News & Rumors: Xbox One (codename Durango)

[eastmen]

Are those porta potties on the outside? Preparation for people shitting themselves en masse upon reveal..?

most likely for construction workers and set up crews. I guess also for during the show . They most likely don't want people running to the buildings each time they need to use the potty

 

[Rangers]

On the 360, voice commands used somewhere around 10-12MB, with an upper limit of a few dozen commands, with the ability to switch in new banks of commands. The fewer the commands, the better we can do at discriminating them, it's a completely statistical model, after all. Skeletal I'm not sure about but it was probably in the 10-20MB range.

The upper limit is not known, probably as much as you're given. it all depends on how you store the database, how much data you have, whether you're prioritizing accuracy or speed etc.

devs are going to be like kids in a candy store next gen, going from a handful of ram to gobs.

i mean, 32mb for the os is just ancient these days. suddenly they will get 2gb.

 

[marcberry]

4 days

 

[marcberry]

And the usable data per frame has very little to do with the total texture memory budget. Take any game you'd care to play, in how many of them can you see the entire level at full texture resolution all the time? Say Skyrim, just because you can't see the scenery behind you doesn't mean it's textures are not in RAM. Gears of War, it uses streaming textures now, but you could preload those textures with more RAM, so there's no pop-in. right there, you have a use case for 2x the texture memory from what you can address in a frame. Megatexturing would benefit greatly from more RAM, since you can have more tiles in memory, and manage your tiles more efficiently.

And that's just textures.

http://www.tomshardware.com/reviews/...cn,3104-2.html
GCN 1.0 states that per CU their are 4 16wideSIMDs (vector units) with 64ALUs with one scalar unit. So why in VGleaks are their 4 VSPs ( vector scalar processors) per SIMD?
That's not right.

 

[Betanumerical]

http://www.tomshardware.com/reviews/...cn,3104-2.html
GCN 1.0 states that per CU their are 4 16wideSIMDs (vector units) with 64ALUs with one scalar unit. So why in VGleaks are their 4 VSPs ( vector scalar processors) per SIMD?
That's not right.

Its because microsoft/vgleaks love to rename things, it means the same thing either way.

 

[marcberry]

Its because microsoft/vgleaks love to rename things, it means the same thing either way.

That's call a no answer. Nice try.

 

[Rangers]

i guess the latest rumors if you can even call them that being hinted around neogaf are 20 cu's/12gb/64mb esram....one or all.

i dont even want 64mb esram (too expensive for the value)

12gb ram keeps getting thrown around. heck i would think 16gb would be more likely than 12 since just doubling chip density gets you 16 not 12.

reiko is hinting around something about compute. both sides really going for compute. etc etc.

if i was to make a guess based on his posts i might think he heard something like "MS added a discrete-ish 6-8 CU compute module at the last second" or something.

might as well go wild with the rumors with 4 days left.

 

[(((interference)))]

I'm guessing out of all of those, the 12 GB is the most likely. No idea how they could increase the CU count or ESRAM now.

Don't see how much benefit 12GB will bring though, they're BW limited more than anything.

Strange that the rumours aren't mentioning clock increases, as that's the most plausible out of any changes.

But we'll see in 4 days I guess.

 

[Xenio]

i guess the latest rumors if you can even call them that being hinted around neogaf are 20 cu's/12gb/64mb esram....one or all.

i dont even want 64mb esram (too expensive for the value)

12gb ram keeps getting thrown around. heck i would think 16gb would be more likely than 12 since just doubling chip density gets you 16 not 12.

reiko is hinting around something about compute. both sides really going for compute. etc etc.

if i was to make a guess based on his posts i might think he heard something like "MS added a discrete-ish 6-8 CU compute module at the last second" or something.

might as well go wild with the rumors with 4 days left.

64 eSram? well I think that 32 MB should be enough, doubling this will double a LOT of transistor without a serious improvement, so will not happen unless they will go with dual APU that will hardly happen too.

adding CU's? another thing that hardly will happen unless the 12 CU's are reservation for gaming and there're CU's not documented to developers (Es. 4 for system/video/kinect computing and 2 for improving yelds, this is, why not, possible but expensive until serious shrink process will happen)

in the end they can do the trick saying @E3 or @ reveal day: "This is the Xbox [insert name], the only console in the world with real 1 GHz GPU and 16 (or 12, anyway >8) solid GB of unified RAM"

hardly the user will be capable to understand that 18 CU's @ 800 MHz are slightly faster than 12 CU's @ 1 GHz and what he understand will be that Xbox is faster and with more ram -> superior

this kind of marketing tricks are in the style of our dear console manufacturers, only few thousand persons in the world will understand the meaning of those numbers.

 

[Betanumerical]

That's call a no answer. Nice try.

Its called the truth, if you looked through the diagrams and descriptions of everything you'll see that everything perfectly aligns for it be stock standard GCN. Now because vgleaks/microsoft has taken some liberty in naming parts of the same architecture (for god knows why) something different some people (and I'm not saying you) have taken to mean that its for some reason some super new modified architecture, its not, its bog standard GCN once you do the maths.

Its a little frustrating to be honest, I wish they didn't change the description.

 

[Rangers]

I'm guessing out of all of those, the 12 GB is the most likely. No idea how they could increase the CU count or ESRAM now.

Don't see how much benefit 12GB will bring though, they're BW limited more than anything.

Strange that the rumours aren't mentioning clock increases, as that's the most plausible out of any changes.

But we'll see in 4 days I guess.

well i dont know how much you keep up with neogaf, but for example this

http://www.neogaf.com/forum/showpost.php?p=57921432&postcount=12283

http://www.neogaf.com/forum/showpost.php?p=57915720&postcount=12247

i'm not really sure how reliable reiko is. not super reliable i think. a little prone to being prolite-ish

from what little i know of reiko's new source, i'm a bit skeptical.

 

[kots]

i'm not really sure how reliable reiko is. not super reliable i think.

Isn't he a rabid fanboy ? Now he's an insider too ?

 

[Rangers]

Isn't he a rabid fanboy ? Now he's an insider too ?

He is not an insider just a rumor-monger

 

[Deleted member 7537]

20 CU and DDR3?

 

[Xenio]

20 CU and DDR3?

what are you asking for? for feasibility or you're ironic or are you're just guessing?

anyway I think 20 CU's is out of possibility, the only remote possibility can be durango has the same gpu of what you know with 18 CU, with 4 for kinect (skeletal tracking, matching etc) and for system (realtime video effects, morphing, computing); but this is unlike, 4 CU's are too much.. and two CU for conservative yield

PS. if this gpu is a 18 cu's with 2 disable and 4 reserved, I think we can forget at all the upclock to 1 GHz, it's too fat -> too much heat

 

[Deleted member 7537]

what are you asking for? for feasibility or you're ironic or are you're just guessing?

anyway I think 20 CU's is out of possibility, the only remote possibility can be durango has the same gpu of what you know with 18 CU, with 4 for kinect (skeletal tracking, matching etc) and for system (realtime video effects, morphing, computing); but this is unlike, 4 CU's are too much.. and two CU for conservative yield

PS. if this gpu is a 18 cu's with 2 disable and 4 reserved, I think we can forget at all the upclock to 1 GHz, it's too fat -> too much heat

just wanted to point out that binding a 20 CU GPU to a slow RAM doesnt seem to be a very efficient design. that is why i'm skeptical about these huge changes in specs, they would require a full redesign of the console.

 

[DJ12]

Isn't he a rabid fanboy ? Now he's an insider too ?

Come on now you know the only thing you need to do to be an insider is say something fanboys want to hear.

 

[Xenio]

just wanted to point out that binding a 20 CU GPU to a slow RAM doesnt seem to be a very efficient design. that is why i'm skeptical about these huge changes in specs, they would require a full redesign of the console.

it's a 68 GB/s DDR3, how can you label it "slow"?
and you're forgetting about eSram too, why should they redesign the whole console, included memory system, for 12-13-14-15-X CU's gpu when the memory system was already designed as they want it to be?

 

[upnorthsox]

Are those porta potties on the outside? Preparation for people shitting themselves en masse upon reveal..?

So the porto-potties are so secret sauce laden that they too come with covers???

Seriously though, my guess would be chairs.

 

[marcberry]

Its called the truth, if you looked through the diagrams and descriptions of everything you'll see that everything perfectly aligns for it be stock standard GCN. Now because vgleaks/microsoft has taken some liberty in naming parts of the same architecture (for god knows why) something different some people (and I'm not saying you) have taken to mean that its for some reason some super new modified architecture, its not, its bog standard GCN once you do the maths.

Its a little frustrating to be honest, I wish they didn't change the description.

I can see you have not done your homework, so i will help you, but where will i start? He is the short way. Here is GCN 1.O in detail.
http://www.tomshardware.com/reviews/...cn,3104-2.html
After that read this one. You will find that all this is FACT.
http://www.amd.com/la/Documents/GCN_Architecture_whitepaper.pdf
1. AMD uses CUs (not SCs but we will for now) inside of that are called Vector Units that are 16wideSIMDs in sets of 4, we see this in VGleaks as the SCs to 4 SIMDs. This is GCN101.
2. Inside each CU has four Vector Units (VUs), each with 16 ALUs, for a total of 64 ALUs per CU. We see this in VGLeaks: Compute: SC to 4 SIMDs.
3.Inside each CU their is 1 SCALAR processor for all 4 SIMDs.
Every thing i just told you was a FACT..... This is in a CU of GCN1.0.
Now here is where SuperDae missed. NOWHERE in the documents anywhere on the internet can tell us WHY is their 4 VSPs in each SIMD and not just 1 in the CU, but we see this in VGLeaks: Compute: SC to 4 SIMDs to 16 VSPs, not 1 for the CU but 16. You can not dismiss this. "its bog standard GCN" not with 16 VSPs ( Vector scalar processors).
If it had only one VSP your right, it's GCN101. It's not.

Do you wont me to do the math? it's ......