The controls need to be customized off the bat, but I definitely would not say a gamepad is better than it. Turning is just too slow with sticks.
Nintendo's hardware philosophy: Always old, outdated tech?
- Thread starter Tahir2
- Start date
Only Yu Suzuki had nothing to do with the Shenmue 2 port to Xbox, it was all Microsoft.
Microsoft did brag about portability and alot of PC game devs made alot of promises, yet they never kept them and worst of all WinCe was what opened the door to boot discs and piracy.
Would'nt it make sense, extremely good sense if say AMD/ATI was much more aggressive with Nintendo to try to contract the Radeon 5870 based console GPU for a 28nm engineering process (down from 40nm) or also have been very aggressive with insisting on the Radeon 4890 GPU also on 28nm (down from 55nm)
Basically I estimate that if a graphics card were to be produced that is based on the 4890 it would if fabbed at 28nm be a single slot graphics card clocked higher but not needing even a single six pin power plug while I estimate the same for a 5870 card only it needing a single six pin power plug like the Geforce GTS 250.
Basically I estimate that if a graphics card were to be produced that is based on the 4890 it would if fabbed at 28nm be a single slot graphics card clocked higher but not needing even a single six pin power plug while I estimate the same for a 5870 card only it needing a single six pin power plug like the Geforce GTS 250.
D
Deleted member 13524
Guest
By now, the probability of Nintendo adopting something as remotely powerful as a RV770 is almost none.
Let alone a Cypress.
As far as making sense, IMO a Fusion is what would make sense, or a Fusion + a discrete GPU.
But Nintendo is mimicking the X360 with a 3-core PPU-esque CPU.
Let alone a Cypress.
As far as making sense, IMO a Fusion is what would make sense, or a Fusion + a discrete GPU.
But Nintendo is mimicking the X360 with a 3-core PPU-esque CPU.
Agreed I expect the GPU perf to be a bit higher than hd 4670 (more SIMD arrays but lower clock).
Well I realized that there are a lot confusion in semantic, I for example use fusion not as the AMD brand but describe a CPU and the GPU put on the same die. SnB, last 360 chip is "fusion" for me.
I don't think AMD would prevent Nintendo to put a R7xx derivated on the same die as a POWERPC CPU as long as they don't use the brand "AMD fusion".
May be I (and may be other) should be clearer on the matter like "Fusion" "fusion like"? Actually is there a proper and convenient word to describe a chip that includes the CPU and the GPU?
D
Deleted member 13524
Guest
I'd say a lot less than HD4670. Less shaders to make way for eDRAM, while keeping aproximately the same amount of transistors.
That should allow better BC to Wii titles, or even running them in 720p + MSAA, and a reduced learning curve for X360 developers.
Given that, during the transition from R6xx to R7xx, ATI found a way to significantly reduce the amount of transistors needed for each shader processor, I wouldn't be surprised if the Stream's GPU has few\no more transistors than Xenos, while being a bit more efficient and carrying more shaders.
Yap, it's APU (Accelerated Processing Unit). It's what AMD calls it, but Intel didn't adopt it, nor did Microsoft with the XCPU.
I don't agree with that. My belief is that N made the Wii a huge but cheap bet, they successfully opened the road for motion controler. Now the other manufacturers have reacted. I believe that N realized they won't be able to match kinect V2 for example, that they can't go without core gamers, that they can't pass on a better online service and the associated revenue. I don't believe either that N studio have what it take to push out AAA quality title pushed bu the biggest third editors, they need EA, Ubisoft, etc.
To make it short I believe N realized they are cornered (by MS/Sony and IOs/Android/WM on the mobile market) without a proper reaction they can lost everything (or move forward as a software editor). Hence their anticipated move, they are in need for a more balanced business model which includes core gamers and more importantly support for third party editors. As they say in the only legit (or so it clloks like) slide, they listened to what editors wants as it their only way out looking forward. I don't expect them to push a monster but at least to offer a "sane" system, my belief is that somehow matching Llano perf is a good call. They need this as a basis looking forward, PC enthousiast gamers my think Llano GPU perfs (or future Intel counter part) are irrelevant but editors may think other wise and it's a possible reason behind all the talk and rumours lately about a rebirth of PC gaming. My bet is N want to be in the ball park in regard to perfs to what Amd and Intel will offers while topping nowadays consoles.
In regard to EDRAM I think N will pass as while it's convenient it's not standard (especially looking at PC), ATI RBE make a great use of available bandwidth deferred renderer push that trait even further, even a 128bits bus + GDDR5 will fill their need (MSAA at least in console realm can be substituted by screen space forms of AA as MLAA, it's clear by "real" people return that ti's good enough), more on it I don't expect the bandwidth requirement for the CPU to be that much higher than the tenth of GB/s available to xenon.
I also believe that N if it could would have launch this fall (say they informed editor last falln they could have run a cheap port of COD this fall) they may be waiting for mass availability of 32nm process. They clearly have a shot to ship a 200mm² (or a bit less) single chip, this will be cheap enough (if HD 5770 are any clue), if they are trying to make sense to 2GB of RAM would be a better place to spend money than in extra execution resources. My growing feeling is that N will push something decent in perfs with a possibility to be impressive at what it does either in perfs per Watts and mm² (so more of a successor to the gamecube than to the wii).
For the same reasons (being a bit cornered by various actors) I expect N to have a pretty relaxed / opened in regard to the system, we have clues that N may leave the HDD choice into consumers hands for example, I don't know what will be their attitude in regard to "other OS feature" but I've hope here too. Same for Steam, it would not surprised me if they land a deal with Valve sooner than latter.
As a summup I expect N to do "the right thing" on a lot of accounts as they don't have that many chances to stay relevant otherwise (sorry for N lovers but while super profitable N is fighting giants & my belief is that it could show (without the proper move) sooner than later).
thanks I completely forgot about this
What Nintendo SHOULD have done was this.
Make an announcement that the Wii will be significantly dropped in price ($99) as well as a 10%-20% drop in the cost of software and periphials. Then at the same time make note that they will be entering the console race shortly after (6 mos) MS and Sony with a comparable console that offers Nintendo's unique take on the future of gaming but at the same time employing cutting edge technology.
They could then release a system 10%-30% less powerful then both MS and Sony (because honestly at that point in time that small difference won't matter nearly as much to many people) and use that extra cash to do something unique. They would be the last console of that generation to be released so many might hold off on purchasing either the MS or PS console because "its right around the corner"
Make an announcement that the Wii will be significantly dropped in price ($99) as well as a 10%-20% drop in the cost of software and periphials. Then at the same time make note that they will be entering the console race shortly after (6 mos) MS and Sony with a comparable console that offers Nintendo's unique take on the future of gaming but at the same time employing cutting edge technology.
They could then release a system 10%-30% less powerful then both MS and Sony (because honestly at that point in time that small difference won't matter nearly as much to many people) and use that extra cash to do something unique. They would be the last console of that generation to be released so many might hold off on purchasing either the MS or PS console because "its right around the corner"
D
Deleted member 13524
Guest
liolio:
I generally agree with you about what Nintendo should do.
My comments about what Nintendo is going to do are based solely on the information we've been gathering regarding its final performance.
A 640-shader, 32 TMU and 16 ROP RV740, plus an UMA of 2GB 128-bit GDDR5 would put the Nintendo Stream with 3->4x greater graphics capabilities than the PS3 and X360 and clearly allowing a disruptive difference in its games' visual quality -> even if they just copied the 163 million transistors Xenon.
With all the rumours converging to a console that will simply "match" the current crop, everything points that Nintendo isn't aiming at that level of performance. They just want to show off the new gimmick (controller) and at the same time regain support from decent 3rd parties.
Let's not forget that Nintendo has always had a steady 5-year longevity for their home consoles.
With the PS4\X720 launching on 2014 with "exquisite" hardware, it'll take at least until late 2015/2016 for the developers to really come up with "disruptive-looking" games, and that's only a year away from Nintendo launching the successor to Stream, with a new gimmick and hardware that matches the other two (at a fraction of their initial price).
Of course, none of this prevents them to invest heavily in a new and up-to-date network environment, as that is equally crucial to keep up with PS3 and X360.
I generally agree with you about what Nintendo should do.
My comments about what Nintendo is going to do are based solely on the information we've been gathering regarding its final performance.
A 640-shader, 32 TMU and 16 ROP RV740, plus an UMA of 2GB 128-bit GDDR5 would put the Nintendo Stream with 3->4x greater graphics capabilities than the PS3 and X360 and clearly allowing a disruptive difference in its games' visual quality -> even if they just copied the 163 million transistors Xenon.
With all the rumours converging to a console that will simply "match" the current crop, everything points that Nintendo isn't aiming at that level of performance. They just want to show off the new gimmick (controller) and at the same time regain support from decent 3rd parties.
Let's not forget that Nintendo has always had a steady 5-year longevity for their home consoles.
With the PS4\X720 launching on 2014 with "exquisite" hardware, it'll take at least until late 2015/2016 for the developers to really come up with "disruptive-looking" games, and that's only a year away from Nintendo launching the successor to Stream, with a new gimmick and hardware that matches the other two (at a fraction of their initial price).
Of course, none of this prevents them to invest heavily in a new and up-to-date network environment, as that is equally crucial to keep up with PS3 and X360.
I don't expect N to abandon the Wii, it still sell and they can easily cut the price when they decide that the sales are too low. They have no reason to EOL the Wii, I don't see why they would not considr a price drop at some point either.
They should (I'm confident they have) aimed at more than 10-30% percent differrence. It's damned cheap to beat ps360. If they were to provide such a tiny increase in perfs they should have aim for this fall as a launch date as they can do this at acceptable cost even on 45nm process.
If their chip is as I expect in the same ball park in perf as Llano it will clearly out do the PS360, the extra RAM (even 1GB) will make the difference meaty. Llano has supposedly an acceptable TDP, xenon based CPU and the the same number od SP would be even better in this regard. It's clearly within reach even keeping in mind N conservative nature in regard to hardware.
I know N has been deceiving but the few info we got in regard to the system doesn't make sense if N is to push a PS360 v1.2. More importantly is next to suicidal imho, the new controller while a real good idea ain't the wii-mote or kinect, they need to be reasonnably attractive to gamers and early adopter to drive adoption.
In fairness, console power alone has proven to not be the determining factor of the success of a system. The Wii's issue with third party developers had more to do with marketing and problems with porting down 360/PS3 game engines into a system with such a different artitecture and "shader" system. If the rumors of it being similar to the 360's artitecture is true, it shouldn't be nearly as much of an issue porting down from the PS4/xbox720 as it was for the Wii.
That said, at the very least, Nintendo will most likely address the hardwares issue of the current-gen consoles. Having more RAM available by itself can make a big difference. The system will also have to be powerful enough to stream one or more additional screens without hindering the performance below its targeted rendering range.
Last edited by a moderator:
I tend to speculate keeping in mind the most legit information we have:
Especially the "unparalleled next gen perf", so while supposedly not qualifying as "real next gen" (such a mess of a naming...) I expect it to be really consistantly above the ps360.
I read again reviews of the HD 4770 and it's too big of chip for what N seems to aim at. Something I don't agree is your 3->4x improvement vs the ps360, the hd 4770 is imho more than that, taking FLOPS alone it's already 4x times xenos perfs, but it says a only tiny part of the story as the chip makes much more efficient use of its resources, then texture units have nothing to do with ps360 ones, RBEs does a much better job out of the availlable bandwidth than RSX ROPs and so on. Clearly if devs were to go a bit closer to the metal they could extra a lot more out of it. Anyway it's too much of a chip anyway.
On the other hand something close to Llano/ HD 4670 (Llano has more unit clocked lower) could imho provide a meaty 3->4x jump in perfs over the PS360 not in something as obvious at framerate at a given resolution, but taking in account the quality of shading/effects, AF, etc.
Having the GPU some cyle away from the CPU could also help as I remember reading how communication latency between Xenon and Xenos prevent some devs using the memexport more often (synchronization issues, etC.).
You may be right but I tend to think the N could surprise people a bit in this regard, cause they have to (as we agree on what they should do). N won't survive an hardware gamecube like sales (I know the GC made money but they would lose so much mind share, third party supports, etc.).
What I consider "worse it" out of the rumours we had (as so many sites are after clicks...) is a POWERPC 3 cores CPU, a R7xx based GPU, 8GB of flash storage and BRD disk/player. See what happen this gen, most site don't have a clue, they don't get FLOPS even simple thing as resolution (if they get it they don't care "clicks FTW...). I would not be surprised reports are simply based on resolution aimed by N for rendering either it's 1080p or 720p (actually rendering @1080p is already twice the processing/bandwidth requirements).
So it makes even more sense for N to make its system reasonnably attractive to gamers. I want to upgrade my pc but I no longer want PC (I may get a tablet as I realize that I don't need windows/linux I simply don't do much with my PC), my wife would happily welcome a N console (for mario kart and some other mario game she likes) my-self a Llano grade hardware along with the possiblity to play on the pad (quiet convenient concept if done right imho) may fell my need for some years (espceially if Diablo ii comes on the system
).They will have too, I wonder if after the PSN debacle an Valve /Steam offer would be even more tempting for N as it could allow them to catch-up faster with MS/Sony.
Last edited by a moderator:
D
Deleted member 13524
Guest
I wouldn't say marketing was a problem for 3rd parties. Nintendo sold over 50 million consoles in a little more than two years. Anyone would want to tap that userbase, as regardless of how the console was being marketed, they'd certainly get profits if the game was good.
But yes, the Gamecube's backward compatibility and Nintendo's "hard-head"-ness of going with an evolution (or just overclock, according to some) of Flipper instead of choosing an off-the-shelf\custom Shader Model 2/3 compatible GPU has cost them all the most important 3rd party AAA games and quite a chunk off the console's potential lifecycle (looking at the sales numbers, I'd say the Wii "died" in 2010).
How would you know?! What makes you think the next-gen of consoles will even consist of a multi-core CPU and a GPU?
By 2014, AMD will have passed all the FP calculations to the IGP in their APUs, probably with the IGPs approaching the high-end discrete offerings and sharing huge L3 caches that could even be used for framebuffer; Intel could be bringing some kind of "Larrabee 2" with spectacular rendering\computing performances; nVidia will have already released their autonomous ARM-assisted Maxwell GPU with an estimated 15x greater performance/power than Fermi.
What if those consoles use only ray-tracing engines? Or are solely based off computing performance rather than rendering performance?
Last edited by a moderator:
When I said marketing, I meant it as how Nintendo approached third parties to support the Wii. Nintendo attempt was to attract them with lower development cost, a new interface, and to be able to use previous-gen assets, but the market didn't turn out like that.
I believe that it's a possiblity that MS/Sony will be a bit more conservative for their next consoles, but you're right that my statement would only work with that as an assumption. I suppose we should just hope that Nintendo is developing this console with that in mind.
I still can't imagine why Nintendo would use something from the RV770 era so I'm calling bullshit on any rumors that mention that. If it's not based on Hollywood, it would surely use the latest technology (well latest as in developed in the past few years like Cypress/Cayman). Performance could still suck, like how Xenos was really nifty but slower than a X1900.
Maybe they'll build a console with AMD Llano? That little budget chip will lay waste to PS3/360.
Maybe they'll build a console with AMD Llano?
That little budget chip will lay waste to PS3/360.
D
Deleted member 13524
Guest
There's a rather easy explanation for that.
If Nintendo has no intentions of using the GPU for computing tasks, there's the fact that performance-per-transistor hasn't really increased since the R7xx family. In fact, it has even decreased in the R7xx->Evergreen transition.
Sure, there have been performance/power improvements, but that may also be related to using smaller process nodes or just taking advantage of the same node being more mature. Furthermore, if there are architectural-based power optimizations they could be added to a SM4.1 GPU in a custom design.
Besides, if the console only means to have 3rd party ports from the other consoles with SM3.0, even the fact that it supports SM4.1 could be considered overkill and a waste of transistors.
OpenGL ES 2.0 with SM3.0 compatibility could even be considered better for that target (saving those transistors aimed at supporting more shader instructions in order to add more shader processors, for example).
UE3, CryEngine 2.0, idTech5, MT Framework, Skyrim's Creation Engine, etc.. all the most popular game engines have been adapted and optimized for OpenGL ES 2.0\DirectX9\SM3.0 because of PS360 and higher-end ARM SoCs.
Why spend money+transistors+power+die area for something that won't be used?
Last edited by a moderator:
I read again some reviews of the 4770 and the 5770-5750, it's interesting. It's old tech but in no way worse either in perfs per Watts or mm2. Actually 5750 and 4770 are dangerously close, maybe a comparison between the 4770 and the 5770 is more relevant as they are both "complete" chip but the 5770 is clocked ~13% has 50% more bandwith so it's tough to make direct comparison in 1680x1050 it achieves 28% perfs at 1920x1080 26%// On the other the chip is 26% bigger.
I can't see what is so bad in R7xx vs R8xx as far as bang for bucks are concerned. So far AMD has not push any Cayman derivated in the mid and low end segment.
I don't think AMD has any intensive to sell the thing for cheap.Maybe they'll build a console with AMD Llano?
EDIT the numbers come from this review
The difference is featureset. DirectX 11, GPGPU, UVD and OpenGL enhancements. Some of these brought in the capability for MLAA and sparse grid SSAA. 4770 simply lacks a lot of functionality of 5770. On the PC this hasn't mattered a lot because games aren't flat out requiring the features yet.
Cypress was a nice jump over RV770 and RV670 in features. RV770 was mostly a huge improvement in performance per die area.
I don't think MLAA is impossible on R7xx it's more like AMD didn't feel like porting it. GPGPU is possible. If it's not relevant on PC it's not either on the Wii2 as well as on PS360.
Actually with lower access to hardware I would no be surprise if devs manage to do what they want.
For UVD well it's still a custom chip N may have include some UVD unit even if it's not the last one (3 right?).
Even tesselation used reasonably (ie with reasonable amplifying factor and an adaptive approach) should perform properly.
We can't dismiss costs AMD may ask more for more recent architecture may and there the silicon cost.
Overall I would favor a extra array as even 400sp is not that much to better functionality.
