 'bang' for the 'buck'), I'd have to say that PB wins hands down!

Russell

***Commodore 64 - Over one million cycles per second, 16 vibrant colors, 3 incredible audio channels and 38,911 Basic Bytes Free! Who could ask for anything more?***

Restored from previous forum. Originally posted by theogott.

About the last topic I could agree .
Then there are regular GREAT updates that are free !


*************************
The best time to do things is now !