* * *

Recent Posts

Pages: [1] 2 3 ... 10
General / Re: Slowness of TOpenDialog in Windows
« Last post by josh on Today at 02:28:37 am »

I find that it takes a few seconds to show whilst run in the IDE,
But when the application exe is run outside of the IDE no such delay is visible.

Is the true in your case?
Third party / Re: Pascal for TensorFlow
« Last post by lainz on Today at 02:23:00 am »
Thanks I'm interested. I need to find the time to try.

I once was on a live stream of Microsoft Latinoamérica but it was only - as I see it - a way to sell his own service of machine learning. It was too short and I get nothing with it.

Hope this is something I can afford and use.
General / Slowness of TOpenDialog in Windows
« Last post by bobonwhidbey on Today at 02:22:35 am »
TOpenDialog takes a few seconds to display in Laz while IIRC it was much faster in Delphi. Is there anyway to speed up this component, or perhaps there is a different component?
Other / Re: TPowerThread
« Last post by mercurhyo on Today at 01:26:50 am »
so... I do NOT create a TEvent every push but an acccessor to FReady  :D in a safe way, at that point of non optimized template

nothing to do with that post, I say hello here to my son Alexandre P.-C.
General / Re: compiler cannot find IdSync
« Last post by BjPascal on Today at 01:19:26 am »
Thank you both !
 I added it to indylaz and recompiled and it now works fine
Other / Re: TPowerThread
« Last post by mercurhyo on Today at 01:14:51 am »
Why do you create/release a tevent for every push?
simple :
when you 1st create a TEvent, you are the owner (depending platform and obscure PSecurityAttributes)
when you want to read the state it is a good habbit to create a checker, and not modify what the owner have set or reset (it is the 'safe' way). I know I can use the original FReady in that particular case directly. But as it is a case study made to be extended/improved... it is far from optimized (also I do not need PeekItem yet but I wrote it)
Android / Re: Android Module Wizard
« Last post by BjPascal on Today at 12:55:19 am »
using Lazarus 1.9.0 and FPC 3.1.1 on Win7 x64

Installing LAMW with OPM failed to compile amw_ide_tools
and reports:
   Cannot install package: "amw_ide_tools.lpk"

Any idea how to fix it ?
Beginners / Re: How make Random Numbers (from,to)
« Last post by VTwin on Today at 12:27:05 am »
EDIT: Thanks, very interesting. It does make sense if it is based on bit patterns, and that masking will be quicker.
Note the values itself are always valid. Performing calculations with them can give problems.

Makes sense!

For anyone who is curious, I found an interesting article discussing IEEE NaN and Infinity:


Third party / Re: Pascal for TensorFlow
« Last post by Phil on Today at 12:20:51 am »
I was curious about how well TensorFlow scales, so I timed some runs of the MNIST.pas image classification demo program. This program is not memory limited on systems with at least 4GB RAM and does not do any disk I/O within the training loop, so it should be a good example of a compute-bound program.

Note that MNIST.pas works with a 60000 x 784 input matrix (188 MB of memory).

On my old Late 2008 MacBook, MNIST took exactly 60 seconds to complete.

Next I tested MNIST on a 2017 MacBook. This is still a dual-core CPU, but a Core i7 which is clocked faster. It took 25 seconds, meaning a nice 2x bump as expected for a newer chip.

Finally I tested MNIST on an 8-core AWS instance running Ubuntu. In tests of other compute-bound models, this system has similar performance to the 2017 MacBook when utilizing only a single core. This system ran MNIST in only 12 seconds, suggesting that it was able to utilize more cores to do the arithmetic faster, without requiring any changes to the Pascal code.

Note that the TensorFlow library has a permissive Apache license. The Pascal for TensorFlow units are released under the same modified LGPL license as FPC RTL, meaning there shouldn't be many restrictions on your use of TensorFlow in any app you might write.
TAChart / Re: Transparent series
« Last post by wp on Today at 12:13:38 am »
A shaded error bar plot as mentioned in above OriginLab reference can be created by a combination of a lineseries and a stacked area series: the line series contains the "normal" data values", and the area series takes care of the error limits. The area series must contain two y values. The first y value is the lower error limit, and the second y value is the difference between upper and lower error value - when these two values are stacked on top of each other the error band is clearly visible. Since the area series always fills the area between the curve and the zero level the area underneath the lower error limit must be drawn with a clear brush. Variation of brushes between stacks can be accomplished using the TChartStyles component which must be assigned to the Styles property of the stacked series.

The attached project shows the necessary code. While everything can be done in the object inspector I decided to do it in code to show the essential steps - study the method "CreateErrorSeries".
Pages: [1] 2 3 ... 10


Get Lazarus at SourceForge.net. Fast, secure and Free Open Source software downloads Open Hub project report for Lazarus