Blog post !

Developed or developing a new product in PureBasic? Tell the world about it.
User_Russian
Addict
Addict
Posts: 1183
Joined: Wed Nov 12, 2008 5:01 pm
Location: Russia

Re: Blog post !

Post by User_Russian »

Code from blog
static integer f_test()
Should be

Code: Select all

static integer __stdcall f_test()
Because the keyword is Procedure, not ProcedureC.
Fred
Administrator
Administrator
Posts: 14310
Joined: Fri May 17, 2002 4:39 pm
Location: France
Contact:

Re: Blog post !

Post by Fred »

It's the x64 compiler and __stdcall has no effect on it.
User_Russian
Addict
Addict
Posts: 1183
Joined: Wed Nov 12, 2008 5:01 pm
Location: Russia

Re: Blog post !

Post by User_Russian »

Sorry, maybe I misunderstood you, but isn't x86 supported anymore?
Fred
Administrator
Administrator
Posts: 14310
Joined: Fri May 17, 2002 4:39 pm
Location: France
Contact:

Re: Blog post !

Post by Fred »

Never said that, I just said the code output you showed was from the x64 compiler where __stdcall makes no sense.
RSBasic
Moderator
Moderator
Posts: 1119
Joined: Thu Dec 31, 2009 11:05 pm
Location: Gernsbach (Germany)
Contact:

Re: Blog post !

Post by RSBasic »

ImageImageImageImage Image
User avatar
skywalk
Addict
Addict
Posts: 3517
Joined: Wed Dec 23, 2009 10:14 pm
Location: Boston, MA

Re: Blog post !

Post by skywalk »

Yay Fred! - What was the compilation time difference ±optimizations?
Can the purifier merge with string compiler optimizations?
The nice thing about standards is there are so many to choose from. ~ Andrew Tanenbaum
User avatar
useful
Enthusiast
Enthusiast
Posts: 267
Joined: Fri Jul 19, 2013 7:36 am
Location: Novosibirsk

Re: Blog post !

Post by useful »

Reading about you snacking on wine with delicious cheese (experimenting with a new backend) is interesting! :D
But we want to try something to compile, too.
Is it possible in summer?
Or not before the new year?
User avatar
StarBootics
Enthusiast
Enthusiast
Posts: 706
Joined: Sun Jul 07, 2013 11:35 am
Location: Canada

Re: Blog post !

Post by StarBootics »

Hello everyone,

For me the compilation time is not really an issue as long as compiling a program don't take an hour or so. I'm playing with C and C++ codes and it take some time to get used to little longer compilation time in comparison with PureBasic, but it's bearable. So I'm willing to sacrifice some compilation time for better program performance overall.

Because of this bug I have to compile +64000 lines of code program with the standalone debugger and let me tell you that it took almost 6 seconds to have the program up and running. I hope Fred, even if is very busy with the new backend, will have this bug sorted out for the next release. But for now when I work on this big project and I need to compile with the debugger I just wait for it.

Best regards
StarBootics
The Stone Age did not end due to a shortage of stones !
User avatar
majikeyric
Enthusiast
Enthusiast
Posts: 166
Joined: Mon Oct 21, 2013 5:21 pm
Location: France
Contact:

Re: Blog post !

Post by majikeyric »

OMG the execution speed increase is AMAZING ! :shock:
User avatar
StarBootics
Enthusiast
Enthusiast
Posts: 706
Joined: Sun Jul 07, 2013 11:35 am
Location: Canada

Re: Blog post !

Post by StarBootics »

majikeyric wrote: Thu May 13, 2021 4:22 am OMG the execution speed increase is AMAZING ! :shock:
Yes Indeed ! And this speed improvement make me wonder why the FASM backend should remain after all problems will be worked out about the C backend. Even if the FASM backend remain available probably nobody will use it.

Best regards
StarBootics
The Stone Age did not end due to a shortage of stones !
User avatar
STARGÅTE
Addict
Addict
Posts: 1509
Joined: Thu Jan 10, 2008 1:30 pm
Location: Germany, Glienicke
Contact:

Re: Blog post !

Post by STARGÅTE »

StarBootics wrote: Thu May 13, 2021 11:35 amAnd this speed improvement make me wonder why the FASM backend should remain after all problems will be worked out about the C backend.
Because the compilation time is much faster in FASM. During development I would prefer a fast as possible compilation time. The fast execution time is preferred for the released version. A compilation time of more than ~10 seconds would be a no go for me.
But of cause, this is only my opinion!
I'm also very happy to see this large improvement in execution time with the C-compiler.
If it can also beat my assembler-only procedures, the future will show.
PB 5.73 ― Win 10, 20H2 ― Ryzen 9 3900X ― Radeon RX 5600 XT ITX ― Vivaldi 3.6 ― www.unionbytes.de
Lizard - Script language for symbolic calculations and moreTypeface - Sprite-based font include/module
User avatar
skywalk
Addict
Addict
Posts: 3517
Joined: Wed Dec 23, 2009 10:14 pm
Location: Boston, MA

Re: Blog post !

Post by skywalk »

Yes, this is why I asked how much compilation time the optimization settings added?
Amalgamated Sqlite.c can take almost 30sec depending on situation.

For me, it does not make sense to debug with FASM if C will introduce subtle errors or behavioral changes.
My approach will be to debug with C and -0 optimiziers and then release with the full shebang! :twisted:
The nice thing about standards is there are so many to choose from. ~ Andrew Tanenbaum
User avatar
StarBootics
Enthusiast
Enthusiast
Posts: 706
Joined: Sun Jul 07, 2013 11:35 am
Location: Canada

Re: Blog post !

Post by StarBootics »

skywalk wrote: Thu May 13, 2021 2:11 pm Yes, this is why I asked how much compilation time the optimization settings added?
Amalgamated Sqlite.c can take almost 30sec depending on situation.

For me, it does not make sense to debug with FASM if C will introduce subtle errors or behavioral changes.
My approach will be to debug with C and -0 optimiziers and then release with the full shebang! :twisted:
I guess we will have to wait'N'see how the compilation time will look like.

Best regards
StarBootics
The Stone Age did not end due to a shortage of stones !
BarryG
Addict
Addict
Posts: 1519
Joined: Thu Apr 18, 2019 8:17 am

Re: Blog post !

Post by BarryG »

BarryG wrote: Fri Mar 26, 2021 12:29 pmI'm hoping the new compiler with the new code base will generate entirely new code patterns that won't be recognized as "malware" anymore.
Yes! It does help! See my report here -> viewtopic.php?f=7&t=77459
Post Reply