In @digitalwanderer 's "What brought you to B3D?" thread, someone had their first post quoted and it made me go back and find my first post. It was in the console forum on the subject of Nvidia's involvement in development of the PS3 (back before we knew exactly what that was going to be). It was a bit off the mark.
I thought it might be fun to have a thread where we dig up either our literal first post, or our first substantive post and see what we were thinking about at the time.
https://forum.beyond3d.com/members/digitalwanderer.327/
Well, if you look at the architecture if the PS2's Emotion Engine vs. The XBOX GPU (or any PC GPU for that matter), you can see where an enhanced version of the EE could benefit from the inclusion of Nvidia tech. The architecture of the PS2 is such that it has massive memory bandwidth available but only to a relatively small ammount of memory. This means that you need to limit the ammount of data you have to handle at any one time. This requires a developer to code around this limitation by shuffling the data around to different parts of the system (wherever some free memory can be found). Obviously, this makes the PS2 more difficult to optimize for. This also contibutes to one of the main reasons that PS2 games are, in general, graphically inferior to XBOX games. The PS2 just can't accomodate as high quality textures as the XBOX can.
So suppose that Sony decides it wants to address this limitation, but still build on the strengths of the EE design. They decide that rather than do their own R&D to develop a crossbar memory interface to GDDR4 memory they tap Nvidia who is going to be working on this anyway for its GPU's. Now Sony can possibly still include embedded DRAM like on the original EE, yet still have access to a larger memory pool when required.
This is just one possible way that Sony could be utilizing Nvidia tech in their PS3 GPU. Obviously, we don't know what the extent of the cooperation is, but Nvidia do have a lot of talented engineers. It may just be that Sony felt that 2 heads are better than 1 and the combnation of their engineers with Nvidia's could create a superior end product than what they could create on their own.
I thought it might be fun to have a thread where we dig up either our literal first post, or our first substantive post and see what we were thinking about at the time.
https://forum.beyond3d.com/members/digitalwanderer.327/