Xbox One (Durango) Technical hardware investigation
- Thread starter Love_In_Rio
- Start date
- Status
What's the something being discussed? Some kind of pipeline tweak or half of the eSRAM's interface?
The latter case seems like a large thing to gloss over.
The latter case seems like a large thing to gloss over.
Did that get them 50% more bandwidth on the same bus at the same clocks?
It was a military FCS and the form factor was extremely tight.
Apparently it was easier to route the board that way, at the minor expense of not being able to do unaligned reads or writes or read or write bytes from odd addresses. The latter made the displat more than a bit annoying to work with.
It wouldn't have been an issue if they'd bothered telling anyone that was what they intended to do, but it ended up wasting a lot of time and money.
I would tend to agree, but it's very much dependent on how most of the communication between the hardware and dev support software teams is done.
I could see it being done with block diagrams and people making assumptions, but I agree it seems unlikely that something of that magnitude would be overlooked.
Love_In_Rio
Veteran
Any hint about cpu clocks in X1 or PS4?. i still think at least Sony tried to go with 1,8-2 GHz to make the change from Cell for first parties more bareable (see the Guerrilla pdf about Shadow Fall where they are trying to mimic cell jobs in 6 jaguar cores, surely they will miss some of the cell performance ).
Shifty,Brit!. Readmit MikeR!.He was the link!.
Last edited by a moderator:
That sounds really peculiar to me -as if it were unusual-, and unexpected. I don't know how bandwidth is measured and what they use to measure it.
I used to think that Sony, Microsoft and Nintendo began to work with the vanilla hardware AMD provided them with. And then they started tweaking here and there and expending loads of money in R&D -what Microsoft seemed to do-.
I believed all that money they invested on R&D was meant to show how far they could make the vanilla hardware go without overclocking it.
I believed that they were always comparing the original vanilla hardware to the hardware on development and measured the results side by side almost everyday.
In the wired articles on the announcement date there was mention of the Cadence system for simulation: http://www.engadget.com/2013/05/21/building-xbox-one-an-inside-look/
Does anyone here know more about that Cadence system (the name, specs?) The wired article is a bit limited... "Space-ship like device..."
I ask since I am curious to know just how much custom design work and simulation MS might have done. I use Cadence design tools at work, but they are software running on (big) Unix boxes. No "space-ship like devices". Is the wired article referring to some big server rack or a rack loaded with huge Stratix or Virtex FPGAs? (Wild random guess.)
It is a personal curiosity, especially with the recent statement of touching every part of the system for efficiency. Just wondering if anyone has run across anything solid in terms of the MS design and simulation capabilities and where/how they were applied.
(((interference)))
Veteran
Interestingly enough, while I was initially dismissive of this (and all upclock/downclock rumours), it might turn out to be true as I've just heard similar murmurings from elsewhere.
No confirmation yet, but we should know in a few days.
(((interference)))
Veteran
Not if it's only a small 75 mhz upclock.
Betanumerical
Veteran
I haven't really seen a console receive a up clock this large, nor this close to the release date.
Its possible, but this isn't really a minor thing, it would effect the thermals, the yields, it could (depending) effect the power situation.
(((interference)))
Veteran
Well' I've been told before that upclocks are possible with XB1 hardware and it would mainly be an issue of MS being willing to eat the cost of poorer yields rather than running into TDP issues.
Maybe, like the ESRAM thing, this is something they discovered they could get away with from testing final silicon - without any significant impact on their BOM.
I think a change under 10% can be described as a small increase...
Maybe, like the ESRAM thing, this is something they discovered they could get away with from testing final silicon - without any significant impact on their BOM.
I think a change under 10% can be described as a small increase...
Love_In_Rio
Veteran
I believe in an upclock. They are in a desperate situation, at least from a marketing point of view. Don Mattrick departure confirms the mess.
Well, in fact, who knows if Sony didn´t also upgrade the clocks?. These companies do spy each other...and this time it is even easier to know about the another because they both are using TSMC to make the chips. 28nm and GCN is mature enough to put some CUs without many yield loss at 850Mhz instead of 800.
dragonelite
Veteran
I thought their goal was to become silent.
Maybe now their goal is overclock and a bit of noise
(((interference)))
Veteran
Can someone tell me if the ESRAM clock is necessary the same as the GPU clock?
Or could it be possible that the ESRAM is still 800 MHz while the GPU has been bumped up to 875 mhz? (Not that I've heard that is the case)
Or could it be possible that the ESRAM is still 800 MHz while the GPU has been bumped up to 875 mhz? (Not that I've heard that is the case)
- Status
