BitBoys
- Thread starter Quitch
- Start date
Yeah, but isn't that--like--their first piece of functional, purchasable silicon since their founding, despite going through many phases of enormous hype about completely different-natured projects over the years? 
cthellis42 said:Yeah, but isn't that--like--their first piece of functional, purchasable silicon since their founding, despite going through many phases of enormous hype about completely different-natured projects over the years?
Beware of the Savitaipale Revenge ™
Nappe1 may be grinding his AXE
rubank said:cthellis42 said:Yeah, but isn't that--like--their first piece of functional, purchasable silicon since their founding, despite going through many phases of enormous hype about completely different-natured projects over the years?
Beware of the Savitaipale Revenge â„¢
Nappe1 may be grinding his AXE
just read it from my signature!
rubank and keegdsb:
if you really are interested Bitboys history from the begining of this board to up to date, just use search with my nick as poster and "bitboys" as a keyword. you know I am not forever dumping the information on your laps while you sit and wait fingers crossed. it's time to learn using search!
maybe someday, I'll find some time and write "The Ghost Company Story: Another point of view." But until that day, you have to learn how to use search.
EDIT: On a second thought, ( yeah I know, the worst mistake is having second thoughts...) you will be getting something when I am back on my own computer and I have gotten few fans changed to new ones. (bearings are practically gone from my MiproCool fan. it sounds like... umm.. 5800 Ultra.
)Quitch said:
You can't can something that didn't exist.
Bitboys has never been able to send a working product to any company or developer. And I mean never. It's incredible how they still got backing from investors to produce these magical chips.
I guess people are gullible.. Can we say Phantom?
I think the search function is broken.
Rubank and Keegdsb, can you remember when I talked about a day that I would own something that does not exists?
well here it is! just like Waltar proved these chips can't exist and still I have 2 of them
so, here you go, two different revisions of chip that doesn't exist. Anyone that want to see them In Live, can come to Kotka, Finland and look me up.
and yes it is the AXE. A-eng is non-working due missing layer (Inifineon screwed that one) and B-eng is claimed to be fully working. (and because of that it is kept in antistatic place for just in case. I never believed to get even chip, so getting a card, as impossible it seems right now, isn't actually impossible, just extremely unlikely.)
This is the closest you can get to the real Bitboys hardware. (or perhaps if you have cellular phone with NEC's LCD panel, maybe that's closer, because your phone may already have Acceleon G10 on it...)
I have seen 2 different PCBs (beta and reference) for AXE and reference was seen live. other features that AXE (acording to my sources) supported included Video Textures, Capability run Video Stream through pixel shader and dual chip implementation. Memory controller unit was heavily updgraded from Glaze and had 11 Channels. Still, the way Axe rendered things has some similarities to Glaze. it had tiled frame buffer and in dual chip mode each chip would have been rendering tiles. simultaneusly. this way only tile had to be able to fit in eDRAM, not the whole frame buffer. (don't ask how it did polygon sorting or did it before decising which polygon was rendered on which tile...) The schedule for Glaze3D to AXE updgrade was VERY ambigious: decision to continue development happened around Feb-March in 2000 and first samples came back from Infineon around Summer 2001. Transistor budget more than quadrupled. if AXE would have been released in autumn 2001 as planned, it would have most likely been the king of the hill with pretty much equal perfomance with Ti4600 that came in Feb 2002. AXE was planned to be able run in dual chip mode, but not even PCB was designed for that. Most likely it was a back up plan if competitors would have been able to beat single AXEs performance. Problems with chip on Bitboys as well as Infineon's side caused the chip delay and the last straw was Infineon's decision shut down the eDRAM line up. Eventually this actually happend a bit earlier than it actually leaked to net. based on things I know AXE's follower plans, it must have been known at Noormarkku as early as September 2001. it is around that time when I first heard about codename 'Hammer' as well as first rumours were flying about Infineon pulling off. (though those were declined back then.) on March 2002, Bitboys were in very bad financial shape. And that situation actually developed something that now known as Acceleon G10. again off the record sources tell that G10 basic construct was designed by single guy during his winter vacation. this small core was then prototyoped and demoed to some company that interested at once and deal was made about Bitboys developing G10 for them. I don't know which company it was but I assume it was NEC. at the Assembly 2002, I was officially off the record confirmed that Bitboys has a deal with NEC, but it took all way to NEC's announcement about they using G10 instead of dummy driver on their LCD displays before it became official deal.
I haven't seen Pyramid3D card, but I know that there was pretty big stack left them when Tritech went bankcrupt. I also know what it could have been capable of and it's only problem was too much on too early. heck, VS 3.0 could have been almost supported on hardware level. only thing preventing full implementation would be the fact that P3D used fixed point calculation on vertex processor, instead of floating point. program lenghts, memory read / write space, conditional jumps, etc were all supported.
so, why they didn't license AXE stuff to other underdogs like Matrox? where are they now? what's the future?
I'll quote myself from another thread...
here's the stuff for now, hopefully you can make a fine flamewar out of it.
eSa: I'll let you know when *you know what group of ppl* is meeting at Tampere Area... I bet you are enthuastic to see those taffel chips as well.
well here it is! just like Waltar proved these chips can't exist and still I have 2 of them
so, here you go, two different revisions of chip that doesn't exist. Anyone that want to see them In Live, can come to Kotka, Finland and look me up.
and yes it is the AXE. A-eng is non-working due missing layer (Inifineon screwed that one) and B-eng is claimed to be fully working. (and because of that it is kept in antistatic place for just in case. I never believed to get even chip, so getting a card, as impossible it seems right now, isn't actually impossible, just extremely unlikely.)
This is the closest you can get to the real Bitboys hardware. (or perhaps if you have cellular phone with NEC's LCD panel, maybe that's closer, because your phone may already have Acceleon G10 on it...)
I have seen 2 different PCBs (beta and reference) for AXE and reference was seen live. other features that AXE (acording to my sources) supported included Video Textures, Capability run Video Stream through pixel shader and dual chip implementation. Memory controller unit was heavily updgraded from Glaze and had 11 Channels. Still, the way Axe rendered things has some similarities to Glaze. it had tiled frame buffer and in dual chip mode each chip would have been rendering tiles. simultaneusly. this way only tile had to be able to fit in eDRAM, not the whole frame buffer. (don't ask how it did polygon sorting or did it before decising which polygon was rendered on which tile...) The schedule for Glaze3D to AXE updgrade was VERY ambigious: decision to continue development happened around Feb-March in 2000 and first samples came back from Infineon around Summer 2001. Transistor budget more than quadrupled. if AXE would have been released in autumn 2001 as planned, it would have most likely been the king of the hill with pretty much equal perfomance with Ti4600 that came in Feb 2002. AXE was planned to be able run in dual chip mode, but not even PCB was designed for that. Most likely it was a back up plan if competitors would have been able to beat single AXEs performance. Problems with chip on Bitboys as well as Infineon's side caused the chip delay and the last straw was Infineon's decision shut down the eDRAM line up. Eventually this actually happend a bit earlier than it actually leaked to net. based on things I know AXE's follower plans, it must have been known at Noormarkku as early as September 2001. it is around that time when I first heard about codename 'Hammer' as well as first rumours were flying about Infineon pulling off. (though those were declined back then.) on March 2002, Bitboys were in very bad financial shape. And that situation actually developed something that now known as Acceleon G10. again off the record sources tell that G10 basic construct was designed by single guy during his winter vacation. this small core was then prototyoped and demoed to some company that interested at once and deal was made about Bitboys developing G10 for them. I don't know which company it was but I assume it was NEC. at the Assembly 2002, I was officially off the record confirmed that Bitboys has a deal with NEC, but it took all way to NEC's announcement about they using G10 instead of dummy driver on their LCD displays before it became official deal.
I haven't seen Pyramid3D card, but I know that there was pretty big stack left them when Tritech went bankcrupt. I also know what it could have been capable of and it's only problem was too much on too early. heck, VS 3.0 could have been almost supported on hardware level. only thing preventing full implementation would be the fact that P3D used fixed point calculation on vertex processor, instead of floating point. program lenghts, memory read / write space, conditional jumps, etc were all supported.
so, why they didn't license AXE stuff to other underdogs like Matrox? where are they now? what's the future?
I'll quote myself from another thread...
here's the stuff for now, hopefully you can make a fine flamewar out of it.
eSa: I'll let you know when *you know what group of ppl* is meeting at Tampere Area... I bet you are enthuastic to see those taffel chips as well.
Yeah, but um... Were they in any products? That sold? With any kind of volume?
People have Voodoo5 6000's as well, but that hardly qualifies. Heck, the 5800 almost doesn't!
I certainly imagine they produced some silicon, but did they ever bring about product? Was there performance measuring of the chips in a whole device, rather than guesswork from paper stats or testing chips and extrapolating, without really touching the software later? (We all have seen how well THOSE work.)
People have Voodoo5 6000's as well, but that hardly qualifies. Heck, the 5800 almost doesn't!
I certainly imagine they produced some silicon, but did they ever bring about product? Was there performance measuring of the chips in a whole device, rather than guesswork from paper stats or testing chips and extrapolating, without really touching the software later? (We all have seen how well THOSE work.)
cthellis: http://www.acceleon.com
