Xbox One (Durango) Technical hardware investigation

[steveOrino]

I do agree that the size and configuration of the box is up to their designers, but it's AMD's parts that are the scaffolds of these two structures. i'm thinking the chips inside of these two consoles could have been at two different stages when their case designers were assembling them. The chip sizes of the two consoles should be at the same efficiently levels since it's coming from the same provider.

Yeah, I donno, so hard to tell until there are tear-downs of each console. X1's APU is supposed to be bigger... also there is a lot of dead space between the APU and memory in the X1 due to DDR3 trace routes among other things. Sucks Sony did not show whats inside to kill time guessing till these things are released.

 

[COPS N RAPPERS]

Yeah, I donno, so hard to tell until there are tear-downs of each console. X1's APU is supposed to be bigger... also there is a lot of dead space between the APU and memory in the X1 due to DDR3 trace routes among other things. Sucks Sony did not show whats inside to kill time guessing till these things are released.

Since we've seen Microsoft's solution for stomping out heat, What i'm guessing is that sony plans to use is a ventilation chamber setup that draws the air out the back. (given those massive holes for vents.)

the problem that I see for this is that this was similarly used for the original xbox 360's setup, which got a bad rep. my guess Is that they plan to use a substantial heat sink and fan on the necessary parts along with fans out the back.

my worries go to sony because this is their first time working with AMD and yet claiming something big to fit into a small box, and to be commercially sold at a reasonable fee. I'm guessing that sony might also believe that even if their ps4s were to malfunction, they could be easily replaced at the cost of little.

 

[TrungGap]

My guess is that the PS4 will be very quiet as well. Xbox One seems to be designed to be quiet as well as being dead simple to manufacture and service. PS4 seems like it would be a little more complex to assemble and disassemble.

Ding! Ding! Ding! We have a winner!...

Edit: MS wants cheaper manufacturing cost...cheap factory. It's not that MS can't build quality hardware...they probably designed for a cheaper manufacture.

 

[Arwin]

Sony is just better with hardware, and Microsoft with software?

 

[zupallinere]

given those massive holes for vents

That whole back is holes !! I "see" all of that volume of air and I don't worry all that much. I can't wait for the tear down either.

 

[XpiderMX]

Ding! Ding! Ding! We have a winner!...

Edit: MS wants cheaper manufacturing cost...cheap factory. It's not that MS can't build quality hardware...they probably designed for a cheaper manufacture.

The new Xbox 360 model looks cheaper to manufacture too.

 

[Oboro Shogun]

I've been reading about them using 700 series cards for their demonstrations. There's no way they'd do that really. I mean show the games at E3 and then sell a console with on a card that could be as low as 1/4 the horsepower. Maybe for the multiplatform games but not the first party stuff. They've got enough bad press at the moment as it is. I don't feel they were unbelievable looking games. We've got at least 6 times the power of the 360 here. Some 360 games still impress me. Something like Ryze looks very possible on 6 times the power of a machine that will run Alan Wake or a decent Crysis port.

 

[Silent_Buddha]

The significant amount of venting on the Xbox One allows for a number of short and straight paths from the heatink to the exterior of the box. The upper vents may compromise its stackability, but passive or almost passive air flow could be enough for that setup at the low load power modes.

Sony's solution is compact, and it seems to leave the broad panels unvented for stacking or aesthetics, but it looks like it relies on more directed and actively driven airflow.

These are my thoughts as well. Assuming the heatsink for the APU is large, it might be designed to run entirely with passive cooling except when there is a game running. At that point any game sounds will drown out any fan noise in a sound proofed room (like many home theater setups).

For something like that, you definitely DO NOT want any fans running period when watching a movie. In room like that even the quietest fans are quite noticeable. Hence why actively cooled A/V receivers aren't terribly popular (are any still made?). Also why A/V receivers with integrated amplifiers are often very large with most of the space inside being empty except for some relatively beefy heatsinks. It's all about the passive cooling ability when they might potentially be in a home entertainment cabinet with other electronics stacked on top.

My relatively high end Sony A/V receiver/amp features most of the venting on top, but due to how boxy and open the inside is, it doesn't suffer from lack of cooling even with other equipment stacked on top of it.

I really don't think stacking anything on top of the Xbox One will cause any issues with cooling unless you are stacking something on it that wasn't designed to be stacked in a home entertainment cabinet.

IMO, from a home theater POV, I think MS might have been better off making the Xbox One even a little bit bigger to enhance the passive cooling capabilities. But I'm guessing it's as relatively small as it is (for A/V equipment) to not draw comparisons to the size of the original Xbox which many thought was too large.

I fully expect the PS4 to be quiet, but louder than the Xbox One. Especially as it doesn't look like passive cooling could be an option with the design they have chosen.

Interesting that Microsoft went with a hybrid design heavily weighted towards A/V equipment design with a nod towards console design (next gen gaming making a purely passively cooled design extremely difficult). While Sony went with a very console oriented design.

Regards,
SB

 

[SenjutsuSage]

Question: When AMD in making their move to GCN decided to separate their ROPs from their memory controllers because their ROPs weren't as efficient as they wanted them to be, and instead decided to service their ROPs through a crossbar which allowed them to increase the number of memory controllers, thus improving ROPs performance by simply giving them access to more memory bandwidth, did this come at a cost of any kind even though it ultimately increased ROP performance?

And with respect to what the core issue was that caused them to decouple the ROPs from the memory controllers in the first place, and with respect to the potential tradeoff/tradeoffs that likely came about as a result of their decoupling of the ROPs from the memory controllers, is there anything in particular about the inclusion of low latency SRAM that can be used as a GPU memory source that strikes anybody as especially interesting considering AMD's decision to decouple ROPs from memory controllers with Tahiti?

 

[3dcgi]

Question: When AMD in making their move to GCN decided to separate their ROPs from their memory controllers because their ROPs weren't as efficient as they wanted them to be, and instead decided to service their ROPs through a crossbar which allowed them to increase the number of memory controllers, thus improving ROPs performance by simply giving them access to more memory bandwidth, did this come at a cost of any kind even though it ultimately increased ROP performance?

I'm not sure why you think decoupling the ROPs was due to them not being efficient. Sometimes flexibility is an advantage as it allows for new configurations.

 

[Phil]

The Xbox One better be extremely quiet in every day use, considering that it wants to be the center of your livingroom, i.e. will be on most of the time. I'm assuming here that if you have a cablebox connected through the Xbox One's HDMI in, that the Xbox One will have to be running to actually use the cablebox.

 

[SenjutsuSage]

I'm not sure why you think decoupling the ROPs was due to them not being efficient. Sometimes flexibility is an advantage as it allows for new configurations.

Indeed that's true, but there are no potential latency drawbacks to their decision? And is there really no other way to potential improve ROP performance other than to increase the memory controllers and provide more bandwidth to the ROPs.

Low latency SRAM can't be a way of improving ROP performance outside of the method chosen by AMD with GCN? So perhaps they weren't inefficient, but they can become even more efficient under certain configurations?

 

[warb]

The Xbox One better be extremely quiet in every day use, considering that it wants to be the center of your livingroom, i.e. will be on most of the time. I'm assuming here that if you have a cablebox connected through the Xbox One's HDMI in, that the Xbox One will have to be running to actually use the cablebox.

I assume it will be completely silent in standby waiting for the "Xbox on" command. Possibly silent at the dashboard running TV pass through and simple apps. The big fan whiring to life for high power modes.

There are also hefty vents on the new Kinect. The intention seems to be for always on and ready.

 

[Phil]

Yeah, I expect as much to be honest. I just hope it's less noisy than the PS3 slim (not the ultra slim model). Even running the basic XMB, it's far too noisy.

 

[3dcgi]

Indeed that's true, but there are no potential latency drawbacks to their decision? And is there really no other way to potential improve ROP performance other than to increase the memory controllers and provide more bandwidth to the ROPs.

Low latency SRAM can't be a way of improving ROP performance outside of the method chosen by AMD with GCN? So perhaps they weren't inefficient, but they can become even more efficient under certain configurations?

There might be a latency penalty to the crossbar, but I doubt it was significant. I'd only expect low latency frame buffer memory to allow for smaller cache sizes not necessarily to improve efficiency. I haven't given it much thought though so there might be certain uses cases I'm ignoring.

 

[Dave Baumann]

Question: When AMD in making their move to GCN decided to separate their ROPs from their memory controllers because their ROPs weren't as efficient as they wanted them to be, and instead decided to service their ROPs through a crossbar which allowed them to increase the number of memory controllers, thus improving ROPs performance by simply giving them access to more memory bandwidth, did this come at a cost of any kind even though it ultimately increased ROP performance?

Actually there isn't really a change there. The RBE's are more closely aligned with the shader engines and that's pretty much been the case on all unified shader engine products. Its just case of whether the number of shader engines/RBE's map directly to the number of memory channels or whether a crossbar is placed between the RBE's and the memory channels.

 

[jonabbey]

Sorry, what are RBE?

 

[liolio]

Render Back Ends

 

[Esrever]

I assume it will be completely silent in standby waiting for the "Xbox on" command. Possibly silent at the dashboard running TV pass through and simple apps. The big fan whiring to life for high power modes.

There are also hefty vents on the new Kinect. The intention seems to be for always on and ready.

bigger fan and heatsink = quieter system. With a fan than big for a chip with that kind of TPD, it will likely be decently quiet even with a game on.

 

[deanos]

Microsoft: Xbox One is three times as powerful as Xbox 360

Jimmy Fallon at one point asks Phil Spencer if Forza 5 (which they demo) could have been made for the Xbox 360. Spencer responses surprisingly by saying that it couldn’t but how he describes the Xbox One is sure to get a lot of attention.

Spencer said:

Absolutely not. Xbox One is three times the power of what the Xbox 360 was. You’ve got incredible graphics, incredible physics, incredible sound.

Even if the One is under-clocked, only 3x the 360? is this even possible?