"The Real Reasons Microsoft, Sony Chose AMD For The XBOX One And PS4"
- Thread starter Dave Baumann
- Start date
Also, Nvidia probably asked for too much money, and Sony might have gotten ripped off by them this generation.
AMD, being in second, would be more pliable to pressure, in order to help secure market support for its standards and tech (aka, the open technologies).
But, I don't follow hardware, especially detailed industry politics, so I may be completely wrong. Someone correct me?
AMD, being in second, would be more pliable to pressure, in order to help secure market support for its standards and tech (aka, the open technologies).
But, I don't follow hardware, especially detailed industry politics, so I may be completely wrong. Someone correct me?
If ARM doesn't offer the performance in the right timeframe then there is no way for NVIDIA to offer an APU solution with enough performance. APU's have advantages over discrete components when sharing data between the GPU and CPU cores.
D
Deleted member 13524
Guest
If it came down to x86, then the choice for AMD is obvious IMO.
First, since the company has no performance GPUs, they'd have to integrate a GPU from another vendor which would require a longer and more expensive R&D process to get the chip done.
Second, Intel already sells all their chips they make in their own factories with brutal margins, they wouldn't be interested in a high-volume, low-margin deal.
First, since the company has no performance GPUs, they'd have to integrate a GPU from another vendor which would require a longer and more expensive R&D process to get the chip done.
Second, Intel already sells all their chips they make in their own factories with brutal margins, they wouldn't be interested in a high-volume, low-margin deal.
I guess I'm surprised that PowerPC is considered harder to program than ARM or MIPS. Also surprised in the claim that it's not as good as the others in terms of performance/watt. This could very well be true, but I wouldn't mind if some reported did more digging into what happened with IBM and why they weren't capable of delivering a CPU for either console.
I wouldn't read too much into that statement about PowerPC. Forbes is not exactly intended to be a technical platform as much as an odious right wing rag. That said, gotta give em kudos for ignoring the APU-newspeak and talking about SOCs.
The 64-bit readiness argument makes sense though, as do current economies of scale. Although AMD itself does seem to agree there's merit in ARM design for their longer term microserver roadmap, it will still rely on Kyoto at least for the nearer future.
That part sounds quite inaccurate to me:
I'm not sure MIPS is any better than POWER imo.
IMo the real reason for AMD is that:
1) they are one of the few company that provides both performant CPU and GPU
2) X86 software environment is the best.
3) Intel may have done an efforts with regard to pricing when they wwere still thinking about launching Larrabee (and that is my speculation), nowadays... the situation is completely different: they have moved to a pretty aggressive roadmap for Xeon Phi, they have now GPU that can compete toe to toe with Nvidia and AMD products in the segments Intel is addressing right now.
I would be surprised if "POWER is tougher to code for" or if IBM embedded products are not competitive. The part about the software environment sounds right though.
I'm not sure MIPS is any better than POWER imo.
IMo the real reason for AMD is that:
1) they are one of the few company that provides both performant CPU and GPU
2) X86 software environment is the best.
3) Intel may have done an efforts with regard to pricing when they wwere still thinking about launching Larrabee (and that is my speculation), nowadays... the situation is completely different: they have moved to a pretty aggressive roadmap for Xeon Phi, they have now GPU that can compete toe to toe with Nvidia and AMD products in the segments Intel is addressing right now.
Intel is the foundry for Altera, Tabula and Achronix, producing FPGAs. I doubt that volume will come close to what is needed for console production. However, Intel doesn't seem motivated to expand to its outsourcing business and knowing Intel probably dealing with them comes with very unfavorable terms. FPGAs for the most part don't come cheap so it probably alot easier to absorb Intel's price demands.
arijoytunir
Regular
arm will be the base cpu for next next gen!
It's not. It has a few more special purpose registers which matter at the ABI level if you're working on compilers and operating systems, but if you're just doing application (read: game) development—even at low level—it's fairly comparable. All 3 are load-store architectures and have reasonable comparable ISAs. ARM is a bit cleaner, but has it's own little quirks, and if you start delving into extensions (e.g. Thumb/Thumb2) the complexity ratchets up a bit.
ARM and MIPS are probably a lot cheaper to license though and the synthesized cores are incredibly modular, so if you're building some custom ASIC with lots of stuff you want to integrate with them, they are rather ideal.
Why?
BC is going to be supported in nextnext gen consoles so I think x86 is the way to go, imho. The PS4 could be perfectly emulated with a more powerful PS5 and the Xbox One is meant to be indefinitely -or forever- BC compatible.
When you factor in those details all bets are off.
AMD was the only one who has the IP to do a CPU/GPU SOC as well, presumably with savings. Well Intel might nearly be in a position now with 800+ gflops in high end Haswell. But Intel is tough to deal with and they're still short on GPU compared to PS4.
Interesting stuff given the source. My thinking was beeifier cores (EG piledriver, or Haswell) would have been better, but it looks like they were looking downscale not up.
Indeed I think Intel could be a player in the future if they want to be. There was rumors of Sony sniffing around Larrabree.But yea in the future I would expect unified solutions as well which Nvidia and ARm maybe less likely now that both commited to x86 amd AMD of course and Intel if they see the use.
I doubt Intel will ever have any interest in a market which maybe will do 200-250+ million in the next 5 years, where the total SoC cost is less than $100 and profit margins would be far from what they are used to. That's less than what they produce in a year. They would offer up an Atom chip just in spite. 
itsmydamnation
Veteran
I doubt Intel will ever have any interest in a market which maybe will do 200-250+ million in the next 5 years, where the total SoC cost is less than $100 and profit margins would be far from what they are used to. That's less than what they produce in a year. They would offer up an Atom chip just in spite.
where the hell do you get your Numbers from 250million over 5 years @ $400 a console is 60k per console per year for world wide sales. If your talking SOC revenue then even taking a very pessimistic number of 50million consoles world wide LTD @ 5years your taking 5 dollars a chip.........
He means 250 million consoles sold...
OT, I'm surprised that ARM came very close. I used to get in vehement arguments with a guy on another board who was fixated on ARM being (the main CPU) in next gen consoles (before we knew anything about them). I was of the position there was absolutely no way, ARM were too weak.
Turns out he was a lot closer to right than I thought.
OT, I'm surprised that ARM came very close. I used to get in vehement arguments with a guy on another board who was fixated on ARM being (the main CPU) in next gen consoles (before we knew anything about them). I was of the position there was absolutely no way, ARM were too weak.
Turns out he was a lot closer to right than I thought.
I really wonder about 200-250 million consoles over the next 5 years. But MS is talking about 1 billion units even so who knows.
Mobile devices will get there, if they haven't already. Don't know about consoles.
Besides the technical reason, I'm thinking AMD would have done anything to make this deal. If they didn't get the deal, they might not hang on.
Mobile devices will get there, if they haven't already. Don't know about consoles.
Besides the technical reason, I'm thinking AMD would have done anything to make this deal. If they didn't get the deal, they might not hang on.
Well this gen over 6-8 years so far Wii:100m, PS360: ~80m each (IIRC)
So 260 million this gen.
Of course it looks like Wii U will come nowhere near Wii. Personally I think that means PS4/1 will pick up at least much of the slack and sell more, but some seem think they'll meander along to 70m each like this gen.
MS can talk all they want about 1 billion units, I dont think they have a prayer at 499. The key to that is going to be low pricing. And starting at a high point means it's that much harder/longer to get to the low price points.
