Who is the true inventor of GPU main code execution?

[Gorf]

Let me just preface this post with the following......

 

This post i snot intended to incite a flame war nor to cause any further rift in the devloper side

of the community....but enough is enough....I need to set this right....also...and more importantly....

 

Atari Owl is a brilliant coder and I tip my hat to his coding prowess and wish I had half his

knowledge on how to deal with 3D math. However, knowing the kind of decent and honorable man

he is, I am sure he would want to be sure that proper credit is given where due. I also consider

Atari Owl a dear and valued friend and I would not take his due part of the credit from him and

I suggest those of you still bitter refrain from doing the same to me. Truthfully, I could care

less if you have any respect for me...but at least show respect to Atari Owl byt being honest and

accurate...if you are even capable. You know who you are.

 

Now onto the reality of the matter.....

 

It seems there is a certain web site that likes to print many and several inaccuracies about many things

concerning 3DSSS and myself. One is who the 'true inventor and master' to put it as 'they' do, of GPU main

RAM code execution..... Also, There is another inaccuracy claiming, other than Atari Owl, no one else has

use this workaround. I'll deal with the later first. Gorf Classic CD uses GPU main RAM code execution long

before I announced it was possible. Also Surrounded, a released 3DSSS demo is 95% GPU main code RAM executable.

The only 68k code is the set up code which is recommended by Atari and a wise choice and the a vblank handler.

 

To set the former part of the record straight(i'll ignore the other out and out BS and sour grapes

for now) and will post from Atari Owl's project blog site....his own words in writing.....

 

"There is a bug which prevents JUMP (jump absolute) and JR (jump relative - a short distance)

commands working properly within the Jaguar's Main RAM. Atari knew about this and said that it

essentially restricts the Jaguar RISC chips to only running programs within their own very small

local memory.

 

Now back in 2004, when I was working on my voxel engine, I was coding it on the GPU, then I switched it

to the DSP and then back again, this happened several times as I tried out different ideas. During on of

these experiments I noticed data was not changing in RAM where I expected it to... when I looked I found

I had a program resident on the DSP which was running on the GPU – that is the GPU program was running in

external RAM. It was therefore easy to transfer this program to Main RAM and run it there also.

 

Addition or removal of instructions caused problems so I reasoned that this was to do with the alignment

of the jumps and the jump to locations. This proved to work for long alignment of JUMP from and to locations

and I was thus able to JUMP reasonably reliably… jumping between Main RAM and Local proved slightly trickier

until I introduced an extra pipeline/pre-fetch clear with the addition of a MOVEI.

 

No my early code (other than the voxel engine which had worked essentially by chance) proved very awkward to

manage and not ENTIRELY to follow the rules I had set out. Though by trial and error I was able to word align

some jumps to gain reliable execution."

 

Perhaps this tormented person missed this most important admition from Atari Owl himself.....

 

"Gorf in the meantime was able to (completely independently) correctly identify exactly when jump to and

from locations should be word or long aligned – something which would immensely reduce the amount

of trial and error required in my own code."

 

"Gorf and I therefore formulated a set of rules for the use of GPU code running in Main RAM......"

 

Atari Owl is responsible for the masterful feat of jumping to and from Local and main RAM and kudos and

bravo to him.

 

You can read the rest of the FACTS right here....From Atari Owl's own description of who is and is not "true

inventor and master" of GPU main RAM code execution.

 

http://atariowlproject.blogspot.com/2009/12/merry-christmas-and-happy-new-year.html

 

I thank you for taking the time to read this,

 

Sincerely,

 

Gorf

[+LS650]

Does it really matter?

[Gorf]

Does it really matter?

 

I feel it does.....

[Chickybaby]

AFAIK Albert approved this thread and Gorf obviously fel it important enough to submit for approval.

Everyone is entitled to their opinion as long as it is not harming another and respectful. If anyone does not feel that this is important to them then please don't read the thread. If anyone does not want to read what a member has to say then that is why there is an 'ignore user' feature. Thank you.

 

Does it really matter?

 

I feel it does.....

[Gorf]

AFAIK Albert approved this thread and Gorf obviously fel it important enough to submit for approval.

Everyone is entitled to their opinion as long as it is not harming another and respectful. If anyone does not feel that this is important to them then please don't read the thread. If anyone does not want to read what a member has to say then that is why there is an 'ignore user' feature. Thank you.

 

Does it really matter?

 

I feel it does.....

 

I appreciate your clear-minded response to this subject matter.

[carmel_andrews]

the question is, in order for Mr owl (and gorf) to be able to code using this technique they must have been shown the concept from somewhere (be it a website or atari text/doc somewhere) so, whoever was the originator of that text/doc is therefore the originator of this technique

 

since not many people know of this techique though, using checks and balances i guess it would be fair to say that Mr. Owl is a modern pioneer of this technique

[Starcat]

@Gorf:

I don't know the background of this thread, but I think it's good to try and correct misinformation that is out there and it's a shame people spread misinformation in the first place. It's just sad if people can't give credit to or appreciate great work. Credits to those who deserve it.

 

@carmel_andrews: I don't agree with you there. There are plenty things about the Jaguar which are not documented.

I'm not sure if the GPU main ram execution was ever done before. In fact I never heard of it working before Atariowl and Gorf started talking about it. The only thing I recall is a rumor that Dactyl Joust used the main ram to some point, but we don't know if they actually succeeded to use it like this or if they just wrote code without jumps.

 

If a coding tweek was done before, but it's never documented and thus not recreatable, it's quite worthless to the community.

Now if somebody comes up with a way to make it work by himself, he's definately the one who deserves the credits for it.

The way I see it Atariowl and Gorf are the pioneers of bringing this technique to the Jaguar community and I highly appreciate it that they even shared their information with the rest of us.

[Goochman]

Regardless of origin - thanks for all the breakthroughs - here's hoping more gets done with them

[carmel_andrews]

Starcat, you obviously didn't quite understand where i was coming from

 

Now i am guessing that there are repositories of jaguar programming documents for developers that were produced by atari and/or a third party (during the jaguars lifespan or at least until atari r/m'd with jts)

 

now i am sure that programmers/developers like Atariowl are in possession of these documents and from these documents can go about putting programming or coding theories and techniques into practice (such as GPU main code execution) i.e the actual concept doesnt exist within the developers documents etc, but various concepts/idea's within the various programmers/developers documents put together you arrive at new techniques simply buy putting various concepts and ideas together into a new form (based on the information/documents the likes of atariowl have to work with)

 

To using an example, say the classic newton and the apple falling on his head, now obviously newton didn't invest or discover gravity , it already existed, but by observing the way the apple fell on his head, newton was able to extrapolate the properties of gravity and it's effect on earthbound objects

 

To apply that to my point, whilst this 'concept' of gpu main code execution did not exist before the likes of atariowl/gorf started devising that technique, the various concepts and ideas that go into a technique like gpu main code execution did exist prior to this technique coming into existence

Edited June 22, 2010 by carmel_andrews

[The_Laird]

the question is, in order for Mr owl (and gorf) to be able to code using this technique they must have been shown the concept from somewhere (be it a website or atari text/doc somewhere) so, whoever was the originator of that text/doc is therefore the originator of this technique

 

since not many people know of this techique though, using checks and balances i guess it would be fair to say that Mr. Owl is a modern pioneer of this technique

 

There is no question to be raised and you really don't know what you are talking about Carmel with all due respect. The Atari docs are trash and I know for a fact there is nothing regarding this technique in them. I know Owl personally and have met and talked with him in person and have no reason to doubt anything he has told me and may I add he gave massive props to Gorf to first coming up with the tecnique. They found this because they know the Jag hardware better than anyone and studied it in detail just as other coders have over the years to get the best out of their respective machines. There are many examples over history of good programmers doing things with machines that nobody (including the hardware producers) thought possible such as avoiding attribute clash on the Spectrum, displaying 256 colours on the Genesis or one of many examples on the 2600 wish was pushed well beyond what Atari themselves thought possible.

[Gorf]

the question is, in order for Mr owl (and gorf) to be able to code using this technique they must have been shown the concept from somewhere (be it a website or atari text/doc somewhere) so, whoever was the originator of that text/doc is therefore the originator of this technique

 

since not many people know of this techique though, using checks and balances i guess it would be fair to say that Mr. Owl is a modern pioneer of this technique

 

 

Now how can you possibly say this? Surrounded! and Gorf Classic used this technique

long before anyone else made any such claim. In fact it was then after I made this

announcement that Owl contacted me and said he too may have found this workaround.

The rest is in the blog at his site. There was no such website or text document. If

you think so..please do find it and post a link. Make sure you know what you are posting.

As far as I know it does not exsist. I was the first one to bring this idea to the

attention of the community. I nor Owl would not be the first ones to discover a feature

or ability of a machine that was once thought to not be possible.

 

It is certainly possible that someone else before either of us found a way around the bug,

but they never told anyone. I assure you I independently invented the technique through

trial and error. There are no documents previous to my discovery and announcement that I

know of. I then compared notes with Owl, whom also independently discovered this technique

and we found out that we both had something to offer each other to 'perfect' the bug fix.

 

I think Starcat put it perfectly. I/Owl did not create anything...we discovered something

that was already there but just not known, or at least not publically. In fact my sole

motivation was to prove Atari was wrong and that there HAD to be some way around the bug.

Luckily for me, the bug is a hardware issue and consistant which makes the workaround

possible. In fact even before Surrounded, I discovered the workaround by writing a wireframe

tie fighter like demo to test its integrity. Surrounded was devised to make sure it was

stable, using it for a much more robust application..ie .. a fully operational game demo...

its actually a ful game and the release is 'locked' into demo mode.

 

One more addition..... Dactyl Joust indeed supposedly used main code, but from what I

understand, it was code that was still flipped into the local to run. Forgetting who

actually did that game, I will say it was High Voltage off the top of my head, supposedly

they wrote a C compiler that did the flipping of code stored in main to the local. Even so,

they (whomever they are) never let anyone know what they actually did or even how to do it.

All of this as far as I understand is hearsay. It would be great if any former HV guy could

come in and set the record straight on this.

Edited June 22, 2010 by Gorf

[+Random Terrain]

Now that this has been settled, it's time that Al Gore finally got credit for inventing the Internet.

[The_Laird]

Correct Gorf it was High Voltage who were doing Dactyl Joust, would be good to hear from them. Its long been rumoured that Carl Forhan has the code for it too.

[Gorf]

Now that this has been settled, it's time that Al Gore finally got credit for inventing the Internet.

 

Unfortunately for Al Gore, unlike Owl and I, he does not deserve that credit.

 

@ Gooch......

 

Im afraid you wont see much of that ...its been almost a year since that info

was released and so far the only two people making use of it are the same two

who discovered it. Not unlike the source code Curt released.....no one touched

it yet.

Edited June 22, 2010 by Gorf

[RevEng]

Does it really matter?

When someone gives a gift to the community like this, correct attribution of their hard work is a small price to pay. Further, its likely the only reward they'll get for their efforts, other than the satisfaction of the job itself.

 

It may not seem like a big deal to others, but they didn't spend their spare time working on it.

[+Random Terrain]

When someone gives a gift to the community like this, correct attribution of their hard work is a small price to pay. Further, its likely the only reward they'll get for their efforts, other than the satisfaction of the job itself.

 

It may not seem like a big deal to others, but they didn't spend their spare time working on it.

But that goes against everything Atari stands for. Someone invents something without getting credit for it, somebody else steals it one way or another, then the inventor finally gets credit for it many years later. Not enough time has gone by.

[Gorf]

When someone gives a gift to the community like this, correct attribution of their hard work is a small price to pay. Further, its likely the only reward they'll get for their efforts, other than the satisfaction of the job itself.

 

It may not seem like a big deal to others, but they didn't spend their spare time working on it.

But that goes against everything Atari stands for. Someone invents something without getting credit for it, somebody else steals it one way or another, then the inventor finally gets credit for it many years later. Not enough time has gone by.

 

Yes but let's not make the same mistakes Atari made. Plus Owl and I are not Atari...if that were the case

you could bet the 7800, Jaguar and ST line of computers would be in much better standing...or that would be

our(at least my) hope.

 

@ RevEng......thanks for the confirmation of my point. It's well appreciated.

[Fredifredo]

Hello ! I know who is the true inventor of GPU main code execution : Thierry Schembri , the main coder of Zzyorxx II ... in 1995

For more infos just ask him ...

[Gorf]

Hello ! I know who is the true inventor of GPU main code execution : Thierry Schembri , the main coder of Zzyorxx II ... in 1995

For more infos just ask him ...

 

Invite him in...I'd love to hear how he went about it. Too bad Zzyorxx II was never released( unless that happened

recently with all the proto's being let out into the community.) Having never seen the source code to Zzyorxx II,

and never seeing anyone, Thierry Schembri included, reveal it to anyone, Owl and I can still take full credit for

our discovery.

 

Anyone can execute GPU main RAM code. Even a good deal of Jag games do that. What they do not do as far as I know

is jump around in main RAM unassisted. We are talking about error free jumps in main RAM. Besides, the truth is no

one actually invented anything....they, like Owl and I, discovered it. It was always there and possible. It was

Atari that said it was not possible. I found it becuase I did not believe them. Owl was the first I know of to

jump between main and local ram.

 

Owl and I were the first to reveal it to the rest of the Jaguar community. I was the first

to officially release a game that employs it. There are also several ways to do it.

 

a ) The method spelled out by Owl and I...look it up...it's here and on Owl's blog site. This is unassisted GPU main.

b ) Using store instructions to the PC register works too...however you cant use condition codes with this one.

c ) Using the 68k to manipulate the GPU to jump from one spot to another.

 

The truth is I do not really care. All I am doing is correcting a few liars who are bitter and decided to be purposefully

innaccurate about it on their websites. Mainly because they are bitter for being caught in their lies and trouble making

plans within the Jaguar community. So the only way they feel they can save face is by painting me as a liar by further

lying about it. Saying I stole music for Mad Bodies which is as far from the truth as they are. Also they accused Terance

of stealing art work which is another bold faced lie.

 

The only issue they can make claim to is an oversite on Terance's part of leaving an unauthorized MOD in the game. This

was something I put my foot up Terance's ass for doing. Terance did not need to steal anyones music as I can write any

style of MOD for him. He chose to use this mod and others after I offered him original work. I wrote one mod for him that

WAS the title screen mod at one time...Terance then switched it in with the unauthorized MOD and moved my MOD to the High

score screen. My involvement with Mad Bodies was that one mod and sfx/voice effects. I never had much else to do with it

which is how Terance does all his games. He essentially hides all his games from me until they are close to release. That

policy has been put to an end because of his carelessness.

 

So even here these liars are just being bitter liars. The true purpose of this thread is two fold....one to reveal the

truth about who actually came up with accurate main code jumping and to exposes the liars for what they are....bitter

trouble making liars. They know who they are.

 

I believe the title of the MOD I gave to Terance to use as the title screen was called 'Who Knows' and he took it and

modified it to be more spooky than I set it up to be.

 

Here is a nother bold faced lie by this crew of thugs.

 

I would like them to back up with any evidence that Terance stole art work from a game that was NEVER released to the Jaguar.

(that I know of anyway.)I would also like them to point out this Japanese manga art he stole as well. At this point the entire

website is full of lies and liable. NO one bribed anyone with a free copy of Mad Bodies. That was an offer of good will along

with an offer of monetary compensation for each copy sold with that mod on it. But of course when you are hopeless liars you

tend to cover lies with even more lies.

 

Nothing of which they posted on their site has even the smallest bit of truth and they know it. However, they are fine with

playing on the ignorance of the Jaguar community to further their campaign of personal destruction.

[neo_rg]

The truth is I do not really care.

Just let it go then.

 

and stop with all the jibber jabber, this thread will only go one way, and that's locked.

[Gorf]

The truth is I do not really care.

Just let it go then.

 

and stop with all the jibber jabber, this thread will only go one way, and that's locked.

 

I did let it go...a while ago...many months now...However, certain liars using their website

decided not to let it go, therefore prompting me to have to set the story straight. I quite

honestly do not care if the thread gets locked. In fact I suggested to Albert that he do so

when I originally wrote it in the first place. If someone is going to spread liable about me

and 3DSSS, you can bet I will defend myself and us. Most of the community knows better

anyway and sees all the players for whom they really are.

Edited June 24, 2010 by Gorf

[neo_rg]

If you are going to spread liable about me and 3DSSS, you can bet I will

defend myself and us.

please edit this "you" to "people" as it implies that i am involved. the use of

one word in the wrong context may not seem much but it is to me.

Edited June 24, 2010 by neo_rg

[Gorf]

If you are going to spread liable about me and 3DSSS, you can bet I will

defend myself and us.

please edit this "you" to "people" as it implies that i am involved. the use of

one word in the wrong context may not seem much but it is to me.

 

Ah but I'm over reacting? Come now! However, I changed it to 'someone is'

to be clear. Happy now?

Edited June 24, 2010 by Gorf

[neo_rg]

yes. thankyou

[Gorf]

yes. thankyou

 

You're quite welcome.

Edited June 24, 2010 by Gorf