As for the other, well I don't have a good feeling about it.
I assume they will be in talks shortly about either acquisition, or exclusive partner rights.
They will be quite a commodity for whoever picks them up ...
What?
Do you think it would be a mistake for MS and/or Sony to launch in 2012?
- Thread starter kagemaru
- Start date
What?
I was talking about nothing other than how the correlation between transistor counts and Flops is pretty meaningless. Going forwards, we may find less transistors offer better 'flops' due to being able to work more efficiently (similar to a smaller unified shader part outperforming a larger, more Floptacular fixed-shader part).
Yes, but more transistors is not an inherent disadvantage.
The example you cited is merely better utilization of the transistors, not that more transistors held a gpu/technology back.
Last edited by a moderator:
Of course it isn't. Processors are made of transistors. More transistors = more processor = doing more stuff. My point was and only is that a linear multiple of transistors doesn't need to equate to the same linear multiple of flops for processors to be progressing and able to do more work. Diminishing returns, where they do exist, aren't to be found in counting flops.Or putting it another way, the law of diminshing returns in processing isn't the same as that of mechanical work, where resistances see more energy being wasted than put to work. If a car with 100 HP can go 100 KmH, we don't expect a car with 400 HP to go 400 KmH. But with processors these workrate limits aren't there, and if a processor with 100 ALUs and 100 Hz can process 100 GFlop, the same processor architecture at 200 Hz or with 200 ALUs will process 200 GFLops. There is a linear relation. The reason for Flops not progressing as such in GPUs is because they aren't identical flops doing identical work.
Or putting it another way, the law of diminshing returns in processing isn't the same as that of mechanical work, where resistances see more energy being wasted than put to work. If a car with 100 HP can go 100 KmH, we don't expect a car with 400 HP to go 400 KmH. But with processors these workrate limits aren't there, and if a processor with 100 ALUs and 100 Hz can process 100 GFlop, the same processor architecture at 200 Hz or with 200 ALUs will process 200 GFLops. There is a linear relation. The reason for Flops not progressing as such in GPUs is because they aren't identical flops doing identical work.
Agreed.
Sorry for the confusion, I thought you were getting at something else.
'Scuse me Mr Geezer. If they do have Windows 8 compatibility then surely that would entail adding an ARM based CPU into the platform? How would they implement that? Would they go for cross compatibility with a set top box type arrangement and thus put ARM in all SKUs or go for the high end approach and use it as a way of extracting additional revenue on a SKU which comes pre-loaded with Windows 8 embedded?
One thing is for sure:
If xb720 is going to be used as something other than a gaming machine (Windows8 pc/tablet, media-player/streamer, internet hub, dvr/cable-box) It will either have to sell for a higher price than these devices do on the market at the time of launch, or it will have to be sold at or above cost.
MS doesn't want to be in the business of selling a console for a loss with no sure way to make that cost up.
If people are buying the console for all the things it can do, but care nothing for the gaming aspect, that spells trouble for the bottom line. Most of the functions listed above do not have a concern for competing products exist at much lower pricepoints than what a new console is expected to debut at.
Mediaplayer/Streamers - Bluray devices which do this are regularly available for ~$100
Internet Hub - googletv and the like are available for ~$100
DVR - TiVo Premiere XL HD DVR is a dvr a cable box, a music box, a web box, and a movie box which sells for $250
The higher price of xb720 should keep buyers interested in just these services away from xb720 as the lower prices of these devices is much more compelling.
But when we add "Windows8 PC" to the mix, the waters start to get muddy.
Looking at Bestbuy prices, PCs start at around $300.
Core introduction of xb360 was $299 - which sold at below cost.
So one of two things need to happen:
1) Higher initial MSRP to separate from PC's (and help shave the losses)
2) Introduce Windows8 as an add-on pack for $100
The other obvious answer would be to strip the hardware so that the BOM is below $300, but this would handicap the hardware and detract from the purpose of driving software sales which then lead into users buying into all of the other extras these devices(and consumers) are intended to dive into.
One interesting pricing method that hardware makers haven't delved into is the iphone method.
If these consoles can be tied into a cable company provider (Brighthouse, Timewarner, Cox, DirectTV, etc.) with a longterm agreement, (much like Apple does with Verizon, Sprint and AT&T) then MS/Sony can use the extended capabilities of the machines to provide the role of cablebox/dvr as well as all of the above AND gaming.
Example:
IPhone = $200 with 2yr contract (AT&T) or $600 by itself
xb720/PS4 = $200 with 2yr contract (cable) or $500 by itself
The other added benefit of this method is obviously a much larger 17" wide case which enables easier cooling.
The steps MS has been making in this regard are starting to hint in this direction. Perhaps MS will take the dive and fully embrace the advantages such an arrangement would have for their bottom line.
Windows 8 isn't ARM exclusive.
Dominik D
Regular
It's extremely unlikely for TripleX to be on Win8*, either ARM or x86/amd64 (or, lulz, IA64). Having it run on any other HW means 1) soft still has to be rebuilt and retested for the new architecture; 2) Windows code has been running on that architecture for at least 3 years. That last part means that there would have been rumors by now and lots of them. Win 8 compatibility seems super improbable to me. Managed APIs compatibility - sure, but nothing more.
*or Win8 compatible beyond managed code, which is already pretty much true due to XNA
By Win8 compatibility, I mean the Win Apps over Windows Store and not 100% Windows OS on 360.
Replace the cable/net with Live gold and you've got a point
E.g $500 for console + (random number) $50/year Live vs $300 for console + $100/year live with 3 year contract
This is why I think it is a stupid idea for Microsoft or Sony to lauch a console in 2012 or maybe even 2013.
For Nintendo its a different story and it does not really matter, Nintendo is a smaller company but Microsoft and Sony need to lauch at the earliest at 2013 and at the latest on 2014 after we are all safe when the world ends and 22nm or 28nm (for Microsoft) is able to ramp up into full chip production for CPUs and GPUs.
The main reason is thermals and power consumption, the same major problems that plagued the current generation which I also believe should have been delayed by two or three years or able to have had a 65nm CPUs and 55nm GPUs would have been of great help but we are still enjoying this generation despite its flaws because we just do not know what flaws could have popped up on a delayed launch.
Just go and look at the launch PlayStation 3 60GB dissassembly videos or photos, notice how Sony put FIVE high quality copper heat pipes into the heatsink-fan cooling system, not only that but the fact that Sony also put some rather robust heat spreaders on top of both CellBE and RSX/GDDR3 setups, even the EE/GS chip had a heat spreader (I disassembled my PS3 60GB to apply Artic Silver 5 back in January of 2011) even though the PS2 chip is not producing the same level of heat that CellBE/RSX do given they are new archs on 90nm process.
Basically it is imperative that both Microsoft and Sony simply wait for mature 28nm process production and 22nm process production to not only be able to put out a much more efficient high end CPU and GPU set up, I also wager that we will still see some type of high quality copper heatpipe heat sink set up on at least one or both of those consoles (note how I just cannot defend the cooling design on the Xbox 360)
After such a goal is reached then either MS or Sony can worry about RAM setups, storage mediums, etc as well as benefit from higher end SATA, USB 3.0 and whatever else is expected to come down in price.
You haven't been paying attention...
28nm is starting mass production now and will be at retail q1/2012 in high end GPU's.
20nm is expected to start mass production q3/2012 and at retail q4/2012.
By q4/2012, it will be no problem for measly ~100watt GPUs and ~40watt CPUs @ 28nm in nextgen consoles.
As for cooling solutions, MS learned their lesson with their flawed launch cooling design.
If that means huge copper heat pipes, or more intelligent design (as is the case with current slims) then that is what will be done.
Replace the cable/net with Live gold and you've got a point
E.g $500 for console + (random number) $50/year Live vs $300 for console + $100/year live with 3 year contract
If Live is to start offering cable programming, then sure they could afford to charge enough to offset the cost.
Somehow though, I doubt the cable companies that are providing bandwidth for the internet-connected xbox720 will be all too happy to see such an arrangement though.
It is a difficult proposition to enter the cable TV market in the USA (not sure how that process is in other regions).
Here it is a near monopoly/duopoly per region.
Getting regulatory approval for CableTV directly on live will not be easy.
Anything less than Cable TV though wouldn't be enough to justify a ~$50/mo supplimental 2 year contract to offset the initial high cost of the console.
Partnering with the existing cable companies would be a much smarter way to go.
Whereas I can somehow see someone launching along bigN in fall 2012 I can't see it to use @28nm/32nm lithographies. It's imho too early I expect production capacity to be to constrained and I don't remember manufacturers going with fresh/next too immature process. Too much is at stake I can't see anyone taking risk after the RroD.
So for me if it launches in 2012 hardware will use one of the 45nm 40nm processes available.
For me the gen has already been really long and the aforementioned processes offer enough of a kick to avoid a system to look like a Wii to the GC.
The more I think about it the mote I would welcome something conservative (reasonnably though) and cheap along with returning to five years life cycle, future proofing the system for more time is unrealistic/too expansive in many regards.
Also on topic: Sony launching in 2012 would conflict too much with Vita. They need a good effort on getting Vita on the map, and they don't particularly strike me as the kind of company that could pull off both of them in the West in the same year.
On the other and, that makes it slightly more interesting for Microsoft to launch in 2012 to give Wii U a tougher time and again a good heads up on Sony. The only downside there is the Kinect platform would probably suffer if they include Kinect 2 in the box, so they'd have to tread that one carefully so as to not get current Kinect owners too annoyed. In that respect, an 'inbetween' snack would perhaps make sense, and have Kinect 2 launch as a Windows oriented device first. That could also give developers a good heads up for the next gen launch on Microsofts console.
On the other and, that makes it slightly more interesting for Microsoft to launch in 2012 to give Wii U a tougher time and again a good heads up on Sony. The only downside there is the Kinect platform would probably suffer if they include Kinect 2 in the box, so they'd have to tread that one carefully so as to not get current Kinect owners too annoyed. In that respect, an 'inbetween' snack would perhaps make sense, and have Kinect 2 launch as a Windows oriented device first. That could also give developers a good heads up for the next gen launch on Microsofts console.
MS launched xb360 on 90nm 11/2005.
ATI had their first 90nm GPU on retail shelves one month prior.
Contrast this with 2012:
q1/2012 28nm Nvidia & AMD GPU's will launch.
This leaves QUITE a bit of time to iron out the process node for a year end release.
In fact, it would be roughly the same as the timeframe between xbox1 gpu nv2a built on 150nm process and the first PC GPUs built using the same 150nm process coming out in q1 of the same launch year.
MS may not come out in 2012, but the process node will not be the roadblock which prevents it from happening.
Also on topic: Sony launching in 2012 would conflict too much with Vita. They need a good effort on getting Vita on the map, and they don't particularly strike me as the kind of company that could pull off both of them in the West in the same year.
On the other and, that makes it slightly more interesting for Microsoft to launch in 2012 to give Wii U a tougher time and again a good heads up on Sony. The only downside there is the Kinect platform would probably suffer if they include Kinect 2 in the box, so they'd have to tread that one carefully so as to not get current Kinect owners too annoyed. In that respect, an 'inbetween' snack would perhaps make sense, and have Kinect 2 launch as a Windows oriented device first. That could also give developers a good heads up for the next gen launch on Microsofts console.
Indeed.
It would be difficult for Sony to match a 2012 launch given their attention to Vita, but I'd think they have enough satellite developers which can work on relatively trivial ps3>vita ports to free up their AAA devs to work on ps4.
Exactly.
Of the two, which one is likely to bring Sony more money in the long run:
PS4 or Vita ...
Yep, if I were in their shoes, I'd be moving my best dev studios off of ps3/Vita games and onto ps4 ASAP ...