An alternative PS3 for similar cost... a better PS3?

[Silent_Buddha]

A better design would of been more memory rather than unifying it. It sounds obvious and yes it would cost more money, but many games could of looked better. Id be interesting to know how much it would of cost to give it 512meg of Vram

Well consider the following a 7900 GTX (G71 with 512 mb of GDDR3) launched at 499 USD. The same price as the 20 GB PS3 launch unit.

GDDR3 was at a premium at that time. So another 256 MB of GDDR3 could easily have been another 100+ USD.

At 499 USD and 599 USD the PS3 was already unattractive to most console buyers. 599/699 USD or more would have been even worse. Unless Sony were willing to eat even more of the costs in order to try to push BRD adoption.

Regards,
SB

 

[eastmen]

Sony's smart play would have been doing 256 megs of xdr and 512 megs of GDDR ram . Remove bluray to pay for it and come in at a much lower cost .

The 512 megs of vram would have produced games far more visualy impressive than what the xbox 360 was to have with a shared 512 megs

 

[TheWretched]

Removing Blu Ray was not an option for Sony at that time. Having Blu Ray was the WHOLE point of PS3.

I think, PS3 simply came out too early. It should've been delayed until late 2007, which would've enabled Sony to get a much better (G80 derivate) GPU and more RAM for a comparable price. Also, the games released in the first year were quite lackluster. Only until stuff like Uncharted was released (which was late 2007) it became interesting. Imagine the system launching with these games.

Though again, the pressure was on Sony to make BluRay win the "war"...

 

[eastmen]

Removing Blu Ray was not an option for Sony at that time. Having Blu Ray was the WHOLE point of PS3.

I think, PS3 simply came out too early. It should've been delayed until late 2007, which would've enabled Sony to get a much better (G80 derivate) GPU and more RAM for a comparable price. Also, the games released in the first year were quite lackluster. Only until stuff like Uncharted was released (which was late 2007) it became interesting. Imagine the system launching with these games.

Though again, the pressure was on Sony to make BluRay win the "war"...

Did bluray really counter all the losses that sony has made with the ps3 ?

As for a delay to 2007 , that would have given MS 2 years to build up a user base along with nintendo. Sure Sony could have put something better out but

1) HD-DVD would have won the war as its stand alone players were selling very well compared to bluray stand alones and the hd-dvd add on would have gained ground .

2) A 2 year old MS console means that by 2009/2010 MS could have simply released another console which would have been much better than the 2007 ps3 . Sony could have been forced into a wierd half cycle limbo hell in which nintendo / MS would allways be a generation ahead / behind them

 

[Lucid_Dreamer]

Sony's smart play would have been doing 256 megs of xdr and 512 megs of GDDR ram . Remove bluray to pay for it and come in at a much lower cost .

I don't see how this line of thinking makes sense. It cost over $800 to make a PS3 at launch. The Blu-ray drive was $125. Removing the Blu-ray drive doesn't even bring you back to $599. The Blu-ray component was the ONLY reason the PS3 was subsidized that much to begin with! Sony already made the smartest play they could with the information they were working with at the time. Remember this was pre-2005.

The 512 megs of vram would have produced games far more visualy impressive than what the xbox 360 was to have with a shared 512 megs

They seem to be doing that anyway, via the exclusives section, IMO.

 

[Acert93]

Removing Blu Ray was not an option for Sony at that time. Having Blu Ray was the WHOLE point of PS3.

I am pretty sure, somewhere somehow, there is a salient argument that Sony developing a successor to the most successful console devices (both over 100M sales) ever was a point, somewhere, in the whole "PS3" planning.

Call me crazy, but Sony didn't say, "We have this thing called Blu Ray. What should we do? Lets invent a device, lets call it the PS3, and we can get this device to consumers!"

In short, I think it may be sane, every slightly so, to argue a PS3 could have been envisioned and designed WITHOUT Blu Ray. Call me crazy, but it may have been possible.

Ps- The flip side of the coin is Blu Ray may have remained vaporware if it was not bundled with the PS3, but that is another point. If Blu Ray as a technology had never matured there would have been a PS3. End of story.

 

[TheWretched]

Well... it's not like Sony earns quite a lot of money through BluRay alone... and that step was needed, as patents for CDs went public domain around that time, too. So it was purely economical for them to force BDROM into the system, by any means. Without PS3, BDROM would've never won the "HD wars".

Of course they wouldn't have had to include BDROM into the system, which could've made it a lot cheaper to produce, but the other costs (i.e. lost BDROM royalties in case HDDVD went further) would've been even higher, probably.

So yeah, I still stand by my point that BDROM was a necessity for Sony to include in the PS3, by any means.

 

[Silent_Buddha]

I think, PS3 simply came out too early. It should've been delayed until late 2007, which would've enabled Sony to get a much better (G80 derivate) GPU and more RAM for a comparable price. Also, the games released in the first year were quite lackluster. Only until stuff like Uncharted was released (which was late 2007) it became interesting. Imagine the system launching with these games.

I'm not sure that would have been a feasible option. Something based on G80 would have slipped to 2007 at the earliest. But that isn't the problem. The problem is the power consumption and size.

Due to hardware put in to not only meet and surpass AMDs advantages at the time (AF quality being one of those) but also the huge investment in unified shaders + GPU compute made G80 far far larger and more complex than G71. So each chip is already going to cost significantly more. Add to that greatly increased power consumption (~50 watts at idle and 75+ watts at load) and we end up with a chip that is very unfriendly to a closed console. PS3 would have ended up being far louder than the X360 and even more expensive to manufacture.

G92, which addresses those problems (8800 GT) would have been within a similar although slightly higher power envelope, as well as much reduced die size although still not as small as G71 I think (don't feel like looking that up ). But then the problem is that the console would end up slipping into 2008. Although slightly slower than G80, the other advantages more than make up for it with regards to the console space.

For a quick and dirty non-specialized (relatively speaking) GPU, G71 is basically the best they could do unless they wanted to delay until 2008. Imagine how much stronger Microsoft's position would be if that had happened?

Regards,
SB

 

[AlNom]

A delay to holiday 2007 would have made a G84 (8600 series, 80nm, 169mm^2, 289MTransistors) derivative feasible. It's a 32SP, 16TexFilt/8TexAddress, 8 ROP design. Perhaps they could have designed a 48SP, 24TF/12TA, 8 ROP part.

An SP cluster cost roughly 79 MTransistors & 42mm^2 die space @ 80nm (judging by TechReport's analysis).

65nm would have hampered supply at that point.

------

G71 was ~191mm^2, but we eventually got an RSX that was 250mm^2 due to a number of redundancy measures, larger texture caches, and perhaps the XDR I/O.

Mind you, there'd probably be extra silicon for a G8x chip considering it had a separate ASIC for display output.

 

[Silent_Buddha]

I'm not sure G84 would have done much for them. It was already slower than G71. It may have been more capable (Dx10 and unified shaders) but would a slower but more capable chip have suited the PS3 better? Especially if it meant shipping 1 year later?

And if you turned on MSAA, performance drops even farther behind G71.

I see it as shipping 1 year later for roughly the same graphics performance.

Regards,
SB

 

[AlNom]

I'm not sure G84 would have done much for them. It was already slower than G71.

That's why I suggested a 3 cluster part. G84 is quite a bit smaller than RSX, but an extra cluster of SPs/TMUs should bring it back up to a similar die size.

ROP-wise, it's pointless to compare to G71 since RSX was gimped itself. More ROPs on top of my suggestion, may be too big.

And if you turned on MSAA, performance drops even farther behind G71.

Are you suggesting putting a full-on G71 into the PS3? I'm not sure why you're comparing to that instead of RSX.

All I'm suggesting is a replacement for RSX within the same die/power budget but on a new architecture. So in that sense, G8x would already be a big improvement with single-cycle 4xAA and other blending/higher precision capabilities. Ultimately, you're still going to be straddled with GDDR3 on a 128-bit bus and 64-bit XDR. Even with an extra year, I don't think they'd have been able to unify memory because nV would have been busy coming up with the G8x line-up. The extra year could have been spent further customizing G7x but it was a dead-end architecture by then.

Performance sucks for G84 because there were only 8 ROPs, but that's no different than RSX. Bandwidth-wise, G84 is similarly gimped as RSX. At least waiting another year would have allowed them to try and use higher clocked memory, though I'm not too sure the PS3-phat could have handled any more heat.

 

[Silent_Buddha]

Are you suggesting putting a full-on G71 into the PS3? I'm not sure why you're comparing to that instead of RSX.

Just because I'm PC centric and it's easier for me to consider things when I base it off hardware I'm more familiar with. No other reason.

Aye, a more capable G84 would certainly have been at best a lot better and at worst, no worse than RSX (there I used RSX instead of G71. ).

I'm just not sure shipping a year later than the already relatively late launch would have done the platform any favors over all.

They'd have given MS a 2 year lead on this cycle with not a lot to show for it. Platform exclusive games certainly would have taken advantage of it. But multiplatform games would have just, theoretically, reached graphics parity sooner. But a 2 year advantage in terms of not only developement time but public mindshare would have been hard to overcome, IMO.

I just can't help thinking that RSX was their backup plan and that they originally wanted a custom solution but either time (most likely), budget, or technical reasons (which plays into both time and budget) forced them to go with RSX.

Regards,
SB

 

[steampoweredgod]

I'm not sure but it does seem like they didn't ask around while developing the ps3. They should've checked in with the various vendors and seen what performance/price was offered, in the preceding years, while having a moving launch target.

Hopefully now that mobile has taken off drastically, with demands for energy efficiency from everyone, there are more options in the table to chose from.

 

[Silent_Buddha]

I'm not sure it was as much shopping around, but I have a feeling they originally wanted to do as much of it in house as possible (like PS2). But that parity with then current graphics technology was proving to be beyond their capabilities/budget. Cell and BRD was already pushing their developement and budget resources. Including XDR wasn't cheap either. So it's likely that as they saw the budget for PS3 ballooning they decided to cut their losses on an in house graphics solution.

Regards,
SB