Overview Features Coding ApolloOS Performance Forum Downloads Products Order Contact

Welcome to the Apollo Forum

This forum is for people interested in the APOLLO CPU.
Please read the forum usage manual.
Please visit our Apollo-Discord Server for support.



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
Performance and Benchmark Results!

High Level Languages: How to Proceedpage  1 2 3 4 

Wawa T

Posts 695
05 Sep 2017 19:06


Steve Ferrell wrote:

    I seriously doubt that any of my comments will drive assembly language programmers away from programming.

 
  i wasnt even talking of assembly coders specifically. neither i think that this your specific comment here will have great influence on coure of matters. but similar attitude ("if you cant code without debug tools, then get lost" - for example) if embraced and displayed by general public, might unnecessarily disencourage people, who would have at least a chance to learn to code better otherwise.
 
 
compiling was painfully slow at times so in the future I will use the AmiDevCPP cross-compiler on Windows:  EXTERNAL LINK   

 
  it has its quirks as well. id rather advise against using this frontend. it also has a specific manner of creating makefiles, which are not really portable outside the devcpp. you could use cahirs or bubbob versions of gcc simply from the shell writing generic makefiles.
     
 

    To test the EXE's I'll simply have the compiler's output place the binary into a shared WinUAE folder on my development system and run AmiDevCPP and WinUAE concurrently. That way I can test immediately after a compile/recompile.
 

 
  thats about the same how i proceed. however it doesnt change the fact that the absence of good debugging tools under 68k or uae isnt really helpful, especially when testing and debugging complex projects such as bigger apps or operating systems (components).
 


Steve Ferrell

Posts 424
05 Sep 2017 20:05


wawa t wrote:

                i wasnt even talking of assembly coders specifically. neither i think that this your specific comment here will have great influence on coure of matters. but similar attitude ("if you cant code without debug tools, then get lost" - for example) if embraced and displayed by general public, might unnecessarily disencourage people, who would have at least a chance to learn to code better otherwise.
   
    it doesnt change the fact that the absence of good debugging tools under 68k or uae isnt really helpful
   
             

             
Why do you imply that I included ALL debug tools in my comments?
  And there are plenty of good assembler (and C/C++) debugging tools for the Amiga, both MMU and non-MMU.  Did they all just disappear and no one told me!!! 
             
And where did I imply that coders should "get lost" if they don't like the fact that the Vampire doesn't have an 030/040/060 compatible MMU?  I stated that if they are relying heavily on an MMU to debug their code that they're obviously exercising poor coding skills/habits and need to go back to the book or go back to school and educate/re-educate themselves in proper coding.
             
Has anyone taken away their existing tool sets?  Has someone recently broken into their homes and stolen their Amigas?  Or stolen their copies of UAE/WinUAE?  All the tools that assembly coders (and C/C++ coders) have had in the past are still at their disposal.  Why do you act as if those tools no longer exist or that they are now somehow unavailable simply because the Vampire doesn't possess an MMU?
             
I'm tired of the false narrative that just because the Vampire doesn't have an MMU that coders will just shut themselves off because they won't be able to debug their code.  That is just as ludicrous as believing that C/C++ programmers will stop programming for the Amiga because the Vampire doesn't support Visual Studio!
     
     
             


E Penguin

Posts 46
06 Sep 2017 00:19


Samuel Crow wrote:

  Wrapping each individual opcode in libcalls would kill the performance advantages due to calling overhead.  Vector units are exceptionally good at handling large streams of data at once.

I was thinking at more of a functional level. Like vector operations - provide a pointer to the vector structure in the function call and data comes back after the processing is completed. A fast fourier function, matrix rotation... Surely these things are faster in the vampire and are generic enough to wrap into a library.



Eric Gus

Posts 477
06 Sep 2017 00:26


Samuel Crow wrote:

  @Eric Gus
  Mike Ness has a Vampire 500+ v2 right now so using GCC with AmosPro would break the memory bank.  I've discussed the matter with him extensively though.  Something is coming but even I am not sure what just yet.

Ok, as long as they are "in the loop" ..

This is good news that they 1. have a vampire 2. they are cooking something up.. AMOS really is my first choice in terms of a dev language, I really like it and nothing screams "AMIGA" to me more than AMOS .. so this really does need to happen.. 

Platforms live and die by applications these days (not so much hardware specs)..  you can have the best hardware in the world but if you lack applications you probably wont be around for long.


Wawa T

Posts 695
06 Sep 2017 07:31


Steve Ferrell wrote:

  I'm tired of the false narrative that just because the Vampire doesn't have an MMU that coders will just shut themselves off because they won't be able to debug their code.  That is just as ludicrous as believing that C/C++ programmers will stop programming for the Amiga because the Vampire doesn't support Visual Studio!           
 

 
  false narrative.. sigh.. id first like to see this vivid coding community you are talking about. thre are few asm coders who like to write demos and chat on forums, but usually we cant find people to fix or implement simplest stuff, be it on amiga or aros68k.
 
  and im not talking about visual studio, like you now are implying, it would be great if one could dependably use for instance gdb, because the current practice is to laborously pflaster the code with debug statements. at least thats what known to me and what i have been told by the most experienced, like toni.
 
  now, im certainly not very gifted in that area, but at least im actively contributing to development according to my skills and i know the current problems and dynamics behind the teams a bit. if you have once coded something amiga twenty years ago, you may forgot what the problems are and not grasp that the general lack of improvement didnt really help the scene.
 
  as example: only lately we have people working again on contemporary gcc (cross) backends for 68k. till then state of the art was still 2.9.x. but usually you cant translate contemporary cross platform sources with it.. right?


Steve Ferrell

Posts 424
06 Sep 2017 08:00


wawa t wrote:

    false narrative.. sigh.. id first like to see this vivid coding community you are talking about. thre are few asm coders who like to write demos and chat on forums, but usually we cant find people to fix or implement simplest stuff, be it on amiga or aros68k.
   

   
Why do you insist on putting words into my mouth?  I never implied nor stated that there's a vibrant community of classic coders.  There clearly isn't a vibrant community and there won't be one in the future either. But there are still a number of classic coders, myself included, who upload software to Aminet routinely in case you haven't noriced and most of us use the debugging tools that you say don't exist for 68K or UAE. 
   
This may be news to you, but new-comers to programming don't seek out obscure hobby boards and dead operating systems as a starting point.  If they do, they're destined to fail or at least in for a very long and painful learning experience.
   
And I know you're not talking about Visual Studio, I'm the one who mentioned it because your premise is ludicrous regarding a complete lack of debugging tools for 68K Amigas and UAE.  Your premise is false.  And you keep putting forth a false narrative on this and other forums that no one will code for classics or Vampires because of the Vampire's lack of an MMU.  You're clearly mistaken.  People are coding for the Vampire now (and classics) and using existing 68K debugging tools, in spite of the "complete lack of debug tools narrative" that you continue clinging to.
   
As for coding 20+ years ago, you're sorely mistaken there as well.  I'm a software developer and business owner in the field of LIDAR and have focused on Windows, Mac and Linux for well over 20 years and continue developing even as I write this. EXTERNAL LINK
 
BTW, when I updated the PLPLOT library, I used the very debugging tools that you say don't exist for 68K Amigas.  Amazing how these "phantom" tools worked for me.  I'll continue using them on my classic A1200 when I'm not cross-compiling from my Windows box.  The PLPLOT update was my first serious classic development in over 20 years but I'm certainly not out of touch with coding on the Amiga and several other platforms
    EXTERNAL LINK   


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
06 Sep 2017 09:04


Steve Ferrell wrote:

BTW, when I updated the PLPLOT library, I used the very debugging tools that you say don't exist for 68K Amigas.  Amazing how these "phantom" tools worked for me.   

Hi Steve, mind that WAWA is mainly a user and not a coder and has no own experience here. So don't be to harsh with him.

Steve has of course a valid point that 99.99% of all software for C64, AMIGA, ATARI, (many other systems) all NES, SEGA, NEO-GEO, or all ARCADE and CONSOLE games were developed without using an MMU.
Looking at these literally hundreds of thousands of games and applications and including the Operating systems - makes it obvious that claiming software can not be developed without using an MMU is silly talk.




Wawa T

Posts 695
06 Sep 2017 11:24


Steve Ferrell wrote:
your premise is ludicrous regarding a complete lack of debugging tools for 68K Amigas and UAE.

there isnt "complete lack", but since you are so convinced that the current state of debugging tools is convenient id welcome your help debugging of a lot of stuff i wasnt able to finalize. in particular on aros 68k which is a community project for advantage of all amiga users, as you must know.

And you keep putting forth a false narrative on this and other forums that no one will code for classics or Vampires because of the Vampire's lack of an MMU.

curious where i stated it here, let alone any other forums. all i said is that things like mutools were of advantage. seems its you putting words in my mouth, while mistaking me for some anti vampire troll on a conquest riding the forums.

   
As for coding 20+ years ago, you're sorely mistaken

thats what i understood from your own statement, that you didnt code for amiga in the meantime, and thats what im specifically referring to. you dont need to be upset about it. it inst meant as an insult.

BTW, when I updated the PLPLOT library, I used the very debugging tools that you say don't exist for 68K Amigas.  Amazing how these "phantom" tools worked for me.

well what can i say. id be very grateful to be tought how to properly debug stuff on 68k. so far i have a number of open issues i dont know how to proceed further, alas.

@gunnar
i dont require any mercyful treatment, i can live with heated discussion, its not that different on aros ml for that matter.
to be honest, while im a "user" as you describe it, most of my occupation with amiga last years was with testing, debugging and some assistance coding, which you can follow from my commits to aros repo.

the subject in question, 68k debugging is something, that im very concerned about. perhaps we should open a thread with advises in this field, however a general one not just apollo related.


Wawa T

Posts 695
06 Sep 2017 11:34


Gunnar von Boehn wrote:

  Steve has of course a valid point that 99.99% of all software for C64, AMIGA, ATARI, (many other systems) all NES, SEGA, NEO-GEO, or all ARCADE and CONSOLE games were developed without using an MMU.
  Looking at these literally hundreds of thousands of games and applications and including the Operating systems - makes it obvious that claiming software can not be developed without using an MMU is silly talk.
 

 
  one needs to take into account that probably many of the commercial titles may have been cross developed on other systems. also in the beginning the target was clearly set (mc68000 chipset ram) ao many bugs became apparent only later. refraining from opinions about the coding practices, fact is that a lot of amiga software is being released broken, never to be fixed, and then sources are lost. because there is lack of actual users and testers, especially such that bother to employ, say mungwall or muforce to see if the application isnt silently trashing someone elses memory. at leats i have seen a lot of that.


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
06 Sep 2017 12:08


wawa t wrote:

say mungwall or muforce to see

Many people seem to have some huge misconceptions here.

Lets state some facts, and lets bust some myths here.

The "idea" of catching an illegal memory access or null-pointer access is very simple. Also implementing this on AMIGA can be done very fast.
And writing a basic "Enforcer" skeleton is task a good coder can do
in very little time on AMIGA.

Of course if you go for many features and gui, dissambler etc - doing the whole will take a lot time to polish. But a 90% solution which gets the basic job of catching errors done - is done in 10% of the time.

FLYPE did demonstrate this before, he wrote a BASIC ENFORCER tool for 68080, using the 68080 MPU, basically on a sunday afternoon.

You see the very same with the FPU.
People were spending days on writing FORUM posts about FPU.
Instead doing anything usefull.

JARI did write a working FEMU skeleton, and you have to mind that it was his 1st AMIGA ASM program.
JARI did get FEMU working and calculating FPU able to run some programs like fractals renderer writing this over a Weekend.

So what do we see here?

There are programming tasks, for which a real programmer needs few hours.

What I see here is that people talk, moan, whine and spending often many times writing essays in forums - that doing the code would have needed.




Samuel Devulder

Posts 248
06 Sep 2017 12:48


+1

As a matter of fact, complaining is always easier.


Wawa T

Posts 695
06 Sep 2017 13:15


Gunnar von Boehn wrote:
Of course if you go for many features and gui, dissambler etc
 

 
  i think being able to identify the instruction in a binary with debug symbols compiled in, where crash or memory hit happens is good enough even for a noob.
 
 
What I see here is that people talk, moan, whine and spending often many times writing essays in forums - that doing the code would have needed.

 
  for the record. i dont think im doing that.


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
06 Sep 2017 13:31


wawa t wrote:

 
What I see here is that people talk, moan, whine and spending often many times writing essays in forums - that doing the code would have needed.

   
  for the record. i dont think im doing that.
 

 
I was not referring to you.
 
But there are enough people with no clue and with bad intentions in AMIGA scene.
 
Lets take MATT and MEYNAF as "bad" example.
Years ago, when they we kicked out of the team for good reason.
They swore to do their best to harm the project.
 
Meynaf did promise us, that he will take revenge and that he will do everything including lying to prevent people from supporting 68080.
 
And this is exactly what both are doing.
Both Meynaf and Matt have in fact, zero clue about APOLLO 68080.
 
Both don't know about how APOLLO works internally.
They have no knowledge about AMMX.
 
Both spend huge amount of time writing nonsense essay in forums - containing no facts, with the goal to put doubts in peoples mind.
 
There is only one reason for this.
They want to harm the project as much as possible as revenge.


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
06 Sep 2017 13:54


There are also non-personal reasons why people try to harm the project by posting lies.
 
Lets look at A1k.org for example.
 
We got "offers which we could not reject" from competing manufacturers - to "cooperate" or otherwise take the risk that people writing non educated, halve wrong posts on A1k forum could "accidentally" harm the public reputation of the Vampire cards.


Michal Warzecha

Posts 209
06 Sep 2017 14:06


I'm not sure You should care about this much. Good product will defense himself. And many people, with me, thinks Apollo/Vampire is good, or even best product in Amiga world.


Chris H

Posts 65
06 Sep 2017 14:18


People and community will more and more come behind this . You guys of Apollo Team always proved that you have not only the capabilities of developing a excellent working CPU and all sideproject stuff. You furthermore as an excellent team will do the job to negate the hatespeech and lies together with the growing number of card owners and developers, like me. It's just beginning.


Steve Ferrell

Posts 424
06 Sep 2017 18:57


Gunnar von Boehn wrote:

There are also non-personal reasons why people try to harm the project by posting lies.
 
  Lets look at A1k.org for example.
 
  We got "offers which we could not reject" from competing manufacturers - to "cooperate" or otherwise take the risk that people writing non educated, halve wrong posts on A1k forum could "accidentally" harm the public reputation of the Vampire cards.

Funny how they won't come right out and name the "competing manufacturers".  There's only one other Amiga hardware manufacturer out there with the resources to make such offers and it goes by the name of A-EON.  It's also interesting how hostile things have become at Amiga.org since A-EON bought the site. Hostile to anyone and anything that is critical of their hardware or anyone who has anything positive to say about the Vampire.  And when you confront the misinformation presented there with facts they counter with personal attacks, and it's always the same handful of users there who are doing that.  What I find even more disturbing is that they show up here from time to time to stir up trouble too.  But it's laughable to see them confront the lead Vampire developer and act as if they know more about the Vampire than its designer!



Gregthe Canuck

Posts 274
06 Sep 2017 19:45


Hey Steve -

At the end of the day just ignore the trolls. As I keep reminding the A=/V= team, just keep shipping product. Nothing else matters in the long run. More product = more users = more interest/advocacy.
 
It is sometimes hard to keep a healthy perspective when you see some really idiotic postings but in the long run they don't matter one bit. It doesn't help when some Vampire users troll back, but that's just part of the 'fun' in the forums.
 
Keep your stick on the ice.


Steve Ferrell

Posts 424
06 Sep 2017 19:51


gregthe canuck wrote:

Hey Steve -
 
  At the end of the day just ignore the trolls. As I keep reminding the A=/V= team, just keep shipping product. Nothing else matters in the long run. More product = more users = more interest/advocacy.
 
  It is sometimes hard to keep a healthy perspective when you see some really idiotic postings but in the long run they don't matter one bit. It doesn't help when some Vampire users troll back, but that's just part of the 'fun' in the forums.
 
  Keep your stick on the ice.

Thanks.  I'll try to follow your advice.  But answer this for me.  Does it seem to you as if the Amiga community has more than its fair share of idiots and trolls?  I follow several IT interests online but the Amiga community takes the cake with regard to trolls, idiots and cyber-bullies.  Hanging out too much at A.org and a few other sites can certainly skew one's view of reality.



Wawa T

Posts 695
06 Sep 2017 21:42


Steve Ferrell wrote:
Funny how they won't come right out and name the "competing manufacturers".

well. id say while (so called "cooperation"), communication and the aquisition methodes arent always that fair as it seems. (id agree with gunnar here). the concurrents have thier own different concepts to propose. lets say fpga-arcade with their arm-68k-emu-board, mboehmer with his pci expansion based on krashans prometheus, the sonnet ppc guys, and so on.. since they are so different and mostly concurrent its difficult to unify these under one roof of approach.

posts 80page  1 2 3 4