Digital Foundry Article Technical Discussion Archive [2015]

The real question is "do we need these huge teams"?

 

That's the infamous "balance" article. They say wait until the games are out to judge if they got the balance right. Well, lots of games are out now and it sure seems as if they failed to get the balance right because they are generally being outperformed by the exactly what you would expect based on the difference in GPU capability. Whatever data they were looking at then that showed them that the CPU was going to be the bottleneck to performance in games this gen seems, so far, to have not been correct.

 

You honestly have no way of knowing what games released on either console have been bottlenecked by the CPU.
Both consoles have issues maintaining their target frame rate in several AAA titles. With both consoles having almost the exact same CPU a title that is CPU bound on one will be CPU bound on the other. This has nothing to do with which console you own or like.
You also have to keep in mind that a majority of games released are either a devs first attempt at releasing a title on these machines or a remaster of titles that were designed with last gens completely different hardware in mind.
The GPU in the Xbox One is fairly balanced for the amount and type of memory this console has. As well as Rops.

 

I think nowadays the huge teams are needed more for asset creation rather than on the coding side.

 

Although content is much easier to create, we have a lot more required, yep, probably can't argue with that.
But content can be outsorced (and is outsorced)...

 

In hindsight, having much stricter limitations was such a nice thing.

 

Yep, the amount of content/assets created these days can be extremely labor intensive. However, it does seem to be the trend that the coding skill required for these game developers has become less and less important. Its more about using the tools than digging into the source code. I say that because its kind of sad that engineers at Microsoft have been able to take somebody else's work, and improve the performance relatively quickly. Game development has taken on the assembly line mentality, lots of bodies doing small pieces of work that are inserted into the final product. Advanced game engines and excellent software development tools make it all work.

 

The CPU in these consoles bottlenecking performance in games is possible. The GPU in the XBOne bottlenecking the performance in games is obvious.

It's possible some developers were erroneously targeting a different balance of CPU to GPU power in initial titles than what was ultimately delivered or were unable to utilize GPU compute to offload work from the CPU to the degree that might be recommended by the hardware capabilities. I would expect though, as we move forward, that this will be addressed.

In short, if it isn't already, the GPU capabilities that are most commonly limiting graphical performance in games I expect that it ultimately will be.

 

This site goes through all the different roles in making an AAA game. http://askagamedev.tumblr.com/post/88963351061/aagds-game-career-faq

Not too many render guys from the sounds of it. Lots of programming to make tools that are used by other people to make content. I don't think there is anything wrong with this concept, you want everyone fully dedicated to their task and this is one way to do it.

 

If razterized polygon aren't good enough for next step in realtime rendering. I think rendering programmer have a few years of fun to come.

 

i don't understand... the fact that Microsoft can improve their own black box APIs means that all other developers suck at engineering? This sounds like a new spin on "lazy devs".

 

Well. Not really. if anything the skill requirement is higher, requiring more systems that are good enough to be rapidly produced.

It's more the case that the scope of projects is much larger now, and the core tech teams generally aren't much bigger. So ultimately you just don't have the time to make things as efficient you'd really like to. Suddenly you are maintaining 25 complex systems that are all working well, but maybe not as well as they could be - instead of 5 systems that absolutely must be optimized to hell and back due to much stricter hardware limits.

Rendering wise, consider the features early last gen that individually were cutting edge and defined the visuals of the game - but are now simply a checkbox requirement.

 

One of my favourite descriptions of the shifting demands on game software developers over the last fifteen+ years went something like: "it used to be programming; now it's software engineering".

 

Not suck, but I am of the opinion that if your body of work can be improved by outsiders in short order, then perhaps your body of work was pretty sub par. Devs aren't lazy, they do the best they can with the resources and time they have, but that doesn't mean their work is always top notch. Think of it this way, if a third party was contracted to port a game from PS4 to X1, and the X1 version was better than the source material, would that not seem ass backwards to you?

 

That's not necessarily the case, especially when the outsiders are experts in the API's being utilised by said code.

Extremely efficient API's can be extremely inefficiently implemented (possibly due to documentation issues), and reviewing code for efficiency will likely be focussed in a way that was not practical for the original development team.

 

I like fun

 

No. Why should optimizing the source material be harder than creating the source material itself?

 

never mind

 

Naughty Dog has somewhere around 300 people last I checked. Must be a lot of MBA's and unskilled workers over there.

 

300 software developers?

I said "some of that culture".