Actually Samsung sampled it already over 2 years ago, they announced in January 2011 completion and availability for testing of 2GB DDR4 modules and Hynix did the same in April 2011.
Xbox One (Durango) Technical hardware investigation
- Thread starter Love_In_Rio
- Start date
- Status
Actually Samsung sampled it already over 2 years ago, they announced in January 2011 completion and availability for testing of 2GB DDR4 modules and Hynix did the same in April 2011.
2years is over 1 year
I knew it was a while ago, but I didn't feel like searching for the date, but thanks for the clarification.
Is this the specifications update VGLeaks mentioned days ago?
I don't think so. Vgleaks specs make sense unlike this. For a moment if this is paritally correct. 1 apu for 360 bc. Instead of gddr4 it is 12 gb ddr3. It could still be very expensive. The most expensive console ever probably. Vgleaks missed 360 soc for BC. The rest I think they are right
January 2013 DDR4 demo from crucial:
http://www.expertreviews.co.uk/components/1297015/crucial-ddr4-memory-demoed
Crucial is just late with their DDR4 clearly, having demoed it 2 years later than competition.
We'll be seeing first PC CPUs with DDR4 in 2014 based on current schedule, it's not impossible MS went ahead and told Samsung and/or Hynix to jumpstart those DDR4 factories for the XBox already.
We'll be seeing first PC CPUs with DDR4 in 2014 based on current schedule, it's not impossible MS went ahead and told Samsung and/or Hynix to jumpstart those DDR4 factories for the XBox already.
Bah. I say it IS impossible. The specs of both orbis and durango have been reserved and conservative. To "jumpstart" factories to produce a completely new memory standard not set for mass market for two years or more (2014 is server premiere; consumer products will follow even later) does not fit with this pattern.
First, Crucial is not a memory supplier, it's a client.
Second, the final specs was published September 2012.
That's called risk. That's where the risk is.
Second, the final specs was published September 2012.
http://www.infoworld.com/t/computer-memory/adoption-of-ddr4-memory-facing-delays-216439
That's called risk. That's where the risk is.
Bullshit memes don't really fit into a rational discussion.
When you make a bullshit statement, I'm not sure what you expect.
Improbable, highly unlikely, financially suicidal are all possible (I have no idea of the actual costs that would be involved). It seems pretty clear from the time lines involved that it IS possible however unlikely.
Sure it's possible. Just like it's possible for the WiiU to outsell both 720 and PS4 combined at the end of console life cycles.
Not worth debating the semantics, or whether it's 0% (impossible) or 0.1% (incredibly unlikely). I think it's more useful to discuss what they'd actually gain from doing it and how much work it would entail. I don't know much about DDR4, but unless they gain bandwidth or density (eg 12GB), then it really wouldn't be worth doing. Someone mentioned that they could do 12GB with DDR3, and it would probably be cheaper and simpler to do it that way. Maybe there are issues with heat, power, PCB size for 12GB of DDR3, but I would doubt it. Relative to other heat producers and power consumers, it should be a very minor change. PCB layout (which is also an issue with DDR4) would be a significant change. If they literally can't fit the extra chips on the PCB and had to jump up a PCB size, that would be a big change as well in terms of the cost of the BOM. So I'd say unless they want to increase the RAM and can't fit more DDR3 on the board, then DDR4 density is not a win. That leaves bandwidth as the only real reason to jump to DDR4. That would entail a new memory controller, a completely new PCB layout, and it would invalidate all of the bandwidth specs they sent out to developers. I think from that alone, you can make a guess as to how likely this rumour is, without beginning to question the validity of the source.
Alphawolf, let me spit on the ground symbolically as I reject your self proclaimed position as the alpha male.
You asked "what risk?". I tried to bring references, including real industry opinions, about when we should expect parity with DDR3, and the risks involved in choosing DDR4. I also pointed out that the specs were not final until Sept 2012. How about you follow from that instead of trying to win arguments with pointless semantics?
Moving on, I'm pretty sure saving two or three watts won't make any difference on the PS nor the cooling on a 100W+ console. The parts involved aren't priced with such granularity.
You asked "what risk?". I tried to bring references, including real industry opinions, about when we should expect parity with DDR3, and the risks involved in choosing DDR4. I also pointed out that the specs were not final until Sept 2012. How about you follow from that instead of trying to win arguments with pointless semantics?
Moving on, I'm pretty sure saving two or three watts won't make any difference on the PS nor the cooling on a 100W+ console. The parts involved aren't priced with such granularity.
Final specs are meaningless unless you care about being industry standard. Non-standard tech is released all the time because waiting for a committee of competitors to agree upon standards is time consuming. The only real risk for ddr4 would seem to be cost, but that could be a long term positive as well.
Every watt makes a difference in a console design, but the cost needs to align.
DDR4 will be in use in other platforms within a few months of durango's release so it's undoubtedly something they've looked at. It's really a matter of whether it makes fiscal sense and with the PC world being what it is, it probably doesn't.
Every watt makes a difference in a console design, but the cost needs to align.
DDR4 will be in use in other platforms within a few months of durango's release so it's undoubtedly something they've looked at. It's really a matter of whether it makes fiscal sense and with the PC world being what it is, it probably doesn't.
A new PCB layout is a non argument as designing one is no relevant problem and who says the memory controller in the APU hasn't DDR4 support from the beginning? And lower performance specs given to 3rd party devs to be surpassed sounds like a non issue to me.
Who knows if that paper is true or not but the way some people argue here why this would be such a huge design issue escapes me. Unless people here worked on the APU design and know for a fact that it does not support DDR4 and know about the deals MS can or can not have with Micron/Samsung/Whatever this is just pointless.
In Feb. most people here still were so convinced that the PS4 can't have more than 4GB DDR5 because either it would be impossible to fit on their MCM or too expensive. Should have humbled a few
french toast
Veteran
What would the point be in having 12gb of memory anyway?? Unless ddr4 gives you gdr5 levels of bandwidth at ddr3 power consumption I cant see any benefit.
How would move that kimd of data around without there being some bottlenecks that make it a waste of memory chips?
If microsoft put in 12 gb of any kind of ram it will be PR stunt and a BOM waste, especially being linked to the rumoured mid range gpu.
How would move that kimd of data around without there being some bottlenecks that make it a waste of memory chips?
If microsoft put in 12 gb of any kind of ram it will be PR stunt and a BOM waste, especially being linked to the rumoured mid range gpu.
french toast
Veteran
Yes I forgot about that, durango is rumoured to be reserving anything up to 3gb for OS+ other functions.
But still I do find that completely unbalanced fitting in with a low power cpu and a mid range gpu, even if you could move that kind of data around/supply OS.
- Status
Similar threads
