Decline PS4 wins the console war against XboxONE, yet it is a hollow victory as Consolesdämmerung is upon us

[DraQ]

AI is really hard to spread to multiple cores actually. AI is very "branchy" code, which is the opposite of what's easy to split up.

You don't really need to split it up, just splitting it, or portion of it off the main logic would be beneficial.
Hi-level decision making doesn't need to be as fast as "hard" logic of the game, it may actually benefit from being a little slower by adding some hysteresis and avoiding exploitable decision switching - for example you have two routes from enemy to to the player, by moving slightly player can force naively implemented AI to alternate and fail to reach player, but if AI is slower it may already be too far down one route for it to be effective when player's movement retriggers the decision making.

And stuff like basic pathfinding should be pretty parallelizable too.

 

[Cowboy Moment]

I doubt there's any point in trying to parallelize pathfinding or AI behaviour in general. In your typical game, you'd normally have more than 4 active AI actors on the map at the same time, with an 8 core CPU you can just create a thread for each and let the OS scheduler arrange things for you (although apparently console optimization involves manually tethering tasks to specific cores...). I'm not a game engine developer, but I would expect that a modern engine already spends a lot of time synchronizing work between threads, so I doubt splitting it up further would help all that much. Also, as Mantle demonstrates, a large amount of cpu time is needed for rendering due to D3D overhead, and this cannot really be parallelized at the moment.

This likely depends on the game as well.

 

[CSM]

Odds are it's selling better because some 360 owners are transitioning to the PS4 instead of the XBO. That's not exactly a victory.

How is more sales for Sony not a victory for Sony?

Plus, I shouldn't have to remind you (but apparently I do) that the PS3 had an absolutely awful launch with few games and retailing for a price nobody wanted to pay. It's not surprising it didn't pick up momentum until there were a decent number of games available and the price dropped. The PS4 doesn't have those problems.

Sales weren't any better for the PS1 or PS2.

 

[CSM]

The problem with the Wii's "expanded audience" is that those people didn't buy all that many games in the long run. Nintendo made a lot of money selling the consoles, but profits tapered off sharply as more and more Wiis ended up collecting dust.

They might have bought less games. But it was still a lot of games.

Lifetime software sales Wii, January 2014: 892.34 Million
http://www.ign.com/articles/2014/01/29/these-are-nintendos-lifetime-hardware-and-software-numbers

Lifetime software sales PS3, October 2013: 905,9 Million
https://www.scei.co.jp/corporate/data/bizdata_software_e.html

 

[Xor]

Odds are it's selling better because some 360 owners are transitioning to the PS4 instead of the XBO. That's not exactly a victory.

How is more sales for Sony not a victory for Sony?

I thought we were talking about the collapse of the video game industry here. If you just want to fanboy for Sony then yeah, they're probably going to do better than Microsoft this generation. For whatever that's worth.

Plus, I shouldn't have to remind you (but apparently I do) that the PS3 had an absolutely awful launch with few games and retailing for a price nobody wanted to pay. It's not surprising it didn't pick up momentum until there were a decent number of games available and the price dropped. The PS4 doesn't have those problems.

Sales weren't any better for the PS1 or PS2.

The PS1 and PS2 didn't have worldwide launches, but were instead only available in Japan for months, so obviously their sales would be lower initially. If you look at the PS2's numbers they shoot up pretty quickly as the PS2 launched in other countries. The PS1 also had the added difficulty of being Sony's first console - no loyal customers, no brand recognition, no established franchises, etc. And that console still sold almost as many units as the PS3.

 

[tuluse]

The PS2 sold literally as quickly as they could make them once it launched in North America. The thing was that they *only* made 100k per week. They can make many more PS4s per week than that.

 

[Xor]

The PS2 sold literally as quickly as they could make them once it launched in North America. The thing was that they *only* made 100k per week. They can make many more PS4s per week than that.

Yeah, I didn't even mention the supply issues. The same thing happened to the Wii.

 

[CSM]

I thought we were talking about the collapse of the video game industry here.

Which doesn't require you to make shit up.

 

[Xor]

 

[Destroid]

I doubt there's any point in trying to parallelize pathfinding or AI behaviour in general. In your typical game, you'd normally have more than 4 active AI actors on the map at the same time, with an 8 core CPU you can just create a thread for each and let the OS scheduler arrange things for you (although apparently console optimization involves manually tethering tasks to specific cores...). I'm not a game engine developer, but I would expect that a modern engine already spends a lot of time synchronizing work between threads, so I doubt splitting it up further would help all that much. Also, as Mantle demonstrates, a large amount of cpu time is needed for rendering due to D3D overhead, and this cannot really be parallelized at the moment.

This likely depends on the game as well.

Pathfinding can be pretty demanding in RTS games that deal with a lot of units in real time. Dwarf Fortress pathfinding is also pretty demanding, although that's probably pretty suboptimal code/algorithms from toady.

 

[Cowboy Moment]

Oh right, almost forgot.

Over 150 million over its lifetime, and it's unlikely any console will ever top that. PS3 and XboX360 are both at ~80 million, and are unlikely to go much higher given their current monthly sales.

PS3 sales were about half at this point in his life cycle.

http://vgsales.wikia.com/wiki/PlayStation_3

Just shut up.

Bro I think you are having trouble expressing yourself. I will assume you actually meant PS2 because otherwise your post makes no sense.

Now, putting aside the extremely silly premise that the world of consumer technology has not changed at all in the last 7 years, and therefore the PS3 will follow the PS2 sales curve, let's take a look at the actual numbers.

Let's use the beginning of 2014 as our cutoff point. The PS3 launched in November 2006, that gives us 7 years plus one quarter worth of sales. In this time, the console sold a little over 80 million units (link).

Now, the PS2 launched in October 2000 in the US (and then November in EU), so adding our 7 years plus a quarter time period, gives us the beginning of 2008. At this point, the lifetime sales of the PS2 were over 118 million, and probably close to 125 million, assuming relatively even distribution of sales across 2007 (link).

I think it bears pointing out that 125 is indeed significantly greater than 80. Furthermore, monthly sales of last-gen consoles are down 70-80% in January 2014 compared to January 2013. The PS3 sold roughly 50k in the US this January.(link)

Now, if you want to pretend that both the PS3 and PS4 are going to reach PS2 numbers despite all these extremely obvious indicators that the console market has indeed contracted, well, be my guest. I am not so cruel as to kick a poor console peasant while he's down.

Just putting things into perspective.

You're not, you're trying to win a pissing contest.

Egads, my scheme has been uncovered!

 

[Cowboy Moment]

I doubt there's any point in trying to parallelize pathfinding or AI behaviour in general. In your typical game, you'd normally have more than 4 active AI actors on the map at the same time, with an 8 core CPU you can just create a thread for each and let the OS scheduler arrange things for you (although apparently console optimization involves manually tethering tasks to specific cores...). I'm not a game engine developer, but I would expect that a modern engine already spends a lot of time synchronizing work between threads, so I doubt splitting it up further would help all that much. Also, as Mantle demonstrates, a large amount of cpu time is needed for rendering due to D3D overhead, and this cannot really be parallelized at the moment.

This likely depends on the game as well.

Pathfinding can be pretty demanding in RTS games that deal with a lot of units in real time. Dwarf Fortress pathfinding is also pretty demanding, although that's probably pretty suboptimal code/algorithms from toady.

It can, but that's irrelevant to my point. If you have 50 units on the map that you need to do pathfinding for (though modern RTS games seem to do this per group rather than per unit), splitting any of those 50 tasks does literally nothing to help you. As a matter of fact, it hurts, because you now spend more time on synchronizing your results.

 

[Destroid]

I doubt there's any point in trying to parallelize pathfinding or AI behaviour in general. In your typical game, you'd normally have more than 4 active AI actors on the map at the same time, with an 8 core CPU you can just create a thread for each and let the OS scheduler arrange things for you (although apparently console optimization involves manually tethering tasks to specific cores...). I'm not a game engine developer, but I would expect that a modern engine already spends a lot of time synchronizing work between threads, so I doubt splitting it up further would help all that much. Also, as Mantle demonstrates, a large amount of cpu time is needed for rendering due to D3D overhead, and this cannot really be parallelized at the moment.

This likely depends on the game as well.

Pathfinding can be pretty demanding in RTS games that deal with a lot of units in real time. Dwarf Fortress pathfinding is also pretty demanding, although that's probably pretty suboptimal code/algorithms from toady.

It can, but that's irrelevant to my point. If you have 50 units on the map that you need to do pathfinding for (though modern RTS games seem to do this per group rather than per unit), splitting any of those 50 tasks does literally nothing to help you. As a matter of fact, it hurts, because you now spend more time on synchronizing your results.

There are a lot of different ways to do multi-threading, from what I understand most game engines just assign tasks to cores (render on one, physics on another etc), sticking pathfinding/ai on its own core (rather than trying to process it all in parallel) probably yields reasonable benefits. I believe the supreme commander engine does something like this, I don't know about starcraft 2. It doesn't scale, but it saves you a lot of headaches with synchronisation like you mentioned.

 

[CSM]

Oh right, almost forgot.



Bro I think you are having trouble expressing yourself. I will assume you actually meant PS2 because otherwise your post makes no sense.

Now, putting aside the extremely silly premise that the world of consumer technology has not changed at all in the last 7 years, and therefore the PS3 will follow the PS2 sales curve, let's take a look at the actual numbers.

Let's use the beginning of 2014 as our cutoff point. The PS3 launched in November 2006, that gives us 7 years plus one quarter worth of sales. In this time, the console sold a little over 80 million units (link).

Now, the PS2 launched in October 2000 in the US (and then November in EU), so adding our 7 years plus a quarter time period, gives us the beginning of 2008. At this point, the lifetime sales of the PS2 were over 118 million, and probably close to 125 million, assuming relatively even distribution of sales across 2007 (link).

I think it bears pointing out that 125 is indeed significantly greater than 80. Furthermore, monthly sales of last-gen consoles are down 70-80% in January 2014 compared to January 20143. The PS3 sold roughly 50k in the US this January.(link)

Now, if you want to pretend that both the PS3 and PS4 are going to reach PS2 numbers despite all these extremely obvious indicators that the console market has indeed contracted, well, be my guest. I am not so cruel as to kick a poor console peasant while he's down.

I thought we were comparing PS4 to PS3.

I can't into read.

 

[Xor]

Apparently.

 

[Infinitron]

Microsoft deserve to be hated on, but I do think that in a way the Xbox360 was a blessing in disguise, combined with the botched launch of the PS3.

Imagine if the PS3 was as dominant as the PS2 was.

Forget multiplatform games and bad console ports, man. If Sony had ruled the roost for another seven years, PC gaming would have been truly dead because every game would have been a fucking PS3-exclusive.

 

[BlackAdderBG]

Yeah thank god we are getting the Bioshocks and the Obvlions

 

[tuluse]

Microsoft deserve to be hated on, but I do think that in a way the Xbox360 was a blessing in disguise, combined with the botched launch of the PS3.

Imagine if the PS3 was as dominant as the PS2 was.

Forget multiplatform games and bad console ports, man. If Sony had ruled the roost for another seven years, PC gaming would have been truly dead because every game would have been a fucking PS3-exclusive.

There are multiple flaws in this argument.

1) The existence of the 360 doesn't change how profitable it is to sell a game on PC.
2) Microsoft pays a lot of money to make exclusive Xbox games. How much of that money would have been invested in PC gaming instead?
3) PC gaming dying

 

[Infinitron]

There are multiple flaws in this argument.

1) The existence of the 360 doesn't change how profitable it is to sell a game on PC.
2) Microsoft pays a lot of money to make exclusive Xbox games. How much of that money would have been invested in PC gaming instead?
3) PC gaming dying

1) Having one single ultra-dominant platform would have greatly reduced the need to invest in multiplatform development processes. It's like, "Okay, we need to create a 360 and a PS3 SKU, so we might as well do a PC one as well."
2) Well, in my scenario the 360 would have been launched but failed, so they'd have been in trouble. Or even if not, it's Microsoft man, I'm sure they'd fuck it up.
3) Yeah, well.

 

[Cowboy Moment]

Oh right, almost forgot.



Bro I think you are having trouble expressing yourself. I will assume you actually meant PS2 because otherwise your post makes no sense.

Now, putting aside the extremely silly premise that the world of consumer technology has not changed at all in the last 7 years, and therefore the PS3 will follow the PS2 sales curve, let's take a look at the actual numbers.

Let's use the beginning of 2014 as our cutoff point. The PS3 launched in November 2006, that gives us 7 years plus one quarter worth of sales. In this time, the console sold a little over 80 million units (link).

Now, the PS2 launched in October 2000 in the US (and then November in EU), so adding our 7 years plus a quarter time period, gives us the beginning of 2008. At this point, the lifetime sales of the PS2 were over 118 million, and probably close to 125 million, assuming relatively even distribution of sales across 2007 (link).

I think it bears pointing out that 125 is indeed significantly greater than 80. Furthermore, monthly sales of last-gen consoles are down 70-80% in January 2014 compared to January 20143. The PS3 sold roughly 50k in the US this January.(link)

Now, if you want to pretend that both the PS3 and PS4 are going to reach PS2 numbers despite all these extremely obvious indicators that the console market has indeed contracted, well, be my guest. I am not so cruel as to kick a poor console peasant while he's down.

I thought we were comparing PS4 to PS3.

I can't into read.

Is that why you specifically quoted the part of my post comparing PS3/Xbox360 to the PS2? And provided a link to a page with no references to the PS4 whatsoever?

In any event, PS4 will not do PS2 numbers either, that ship has sailed. It might do better than PS3, but it's impossible to tell until we get out of the supply constrained early adopter phase.

 

[DragoFireheart]

Best thing it can happen to us is the market crashing to a point where no exlusives remain and old titles get ported and re-released on PC.

Why would you want consoletard garbage being ported and re-released on PC?!

Dark Souls.

 

[Cowboy Moment]

Best thing it can happen to us is the market crashing to a point where no exlusives remain and old titles get ported and re-released on PC.

Why would you want consoletard garbage being ported and re-released on PC?!

Not all of it is garbage. The AAA graphics arms race has done some serious damage to console gaming, but it still has some pretty good niche games. For example, I'm not much into dungeon crawlers, but handhelds supposedly have some good Wizardry clones. Good action games, fighting games, driving sims - all more prominent on consoles than on PC.

 

[Servo]

Just curious, did anyone here buy the Xbone or PS4 who actually doesn't want to kill themselves now?

 

[Admiral jimbob]

Jesus Christ how did this thread get 166 pages and 185k views in a month and a half? Did something interesting happen? Surely something interesting must have happened. Someone summarise.

 

[Xor]

I have no idea, I just showed up after someone changed the title a page or so ago.