If it makes you less confused, use Valve instead of Apple. Steam took something like $3.5 billion last year and Valve took 30% of that on Microsoft's OS. So "Look at the kind of returns Valve makes from Steam and you'll see why MS should be interested!"
Because Valve can do this, doesn't mean Microsoft can. The gaming market is expanding but finite. For Microsoft to make substantive gains the market needs to grow and/or they need to attract to sales away from Steam.
But looking at established ecosystems, they generally fall when they lose profitability or because a competitor hugely leapfrogs them. Steam shows no signs of the former and they've been leading the charge for decent consumer digital rights for years. They know their audience and serve them
Well. I've been using Steam since the bad early days because Half-Life 2 forced you too.
Microsoft's only hope to displace Steam is to offer something Steam doesn't (and can't introduce) and which is also wildly popular. Ecosystems are popular economic models because it's been shown that content customers rarely leave.
This leaves things like UWP but I don't know if that's a big enough offering. Plus if Microsoft gate Windows technology locked behind their store they're going to get spanked by the EU again.
Gaming is as hard a nut to crack as mobile operating systems and we've seen how well Microsoft have done there even when throwing billions of dollars at the problem. Making a profit is not a matter of money, enthusiasm or will. You can't engineer a commercially successful product. Zune was a better MP3 player than the iPod but that counted for nowt to the majority of consumers.
For starters because it's simply fucking stupid to have it external, especially when the xbone case is as large as it is to begin with. Second, that I don't have any way in my place to hide it, so it's always going to be a huge unsightly PoS lump of plastic on the floor.
When even as something as tiny as an Apple TV has an internal power supply it's ridiculous that MS can't manage.
That a <<2W unit has no problem hiding the power inside the case should be obvious. Adding the PSU inside the case would either increase the size or at least outputs more heat which will increase the noise floor which I regard as a bad compromise for a unit which is not visible for most people in a normal media setup in a living room.
If the PSU annoys you so much I can't fathom how you feel about all the crappy cables behind the TV/AR setup and the silly things like external drives and whatever media/network junk you need these days.
Gaming is as hard a nut to crack as mobile operating systems and we've seen how well Microsoft have done there even when throwing billions of dollars at the problem. Making a profit is not a matter of money, enthusiasm or will. You can't engineer a commercially successful product.
I'm not sure of your argument. I wasn't saying it was easy. I'm saying it can be highly profitable if you succeed, which is why MS should keep trying. Do you disagree?
Demonstrably not true, with PS4. It has internal PSU and substantially smaller case compared to bone. It's all in how you manage airflow. With the bone being largely empty space inside, there's no reason you couldn't have PSU internal and cool it with the same fan used in the external brick if needed, without requiring any additional noise or space, or massive redesign of the internal layout.
If the idea running about that NEO shifted to Polaris because a simple die shrink was off the table because of process changes at 16/14 nm, then that should inform about the design of the XBslim.
Shouldn't a XBslim using Polaris 11 (16cu) at system clock of 1050 Mz (driven by limit of CPU) and using higher speed DDR4 be a complete no brainer? That should be enough for comfortable 1080P/60fps at say $279 and $199 digital only (no BRD or HDD).
That should be plenty to well against NEO and OG PS4. Vega based new console could come next holiday.
If the idea running about that NEO shifted to Polaris because a simple die shrink was off the table because of process changes at 16/14 nm, then that should inform about the design of the XBslim.
Shouldn't a XBslim using Polaris 11 (16cu) at system clock of 1050 Mz (driven by limit of CPU) and using higher speed DDR4 be a complete no brainer? That should be enough for comfortable 1080P/60fps at say $279 and $199 digital only (no BRD or HDD).
That should be plenty to well against NEO and OG PS4. Vega based new console could come next holiday.
I don't know the sound level of current PS4s, but launch era units are louder, yes. Not because of the integrated PSU tho, it sits downstream of the SoC heatsink and doesn't need much cooling.
I don't know the sound level of current PS4s, but launch era units are louder, yes. Not because of the integrated PSU tho, it sits downstream of the SoC heatsink and doesn't need much cooling.
The amount of the cooling required by the PSU can be debated it is likely proportional to SoC load. As SoC load goes up as would PSU requirements. Both PSU and SoC should be sending temperature requirements to the OS to increase venting or not.
How much we don't know but I think it's safe to assume it's adequate cooling but we are unsure as to what the division between the two are.
MSFT made a lot of effort but had proper power management be part of their SOC that would have been a lot easier, for them and SOny as well.
Playing vids should only burn a handful of Watts on a proper TV hub with dedicated hardware, both system were failed in that regard. Whether the PSU is integrated or not is of lesser importance than the primary issue => sucky power efficiency
If the idea running about that NEO shifted to Polaris because a simple die shrink was off the table because of process changes at 16/14 nm, then that should inform about the design of the XBslim.
Shouldn't a XBslim using Polaris 11 (16cu) at system clock of 1050 Mz (driven by limit of CPU) and using higher speed DDR4 be a complete no brainer? That should be enough for comfortable 1080P/60fps at say $279 and $199 digital only (no BRD or HDD).
That should be plenty to well against NEO and OG PS4. Vega based new console could come next holiday.
Sony was planning iterative consoles before the generation started, so I think Neo is just a design based on a shrink and GPU core doubling of the original architecture. While I like that design idea for X1S, I think it will be DDR4 and slightly boosted clocks at most. It just wouldn't make sense that they would bother using an all new Polaris design for 1 year before replacing it with Vega. I think the plan is to go cheap now and wait until Vega and HBM until 2018. Neo might go on sale this holiday, but it only effectively has a one year market life until rumors about newer much more powerful Xbox architecture hit. I think Xbox's bet is to go for the budget play for the next 2 years. Just look at Surface and Lumia (neither of which are expected to introduce new hardware this year) and put Xbox in context.
I'm not sure of your argument. I wasn't saying it was easy. I'm saying it can be highly profitable if you succeed, which is why MS should keep trying. Do you disagree?
As an absolute, yes I disagree. This dovetails back to the point about nobody knowing the rate of return of Microsoft's ongoing investments in gaming. There becomes a point when the resources of fruitless endeavours are better diverted elsewhere. Managers must be able to recognise when the pursuit isn't worth the effort. Like a lion chasing that gazelle which is just too far away, sure maybe it could bring it down but it'll be exhausted. It's reserves sapped. Somebody else will steal the kill.
Knowing when it call it quits isn't failure. Particularly when the money concerned belongs to your stockholders. Of course Microsoft have no other business with projected high growth. So they chase. Like Clarence the one-eyed lion.
As an absolute, yes I disagree. This dovetails back to the point about nobody knowing the rate of return of Microsoft's ongoing investments in gaming. There becomes a point when the resources of fruitless endeavours are better diverted elsewhere. Managers must be able to recognise when the pursuit isn't worth the effort. Like a lion chasing that gazelle which is just too far away, sure maybe it could bring it down but it'll be exhausted. It's reserves sapped. Somebody else will steal the kill.
Knowing when it call it quits isn't failure. Particularly when the money concerned belongs to your stockholders. Of course Microsoft have no other business with projected high growth. So they chase. Like Clarence the one-eyed lion.
Sony was planning iterative consoles before the generation started, so I think Neo is just a design based on a shrink and GPU core doubling of the original architecture. While I like that design idea for X1S, I think it will be DDR4 and slightly boosted clocks at most. It just wouldn't make sense that they would bother using an all new Polaris design for 1 year before replacing it with Vega. I think the plan is to go cheap now and wait until Vega and HBM until 2018. Neo might go on sale this holiday, but it only effectively has a one year market life until rumors about newer much more powerful Xbox architecture hit. I think Xbox's bet is to go for the budget play for the next 2 years. Just look at Surface and Lumia (neither of which are expected to introduce new hardware this year) and put Xbox in context.
What I was reference was the rumor / suggestion that for Sony a redesign was as cheap / cheaper than a shrink. Basically, since the transition to FinFett requires a redesign and not a shrink, then why not go for the Polaris GPU / DDR4 / keep esram and have a box that you can sell as a $199 STB ($279-299 with HDD and BRD) and still comfortably beat the OG PS4 and have solid 1080P60 performance. Then launch the Vega in 2017 with ~ 2X NEO perf at $399 + while keeping the slim for the low end for another 3 years. This year should definitely be all about the budget slim with Vega and HBM2 delays. From a cost standpoint, there should be no meaningful cost difference in the system that I described and the OG XB1 at 14 nm. Therefore, why not incorporate the off the shelf Polaris 11 for improved performance?
Jumping to a different topic, to the GTX 1080 really put a lot of thing into perspective. It's ~320sq.mm monster though it you take a quarter of it you have an improved GTX 750ti that is for 80sq.mm of silicon quite impressive. Even more so when you have the aforementioned GPU fared against the XB1 and the PS4. It is not unreasonable to expect the GP108, the next generation of entry level Nvidia card to match this generation of console (if Nvidia choose so).
It makes my distaste for UMA approach in console stronger and stronger, you could not build an efficient SOC based on such GPU because on a 64 bits bus you would be limited to 2GB of RAM. SOC are great for mobile, for cost and because those designs have low bandwidth requirements. Once you need bandwidth I wonder if the cost part hold true. Both Sony and MSFT jump through of hoops to make it happen, Sony swallow a huge cost and could have get stuck with 4GB of RAM (not that terrible by-self or for game only) and MSFT approach is complicated and failed as it did no saved them the use of a big bus and pretty expensive and fast DDR memory.
Now when you compare that to the Basic Alienware Alpha (which I own for the ref), the overall memory set-up but also the GPU side of things, one may wonder if it was really worth it to jump through all that hoops.
It is worse if you consider that there were available GPU that could have been used in console, Bonaire comes to my mind strongly
As an side in the discrete vs SOC approach, previous consoles were using newer shrunk part when available, not all the IPs are meant to be available at any node down the road (more about that in the end).
So starting from here (entry level discrete gpu offering console level of performance) and that SOC might be a little oversold, if MSFT is to release something new, cheap or not, they should go with a PC under disguise: NUMA, discrete CPU and GPU. It is tricky and costly to alleviate the bandwidth requirement face by a SOC aiming at decent 3D performances. It would be easier to deploy games as UWP under disguised, further easing the convergence between MSFT lines of products.
In retrospective, the XB1 could has looked like that (mostly the same specs and perfs):
CPU: 2 Jaguar cluster as of now (100ish sq.mm of silicon) // GPU:bonaire // Memory: 4GB of DDR3 for the RAM, 2GB of fast GDDR5 for the VRAM.
So wrt to cost only the CPU would have been custom, that is significantly less silicon that either the XB1 or PS4, and the memory set-up cost less than both set-up.
Down the road, MSFT could have updated the CPU complex the Puma+ with appropriate power management. As things went they would have been stuck to Puma+ for the CPU and the refresh for the GPU may have come later than expect (20 nm products being cancelled) yet their will be GPU available @16nm soon ( and the level of performance of console should be achievable with entry level hardware).
Ultimately they would end with two tiny chips, and 2 64 bit bus (to DDR3-4 and GDDR5x). Overall it would have been much more flexible and adaptable to technology advancement hiccups than the system we have. It would have been more amenable to more progressive upgrade, or a tiered approach to the market (aka one system shipping with a better gpu).
So If MSFT is to launch something new, instead of something akin to the PS4 Neo so a bigger XB1 I would favor something like that:
CPU: custom Puma+ (28nm) // GPU: of the shelves Polaris 11 GPU soldered to the mobo // Memory: 6-8GB of RAM - 2-4GB of VRAM.
Yes. This is what I meant. If its true that Sony is going with NEO because AMD offered this as a side effect of the cheaper manufacturing process, the same should hold for XBOXOne. So it should be expected that MS will do the same
