You don't seem to know what you're talking about.
Both the original 3DO and Jaguar flopped so it doesn't take a genius to see the successors also flop. The SuperGrafx didn't really flop. Yes it sold far fewer numbers than the PC-Engine but it was never planned as a mass market console given that it sold for around $500 compared to the PC-Engine which was like $100. Saying the SuperGrafx flopped is like saying the original Neo Geo flopped.
If MS releases a XBO V2 it would cost about the same as the launch XBO did. Also they would drop the price of the original XBO to $300 when V2 is released.
Here's an idea. I don't know how technically or financially feasible it may be, so feel free to slap it down on those counts. If I remember correctly, the 360 has a separate chip, with a high bandwidth connection, that contains the EDRAM and some logic. While not as much as developers would have wished for it did offer the option of a single line of code to turn on AA and walk away.
Is there any reason MS could not offer a premium model with an extra chip, similar in concept to the EDRAM chip on the 360, that would simply crank up the AA, AF and whatever other bells and whistles that are not dependent on rewriting a code base substantially? I think it was one line of code for the 360, literally AA on or something like that. Not hard to include with updates that games already receive. There are going to be redesigned boards for the One at some point anyway. So the premium models get this ability and the others run entirely normal.
Technically possible? Too difficult? Too expensive? Not worth it? I was mostly thinking of incentives for people to not switch. Some are simply tribal, others want to keep their existing friends list, etc. Offer them a reason to stay. As in, it may only run in 900 but I get TSAA and 16xAF. Etc, etc. If the model sold for 100$ more with a hybrid SSD (say 100gb SSD and a 1TB hdd) it wouldn't be that different from what has been done before. (The size of that drive was pulled from thin air, basically whatever budget is left over after accomplishing enough with the chip.) I think the N64 had memory you added to the controller to play 4 person Goldeneye. Bigger/ better hdd's are a fairly established upgrade at this point.
If it didn't fragment the code base and was simply part of an established premium concept I don't see why it couldn't work.
Is there any reason MS could not offer a premium model with an extra chip, similar in concept to the EDRAM chip on the 360, that would simply crank up the AA, AF and whatever other bells and whistles that are not dependent on rewriting a code base substantially? I think it was one line of code for the 360, literally AA on or something like that. Not hard to include with updates that games already receive. There are going to be redesigned boards for the One at some point anyway. So the premium models get this ability and the others run entirely normal.
Technically possible? Too difficult? Too expensive? Not worth it? I was mostly thinking of incentives for people to not switch. Some are simply tribal, others want to keep their existing friends list, etc. Offer them a reason to stay. As in, it may only run in 900 but I get TSAA and 16xAF. Etc, etc. If the model sold for 100$ more with a hybrid SSD (say 100gb SSD and a 1TB hdd) it wouldn't be that different from what has been done before. (The size of that drive was pulled from thin air, basically whatever budget is left over after accomplishing enough with the chip.) I think the N64 had memory you added to the controller to play 4 person Goldeneye. Bigger/ better hdd's are a fairly established upgrade at this point.
If it didn't fragment the code base and was simply part of an established premium concept I don't see why it couldn't work.
yes, what funFor the fun of it.
Do you remember how many years and thousands of posts people made about the 'possible/likely' specs of the ps4/xbone in the run up to their release. Of course 99% were wrong, 90% were just wishful thinking.
I was hoping we were gonna have at least a couple year breather before these posts showed up.
But now it looks like 2 years of xbone1.5 specs (and when these ppl realize in 2016 theres no xbone1.5) well what will the xbtwo look like in 2018, oh what fun
MS should prepare for 1 big, final, channel stuff, like retailers have never seen before; just before they sell the division:
So while buyers are sitting in a boardroom; MS announces
So the investors buy it, and while walking home they see huge traffic pileups. As they move closer they see that the pileups are in front of electronic stores. a huge amount of green boxes is stacked so far outside of the store, it blocks the walkway and even the streets. Upon inspection, the boxes read XBOX ONE Day ONE EDITION
Then the new investors go:
But seriously, that is the best MS can do: increase production, fake the brand value, and sell everything
So while buyers are sitting in a boardroom; MS announces
So the investors buy it, and while walking home they see huge traffic pileups. As they move closer they see that the pileups are in front of electronic stores. a huge amount of green boxes is stacked so far outside of the store, it blocks the walkway and even the streets. Upon inspection, the boxes read XBOX ONE Day ONE EDITION
Then the new investors go:
But seriously, that is the best MS can do: increase production, fake the brand value, and sell everything
Megadrive1988
Veteran
The last gap was 7 and 8 years, not 9.
PS3 (2006) -----> PS4 (2013) = 7 years.
Xbox 360 (2005) -----> Xbox One (2013) = 8 years.
Ubisoft and EA are hoping and expecting, respectively, that this console cycle isn't another 7-8 years but more like 5-6.
Links:
Ubisoft: Don't wait 7 years to launch PS5
EA expects to see a new console generation in 5-6 years time - Xbox One and PS4 successors to roll out in 2018/2019, exec predicts.
Also:
Sony UK thinks PS4 lifecycle might be shorter than PS3
Last edited by a moderator:
LightHeaven
Regular
Yeah, at 1,088 Gb/sec!
If ESRAM2 in Xbox Two was crazy high bandwidth like than then you can probably always find novel ways to make the hardware dance that's virtually impossible on a conventional hardware.
Is that design even a viable alternative for a current gen console?
That made sense in the 360 era, but for a current gen machine having only the ROPs tied to the fast pool of ram (and keeping them really simple like removing compression so they don't get too bigger) wouldn't make much sense. Having the entire gpu being able to access the small pool of ram opens up many more possibilities now that the gpu can use it with compute shaders.
Unless somehow sony could managed to have the entire gpu access the edram with that crazy bandwidth, but I doubt it.
It could certainly have happened;
But not for PS4; Sony could not have another PS "It's so hard to code, sony is teh suxx0rs!! -signed many devs and published on Kotaku, 1UP and other similar websites daily", 3 on their hands.
MS could have certainly went with it; but they did not have the architects, nor the skills in house to make such a thing happen. So it could have been produced for the same price as the current Xbox... 'marvel', but yeah with about 2 times the performance, if used by extremely talented, low-level developers; which MS does not have access to either.
But not for PS4; Sony could not have another PS "It's so hard to code, sony is teh suxx0rs!! -signed many devs and published on Kotaku, 1UP and other similar websites daily", 3 on their hands.
MS could have certainly went with it; but they did not have the architects, nor the skills in house to make such a thing happen. So it could have been produced for the same price as the current Xbox... 'marvel', but yeah with about 2 times the performance, if used by extremely talented, low-level developers; which MS does not have access to either.
Completely unfounded remark. I'm sure if MS wanted a small pool of 1 TB/s eDRAM, they could have made it happen. the choice for ESRAM over EDRAM was about sourcing a supply. eDRM would have had production consequences which is why they didn't go that route.
2 times, eh? How did you calculate that?
you have to read the second part as well:
"if used by extremely talented, low-level developers; which MS does not have access to either."
MS wanted the lowest price; which is why they went with DDR3+ESRAM; not from a performance standpoint, merely a price standpoint. They even went as far as claiming they deliberately did not target high-end graphics in an interview.
The current Xbox One design IMO reflects the skills of its architects; it's just a box box with a big fan. Also looking at the motherboard itself you can see the layout was just functional; nothing more.
I believe a Sony proposed EDRAM system could certainly be developed and still be competitive with the Xbox One BOM; but I just don't see it happening by MS itself.
Also; in the case of an 'exotic design'; looking at PS3 for example; only the most talented first party- devs could exploit that system fully in the end. You had the exclusives really standing out; The Last of Us in 1080P almost passes for a proper PS4 title.
Again; MS does not own studios which have those capabilities; so an exotic design would be suicide for them; multiplatform titles would suffer and exclusive games would look about the same.
In this hypothetical case, we don't know how "small" it would need to be to allow 1TB/s, and we don't know how limited it would have been (directly addressable? only usable for some special instructions?). Maybe it would be completely different from the current ESRAM in XB1.
But I still wonder what is the real technical explanation why the ESRAM is so slow. Many GPUs have external memory faster than this.
But I still wonder what is the real technical explanation why the ESRAM is so slow. Many GPUs have external memory faster than this.
It isn't really slow though, is it? Memory above 200GB/s is pretty rare, no?
It seems to fit the purpose and isn't a bottleneck, but it's certainly not what we thought it could be when we speculated about this. It could have been a special thing like the 360's EDRAM, or the PS3 Cell. But it's only compensating for the DDR3.
I mean we would expect on-die memory to be several times faster than external, in that sense I think it's "slow".
A $300 midrange GPU card is about 200GB/s, right?
I think you can get cards with 300GB/s at that price range?
The PS4's 8GB are 175GB/s, in that sense the 32MB ESRAM barely seems to make a dent. There are latency advantages of course ... But as with everything, the bigger question always remains what the real bottleneck is for a specific game, both technically and in terms of the programming efforts required.
The PS4's 8GB are 175GB/s, in that sense the 32MB ESRAM barely seems to make a dent. There are latency advantages of course ... But as with everything, the bigger question always remains what the real bottleneck is for a specific game, both technically and in terms of the programming efforts required.
Whoever codes the games, how did you calculate that n MBs (where n is an unknown) of 1 TB/s eDRAM instead of 32 MBs ~200 GB/s ESRAM would result in a machine with twice the performance?
I don't know that much about embedded memory. Just curious what else is out there that's faster. There are definitely some GPUs out there with RAM that's closer to 300GB or higher, but that requires a really fast memory clock, which would be a problem in an APU with embedded memory.
Looking at prices, anything over 200GB/s using Nvidia seems to be $350+
Similar threads
