Where was that stated as a design parameter for Durango's APU?
Are you using the top-end clocks for different (smaller, binned) products that are well into diminishing returns territory for price, power, and peformance?
Xbox One (Durango) Technical hardware investigation
- Thread starter Love_In_Rio
- Start date
- Status
Are you using the top-end clocks for different (smaller, binned) products that are well into diminishing returns territory for price, power, and peformance?
We're on the same page with this. 3tf GPU would not have erased the questionable policies, but it at least gives pause to the notion of writing it off completely. In the list of pros and cons, it is certainly a pro.
As is, the tech specs given are a main driver in people saying "why bother?". Especially with questionable policies outside of the specs.
@3Dillante
this isn't the place for this argument but, yeah.
you're doing everything by degree, take it to it's conclusion, why not downclock the gpu to 500 mhz? bet the yields would be stupendous! And how cool running would that be! Why 768 shaders when 400 is way cheaper! Who needs 8 cores btw?
Where is the magical line? There isn't one. Just increments and they all matter.
i could be wrong there's an argument durango is "good enough", but imo nothing would sway the mood right now like we hear xbone gpu is now running 1ghz.
And we're all making nothing but assumptions about the cost. I am just as much as you. Maybe MS has a sheet that says if we upclock the GPU 20% the cost is xxx and that number is really high. But maybe it isn't. neither of us knows.
We do know comparable PC GPU's run 1ghz with ease though.
Also
this isn't the place for this argument but, yeah.
you're doing everything by degree, take it to it's conclusion, why not downclock the gpu to 500 mhz? bet the yields would be stupendous! And how cool running would that be! Why 768 shaders when 400 is way cheaper! Who needs 8 cores btw?
Where is the magical line? There isn't one. Just increments and they all matter.
i could be wrong there's an argument durango is "good enough", but imo nothing would sway the mood right now like we hear xbone gpu is now running 1ghz.
And we're all making nothing but assumptions about the cost. I am just as much as you. Maybe MS has a sheet that says if we upclock the GPU 20% the cost is xxx and that number is really high. But maybe it isn't. neither of us knows.
We do know comparable PC GPU's run 1ghz with ease though.
Also
D
Deleted member 13524
Guest
They added eSRAM to make up for a bandwidth deficit provided by the use of a rather slow DDR3 (which is BTW shared with the CPU so it's never really 68GB/s).
James Car didn't compare the bandwidth of the xbone to a HD7770, he compared the computational units between them, which is correct.
In the end, the HD7770 should be the closest domestic GPU to the xbone's.
No one is saying Microsoft wanted to put a HD7770 in the console but they somehow had to cut on the clock so they added shaders and put some ESRAM to compensate.
they probably just had a CU target AMD was able to meet modularly. The GPU was already going to have to be redesigned for the ESRAM and memory buses anyway.
Personally I call it like a 7770 sometimes just as an approximation, it's not meant to have any great meaning. When I heard PS4 was 1.8 TF and Xbone 1.2, I said ah, so one basically has a 7850 and one basically has a 7770. I know that's not exactly right, but it helps to illustrate things.
Biggest diff (besides memory subsystem stuff) is that XGPU can issue two triangles a clock instead of 1.
Personally I call it like a 7770 sometimes just as an approximation, it's not meant to have any great meaning. When I heard PS4 was 1.8 TF and Xbone 1.2, I said ah, so one basically has a 7850 and one basically has a 7770. I know that's not exactly right, but it helps to illustrate things.
Biggest diff (besides memory subsystem stuff) is that XGPU can issue two triangles a clock instead of 1.
True, but I find it hard to believe that Ms/amd would have such great difficulty in attaining those clocks on a very mature 28nm process and in a relatively large box.
Consider, the 2ghz clock is to be used in tablet/netbook enclosures...
Just saying.
major nelson just tweeted a new video interview with marc whitten about the xbox architecture...
and all they talk about is cloud, kinect, voice, the controller, skype, everything but hard specs.
ugh.
http://news.xbox.com/2013/05/marc-whitten-and-major-nelson-discuss-xbox-one-architecture
i guess the interesting part is they talk cloud right at the beginning. interesting i guess.
yeah, they are definitely hiding.
and all they talk about is cloud, kinect, voice, the controller, skype, everything but hard specs.
ugh.
http://news.xbox.com/2013/05/marc-whitten-and-major-nelson-discuss-xbox-one-architecture
i guess the interesting part is they talk cloud right at the beginning. interesting i guess.
yeah, they are definitely hiding.
The apparent sweet spot of the various designs is approximately where the rumored clocks are.
This isn't a linear situation, so there is a difference in the benefit for lowering requirements below the region where the curve is pretty flat is different from pushing this past the knee of the curve. The slope is much steeper at the top end.
Take whatever number of first-year console manufacturing volumes you desire, say 8-10 million. Now pick a percentage you want to apply to that amount, and throw them away. Assume a chip that size costs something like $50 (I'd expect more, but let's just say).
ok, i pick .00001%.
And again, sweet spot or no, your same logic says lets clock 700. No 600. The safer the better.
It's all increments and it all matters.
So exactly what would 1ghz to the gpu add to the performance of durango
The cost could be the same or less than doubling your ram amount of such an expensive ram.
It really depends on what the yields are. If your only loosing 10% of of your yields at 10 million . Your loosing 1m chips @ $50 is only $50m thats a drop in the bucket for ms in the long run. Will shipping 9m durangos the first year vs 10m be that much of a diffrence and thats assuming the yields stay that way the whole year. they could increase as time goes on within the year.
The cost could be the same or less than doubling your ram amount of such an expensive ram.
It really depends on what the yields are. If your only loosing 10% of of your yields at 10 million . Your loosing 1m chips @ $50 is only $50m thats a drop in the bucket for ms in the long run. Will shipping 9m durangos the first year vs 10m be that much of a diffrence and thats assuming the yields stay that way the whole year. they could increase as time goes on within the year.
yup, and you arent gaining nothing, either.
if being more competitive allows you to maintain a higher price, you save money.
just incremental. hard to exactly quantify, but obviously you could do more shaders, etc.
it's 25% improvement of raw parameters, hard if not impossible to quantify onscreen, but definitely better.
What would the performance be at those speed be. The cost of a higher clocked apu would scale much faster than the cost of other components in the console. the added cost could disapear over the course of months if it was small to start with and surely by the first micron drop it will not even be a thought in peoples mind.
Will be interesting to see what happens. We might just see the 1.6ghz /800 durango as leaked or we could see something else.
I'm starting to think you're hoping for unicorns.
This is Microsoft's accounting for what it has budgeted versus its own design parameters, not what Sony does with its BOM.
It's $50 million to please a smattering of forum posters on the internet.
There is also opportunity cost of any sales that don't happen for X months in a supply-constrained scenario, but I won't get into that.
I doubt pushing the design that close to the top range can happen with only a 1% increase in discards. If it did, they'd be selling 2.2-2.4 GHz Jaguars.
i was making a point. we dont know what that percentage is. you said pick a percentage so i did...at the percentage i picked it's probably worth it.
If we're talking a hypothetical difference between 800MHz GPU and 1GHz GPU then I do think it makes a difference. The first will be (and routinely is) spun as 50% weaker than PS4's GPU, the second would be spun as 20% weaker than PS4's GPU and I think there absolutely is a difference in the kind of impact that has on potential buyers. You may think that the numerous sites reporting things like this has no real effect on how the product's perceived and made no real contribution to the big stock spike for Sony but I don't agree with that.
It isn't really about what the real perceived difference in game quality will be, it's about getting the feeling that MS is at least trying to bring forth about the same level of hardware capability. 20% will "feel" a lot closer to that than 50%, and I do believe this matters psychologically. And the irrational psychological reaction to a product does impact sales.
- Status
