Digital Foundry Article Technical Discussion Archive [2013]

Status
Not open for further replies.
Nothing new, for sure, but I think it's a pretty big deal because if Sony provides a real-time OS with libGCM v2 to use with GCN, also providing low level access to the hardware, I can see how one or two years out, 1st party developers & Sony's internal teams will have had long enough experience to build up interesting tech to actually leverage the console.

Microsoft on the other hand won't provide low level access to the platform, and it might be a disadvantage in the long run. Perhaps it is not essential tbh, but it is a nice possibility to have.

Why won't MS provide low level access?
 
Why won't MS provide low level access?
While I'm not saying they won't, I could see why they wouldn't. Getting developers to use DirectX for Xbox development also reenforces their whole DirectX platform which includes windows games. By providing low level access, they provide developers ways of avoiding DirectX use all together.
 
I suppose so, but why else do you limit your console's performance to a fat API while trying to build an ecosystem based on that same API in other sectors? ;) Note we had a developer a month or three back saying MS is not allowing low-level access
 
I suppose so, but why else do you limit your console's performance to a fat API while trying to build an ecosystem based on that same API in other sectors? ;) Note we had a developer a month or three back saying MS is not allowing low-level access

Well the 360 uses dx9+ and despite this, they didn't restrict dev from having low level access. And where is this dev quote? would like to read it.
 
Well the 360 uses dx9+ and despite this, they didn't restrict dev from having low level access. And where is this dev quote? would like to read it.

IIRC, it was something early in the life of the X360. As well it was likely either a misunderstanding of what was said to the interviewer or a misunderstanding on the part of the person being interviewed.

We've had multiple developers here that have mentioned going quite low level with the X360. Sebbbi for one I believe has done quite a bit with regards to what could be considered low level hardware access.

I believe most developers choose not to do so just because the development tools on X360 were so good and there aren't as many performance cliffs that you can stumble across compared to PS3.

Regards,
SB
 
IIRC, it was something early in the life of the X360. As well it was likely either a misunderstanding of what was said to the interviewer or a misunderstanding on the part of the person being interviewed.

We've had multiple developers here that have mentioned going quite low level with the X360. Sebbbi for one I believe has done quite a bit with regards to what could be considered low level hardware access.

I believe most developers choose not to do so just because the development tools on X360 were so good and there aren't as many performance cliffs that you can stumble across compared to PS3.

Regards,
SB

Oh I do understand all that. Infact there is a thread that dealt with the issue of low level access on 360. From several of the comments above, it seems as if its been said somewhere that durango would not allow such low level access. I was just challenging Shifty's assertion about the implications of using dx11 on durango. My point being that even if they are using dx11 on durango, it will be similar to the one on pc as an api layer but how thick, heavy and flexible it will be will probably be very different as I suspect that the dx11.1x on the durango will expose the underlying hardware while possibly providing access to devs to go lower level if they feel like it.

What I am saying in essence is that durango, for example, will probably dx11 in the same way that it will use the same windows 8 kernel as pc but it is not the same as the full fat one on pc.
 
IIRC, it was something early in the life of the X360. As well it was likely either a misunderstanding of what was said to the interviewer or a misunderstanding on the part of the person being interviewed.

We've had multiple developers here that have mentioned going quite low level with the X360. Sebbbi for one I believe has done quite a bit with regards to what could be considered low level hardware access.

I believe most developers choose not to do so just because the development tools on X360 were so good and there aren't as many performance cliffs that you can stumble across compared to PS3.

Regards,
SB

What exactly is "low level access" anyway? What could possibly be simpler than executing draw triangle commands via the DirectX API besides just filling registers with data?
 
What exactly is "low level access" anyway? What could possibly be simpler than executing draw triangle commands via the DirectX API besides just filling registers with data?

It's a but more complicated than that these days.
You wouldn't believe how complicated the hardware that sets up shader constants is....
Having said that the idea of low level access is somewhat overrated, IMO there is far more performance in optimizing your data for the platform than there is in some mystical low level access.
The somewhat larger issue is how multiple applications and the OS rendering requirements are handled.
 
It's a but more complicated than that these days.
You wouldn't believe how complicated the hardware that sets up shader constants is....
Having said that the idea of low level access is somewhat overrated, IMO there is far more performance in optimizing your data for the platform than there is in some mystical low level access.
The somewhat larger issue is how multiple applications and the OS rendering requirements are handled.

Isn't there some programming law anyway that says something like "given more and more power, the energy needed to extract those last bits of performance goes up exponentially" or am I just misremembering Ahmdal's law?
 
It really bugs me how the article keeps talking about the low level API as if this is something new. This has been a feature of consoles since... forever. PS4 is bringing nothing new to the table in that regard so why keep harping on about it as if it's something brand new?
Maybe because a certain other console won't be going that route.

And it sounds as though Sony will be providing different levels of access depending on the comfort and/or willingness of the devs to go deeper or not.
 
Isn't there some programming law anyway that says something like "given more and more power, the energy needed to extract those last bits of performance goes up exponentially" or am I just misremembering Ahmdal's law?

Doesn't make intuitive sense. Given more and more power, you become less concerned about extracting "last bits" anyway. Given an ocean of power, 95% of it is good enough, you dont need to chase that last 5%.

It's interesting to think about. A lot of the work on PS360 is extracting that last 5%, because they arent powerful enough. Weirdly, I have thought next gen could in some ways decrease dev effort, not increase it.

How much of 600 people on Assassins Creed are dedicated to squeezing one more ounce out of the aging, constrained, 360 for their super ambitious game? Arent frame rates often in the 20's for that game?

I saw a recent Sweeney interview where he talked about some promising algorithm needing 20 teraflops to efficiently implement, vs the "1-2 teraflops" we currently have (matches with next gen consoles). Need for next-next gen already confirmed,

Anyway, already OT...
 
How much of 600 people on Assassins Creed are dedicated to squeezing one more ounce out of the aging, constrained, 360 for their super ambitious game? Arent frame rates often in the 20's for that game?

Not much. I'd be surprised if the engine team numbered greater than say 20 people.

Everyone else is mostly there for art, animation, modeling, level design, etc.

Regards,
SB
 
Well the 360 uses dx9+ and despite this, they didn't restrict dev from having low level access. And where is this dev quote? would like to read it.
Couldn't find it, so it's just the realm of rumour at the moment. You're right that potentially MS could release a console that isn't compatible with the rest of their infrastructure, but I'm with Joker in thinking MS are setting up for a platform other than a console. As my speculation.
 
Isn't there some programming law anyway that says something like "given more and more power, the energy needed to extract those last bits of performance goes up exponentially" or am I just misremembering Ahmdal's law?

http://en.wikipedia.org/wiki/Amdahl's_law

Relation to law of diminishing returns

Amdahl's law is often conflated with the law of diminishing returns, whereas only a special case of applying Amdahl's law demonstrates 'law of diminishing returns'. If one picks optimally (in terms of the achieved speed-up) what to improve, then one will see monotonically decreasing improvements as one improves. If, however, one picks non-optimally, after improving a sub-optimal component and moving on to improve a more optimal component, one can see an increase in return. Note that it is often rational to improve a system in an order that is "non-optimal" in this sense, given that some improvements are more difficult or consuming of development time than others.

Amdahl's law does represent the law of diminishing returns if you are considering what sort of return you get by adding more processors to a machine, if you are running a fixed-size computation that will use all available processors to their capacity. Each new processor you add to the system will add less usable power than the previous one. Each time you double the number of processors the speedup ratio will diminish, as the total throughput heads toward the limit of .

This analysis neglects other potential bottlenecks such as memory bandwidth and I/O bandwidth, if they do not scale with the number of processors; however, taking into account such bottlenecks would tend to further demonstrate the diminishing returns of only adding processors.
 
Status
Not open for further replies.
Back
Top