Ecco il nuovo revenge:

As I said, I'd let you all know what I know. Well sorta, I won't let everything out, due to the possibilty of destroying lives (seriously). But as far as revenge, I'll let out what I can. Also, you must all keep in mind, this is all info I've gotten from Aqueel. So if SOMETHING isn't true, or accurate. Correct it, cuz this is how I know it.
 
Revenge, as we all know now from Aqueel's last saying, was going to be a Gfx chip. A revolutionary one at that. We had old 3Dfx tech, namly the stuff nVidia didn't get. That being Rampage. Rampage, by itself kicked ass. We had got a board made, from a company that will be left un-named, and it was crazy. Aqueel tested in in Quake 3, and @ 1280x1024 with max gfx @ 8xAA it churned out 254FPS. Now, if you ask me that's faster than anything out there. And that was about 8 months ago! Between then and now, we've made alot of revisions to that design. New memory architecture and so forth. Added some things that would make it even faster. It would've been the chip of chips!
 
It was of RISC design, and used very little power, thus putting off very little heat. I bleive, if memory serves me, it would've either not needed a HFS or it would've used just a passive HS. Then there was the Sage chip, or as we called it PHOENIX. We could've added numerous Revenge and Phoenix chips to one board, and it would've just been nuts. It would've obviously needed an external power source if enough were used, but with the right amount, it wouldn't have needed any. We were going to have a few types of boards. (R is revenge, P is Phoenix) 1R and 1P and then 1R and 2 P and then 2R and 2P and 2R and 4P. Crazy shit to say the least. Can you imagine, if a 1R 1P turned 254FPS, what would a 2R4P done LOL. "But it would've cost a fortune". Nope, due to how we had it all layed out, it would've been cheaper than anything out there. We were going to be going from $75-$275. Now the $275 I'm sure would've been the 2R2P but it might've been the 2R4P. Who knows :| Time to bust out documents I have and rip some info.
 
 
These are up to date with some very minor changes that aren't needed to be updated. (plus I really don't know what they are LOL)
 
Quote:96M+ Triangles/s
0.13 micron 8 w power consumption with approx 36 million transistors.
3200MP/s fill rate
32 fully featured pixels per clock.
128 bit internal Precision with 64 bit Zbuffer support.
Hardware Deferred Rendering Capability (Tiling).
1024 bits wide Memory Crossbar divided into 4 256 bit pathways.
96 Texel cache Stages upto 16 shared rendering pathways with 2tmus per pipe.
32 layer multitexture in a single pass.
Upto 1000 K + triangles per frame.
Capable of 1600 Mega pixels in a single pass with 16 textures, Trilinear Anisotropic Filtering and 24 bit textures with 4x Multisampling.
Free Anisotropic Filtering with AA enabled.
Upto 256 tap Anisotropy possible.
Hardware Displacement Mapping and full DX9.1+ Capability.
Upto 78GB/s fill rate.
Full color compression and Bandwidth saving capabilities.
Multi-sample anti aliasing (Gamma corrected and rotated grid jittered samples) upto 16 samples possible.
Super Sampling upto 8X jittered grid.
Full support for cinematic factions such as motion blurs soft reflectance and soft shadows.
Off Chip geometry co-processing Industry innovation in TNL solution
Twin Risc TNL engines for Advanced pixel/Vertex shading calculations  
TNL units hyper threaded for optimal CPU like performance.
PS /VS Shaders 1.0 to 2.0 + specs  
Full DXTC 1 to 5 Support with FXT1 support in Hardware.
Fully microcode updatable Architecture.
500 + MHZ internal and DDR clock.
 Multichip Capable Architecture MCA.
Upto 4 chips in MCA. 9.6 GPixel/s possible.
AGP 8X AGP 3.0 Spec.
32MB to 256MB frame buffer  
 

 
 
Nice huh?  
 
(part 2 to come, stupid MSG board wont allot large posts. How gay)  
  Logged
 
 
 
MaxPain
3dhq Brood Member
New Member

# 783




Dickhead, dickhead, dickhead.... CHECK IT OUT!

       


Gender:
Posts: 24
  Re: REVENGE..... On the Gfx Card World!!!  
« Reply #1 on: Today at 2:43pm »  Quote  Modify  

--------------------------------------------------------------------------------
This is between Aqueel and one of the Investors:
 
Quote:HARDWARE TECHNOLOGY LICENSING AGREEMENT
 
This agreement is made as Dec 28, 2002 by and between 3DHQ ("3DHQ" and its Business source provider Schmidt Co For the amount of  USD 97.5 Million , Ninety Seven and a Half Million dollars.
 
WHEREAS:
 
3DHQ develops  markets computer and console graphics hardware and
Licensee wishes to license such technology on the terms of this license.  
 

 
 
They must have thought it was a good idea too =}
 
 
Here's some patents Aqueel filed for. One of these is Bone's I'll let him go indepth on it if he wishes.
(I'm leaving out Patent #s and other info)
 
Quote:Multiplexed synchronization circuits for switching frequency synthesized signals
 
Multiplexers are used to generate synchronized slave clocks from a common master clock. A first multiplexer and a second multiplexer generate a first slave clock and a second slave clock, respectively, from the common master clock. A third multiplexer and  
a fourth multiplexer are configured as a divide-by-n circuit for providing a third slave clock that is a divided version of the second slave clock. A fifth multiplexer provides a matching delay to preserve the synchronization between the first slave clock and the other slave clocks. A sixth multiplexer is used to select between the second slave clock and the third slave clock in response to a select signal. A flip-flop may be used to provide the select signal and to guard against false selection of slave clocks  
 

 
Quote:Content addressable memory with an internally-timed write operation
 
A content addressable memory with an internally-timed write operation includes a data input for receiving a input word. Coupled to the data input are a plurality of storage registers comprising stored words. Each storage register includes a comparison circuit for comparing the stored word with the input word and producing there from a match output indicating a match when the stored word matches the input word, and indicating a miss when the stored word does not match the input word. Coupled to the storage registers is a miss detector for generating a miss signal responsive to each of the match outputs of the storage registers indicating a miss. Coupled to the miss detector is a write cycle circuit for writing the input word to at least one of the storage registers responsive to receiving the miss signal  
 

 
Quote:Command data transport to a graphics processing device from a CPU performing write reordering operations
 
A system and method for enabling a graphics processor to operate with a CPU that reorders write instructions without requiring expensive hardware and which does not significantly reduce the performance of the driver operating on the CPU. The invention allows the graphics processor to evaluate the data sent to it by software running on the CPU in its intended and proper order, even if the CPU transmits the data to the graphics processor in an order different from that generated by the software. The invention works regardless of the particular write reordering technique used by the CPU, and is a very low-cost addition to the graphics processor, requiring only a few registers and a small state machine. The invention identifies the number of "holes" in the reordered write instructions and when the number of holes becomes zero a set of received data is made available for execution by the graphics processor.  
 

 
Quote:Circuit for processing RISC instructions for 96 bit Pipeline texture stage. Vertex and Pixel Shader Implementation routines 2.0+
 
An apparatus and method for prioritizing interrupt requests in a RISC processor. By utilizing hardware to prioritize the requests, processor time is reduced. The acknowledge signal (Ack(0)-Ack(n)) from a priority resolve circuit (50) selects the given service routine entry to branch instruction generating circuit (56, 5. A lower priority service routine can be interrupted by a higher priority reply. Circuit Accelerates Video Bit signals by dividing itself into multiple synchronized Shader arrays, the shader array is 196 bit precision with 32 alpha channels. Speed tessellation allows RISC processor to emulate all pixel shader versions from 1.0 to 2.0+ HLSL AND CG capable.  
 

 
Quote:EDRAM based internal Hardware Bus with a 1024 bit Memory cross bar divided into 128 divisors.
 
Internal Revenge XC-1 architecture utilizes a 1024 bit internal compression bus. Internal compression allows multiple ACKS numbering to a million per byte to be transferred via complete switching multiplexing. Memory data is compressed up to 24:1 internally with a single chunk size exceeding 1 megabyte/ns. This allows for hardware based compression and faster color compression and Z buffer Routines.  
 

 
 
There you have the majority of tech descriptions. As you can see from above, we were going to use EDRAM. Which would've really made that thing scream. Well.... My brain is fried now lol. Um, I guess ask Qs and me or others will answer them as best, or as most as we can. Remember, don't ask Qs about who else helped us, cuz we won't answer that. It could lead to a mess up in thier life.  
 
 
 
Aleeeeeeeeeeeeeeeeeeeeeeeeeeeeee!!!!!!!!!

Ancora non ci credo

FESTEGGIAMO!!!!!!

Aleeeeeeeeeeee!!!!

nn ha detto che sta x uscire...

Qualcuno traduca, io non riesco a tradurre il tutto!!!

qualkuno traduca infretta!!!!

Nelly

bene ragazzi ora conoscete tutti i dettagli !!!
Posso solo confermare che questo è il revenge , sulla carta




Scritto Da - amigamerlin on 02 Aprile 2003  22:47:48

Max Pain, come promesso, ha postato tutto quello che sapeva.

Riassumendo: hanno 2 chip.
Il Revenge e il Phoenix.
L'idea era di presentarsi con diverse fasce di prodotto: scheda con singolo Revenge, scheda con 1 Revenge e 1 Phoenix, fino ad una scheda con 2 Revenge e 4 Phoenix.
Il chip dovrebbe essere basato su tecnologia 3dfx, nominalmente, tutto quello che nVidia NON ha preso.
Seguono caretteristiche tecniche.

Poi segue quello che mi pare sia un documento notarile, un contratto tra 3dhq e Schmidt Co per 97,5 Milioni di dollari per, credo si tratti di, brevetti tecnologici.

Non c'è scritto che la scheda si farà, non è scritto che la produzione o solo la progettazione sia in corso.

Ho parlato su ICQ personalmente con Max Pain domenica scorsa, prima che pubblicasse questo post e mi ha confermato che il Revenge è fermo per mancanza di fondi.

Che insomma, non ci sarà nessun Revenge, perché nessuno paga per produrlo.
97,5 milioni di dollari sono una cifra esorbitante, pari a 200 miliardi di vecchie lire.
All'inizio del 2 trimestre del 2003, nessuno, e sottolineo NESSUNO, investirà mai cifre simili in un progetto hardware.
Vi risparmio la sviolinata sulla crisi economica globale, ma solo perché non ho voglia di dilungarmi: la crisi c'è ed è pesante.
200 miliardi se uno li ha se li tiene stretti.

Cercherò di ricontattare Max Pain a breve (il fuso orario non gioca a favore) e gli chiederò ulteriori spiegazioni.

In allegato a questo post, il TXT del discorso tra me e Max Pain, datato 29 Marzo.





Scritto Da - Khana on 02 Aprile 2003  23:18:16

Ok, ho contattato Federelli e Max Pain.

Dunque... entrambi chiedono di essere citati all'interno di discussioni riguardo ai loro post e Max preferisce che d'ora in avanti siano postati i link diretti alle discussioni 3dhq invece di copiare e incollare il testo.

Posto qui, d'accordo con Federelli che lo ha reso pubblico sul forum, il file zip con la presentazione del Revenge (non l'ho ancora guardata neppure io, perché non ho office sul PC di casa e devo aspettare domani al lavoro).



Ora si può parlare seriamente di qualcosa.

Questo il tread da cui ho attinto
http://www.tdhq.net/cgi-bin/yabb/YaBB.pl?board=5;action=display;num=1049312536



Scritto Da - Khana on 02 Aprile 2003  23:45:02

Io non posso che essere felice nel conoscere che cos'è il REVENGE, anche se per ora nessuno ci messo i soldini per farlo...


Ma se questo non è un pesce d'aprile io sono tristissimo:

http://www.tdhq.net/cgi-bin/yabb/YaBB.pl?board=5;action=display;num=1049303872

Aquoes sta morendo...

Federelli e Max hanno entrambi confermato.
Sembra non sia un pesce d'Aprile.

Per rispetto a Seyd, chiederei per favore di non postare commenti al riguardo. Poi fate come volete...