Predict: The Next Generation Console Tech

[Grall]

I don't really see how stacks of high performance memory would be feasible anyhow. We know that individual high-speed memory chips can burn several watts apiece, you put a bunch on top of each other and cooling that stack is going to be a damn bother.

Stacking low-performance RAM for use in compact devices where board space is a premium sounds like a hell of an idea, but for a console or anything performance critical? Seems dodgy to me.

 

[tunafish]

I don't really see how stacks of high performance memory would be feasible anyhow. We know that individual high-speed memory chips can burn several watts apiece, you put a bunch on top of each other and cooling that stack is going to be a damn bother.

The ram itself doesn't consume all that much, most of the burn is in the interface. Stacking allows you to build absurdly wide interfaces, (think tens of kilobits), which in turn allows you to use really slow interfaces (and slow dram), while still getting great performance.

Stacking low-performance RAM for use in compact devices where board space is a premium sounds like a hell of an idea, but for a console or anything performance critical? Seems dodgy to me.

Present-gen stacks use very narrow interfaces compared to what is possible.

 

[aaronspink]

I don't really see how stacks of high performance memory would be feasible anyhow. We know that individual high-speed memory chips can burn several watts apiece, you put a bunch on top of each other and cooling that stack is going to be a damn bother.

Stacking low-performance RAM for use in compact devices where board space is a premium sounds like a hell of an idea, but for a console or anything performance critical? Seems dodgy to me.

In the case of stacked or MCM'd dram you generally have a much wider interface, on the order of 512-1024b, plus you generally have a much better electrical environment either through the through silicon vias (which electrically look like just bigger version of the vias already used on chips to move between metal layers and run at high speeds with out any fancy drivers and receivers or via the interposer (be it silicon or another material) which is typically within a cm or two from the sender/receiver and as such can also run at high speed. In either case you really aren't burning any more power than the pre-existing big buses on the chips (for DRAM that can be upwards of 4K bitlines). So with stacking you are looking at several Gb/s and same with on MCM and wide interfaces but all running with lower electrical loss than even a normal DIMM of DDR3 memory.

In both cases, outside of the embedded space, you would be looking to use it as a very high bandwidth high capacity cache but with some significant latency increases compared to on die cache (the core of the design is still dram and as such still would have dram device like latencies aka 10s of nS) but it would or should still have lower latency than actual DRAM with all of its overheads. Realistically you would be looking at 128-512 MB per device. Depending on your stack height with HMC, you could probably get 5-8 dies stacked ideally, with MCM'd dram, you are probably looking at 2-4ish devices. With the MCM'd dram you are probably looking at 2-3 gb/s per dram device which works out to 128-196 GB/s per device.

Depending on design it could either be used as a cache or a MM scratchpad. Either way it has a lot of potential both in embedded, graphics, and servers. For graphics, having a ~512MB-1GB 256-512 GB/s scratch pad to use for things like intermediate results and front/back buffers has a lot of advantages. It also allows you to significantly reduce the bandwidth demands on your main memory pool. It isn't hard to imagine a GPU with a 128-192b DDR4 main memory with a HMC/MCM'd high bandwidth scratch memory. For servers, it allows things like relatively cheap and effective 1-2 GB L4 caches. For embedded it reduces board space significantly esp with HMC.

 

[Blazkowicz]

it would have similarity with the xbox 360's daughter e-dram, actually. xbox's edram is an incredibly fast piece of memory, though small and the HMC could be even faster. in the xbox it has ROPs included though, so it is some kind of "smart memory" ; could the same model be followed with HMC, I don't know.

off-die caches also have precedents, implemented in differing manners : pentium pro, pentium II/III, IBM's big multi-chip modules (such as the one with four POWER5 and four L3 dies, it was beautiful)
I wonder how HMC compares with these, and how the data links look like.

 

[french toast]

@ acert: The two things are not related so I don't see what you mean by putting things into "perspective"...

 

[SKYSONY]

There´s a guy in GAF who has posted this:

XBOX720 has:

• stronger CPU
• slightly weaker GPU
• more RAM (8GB?), but slower?
• stronger CPU and RAM may be more towards being a Windows 8 box?

PS4 has:

• weaker CPU
• slightly stronger GPU
• less RAM (2GB, maybe 4GB at max), but faster?

According to bgassassin it´s a good early summary.


I´m to starting to be doubtful about the rumors tha claimed the Xbox was gonna be more powerful than the Sony system. All the latest sign are pointing to an Xbiix, a much more casual game console than the 360.

 

[Grall]

In either case you really aren't burning any more power than the pre-existing big buses on the chips

Ok, well that sounds good, but how's random access going to be on a slow-clocked, 1k bits wide interface...?

I'm also curious what kind of area penalty that has to be paid to drill 1000+ holes in the die... It would also need power and ground as well of course, control signals etc. Would the cost of those things be much cheaper than traditional high-speed bus transcievers?

 

[french toast]

If I'm honest I think all rumours and leaks so far are a load of crap...they change all the time..someone has got to be wrong.

 

[onQ]

There´s a guy in GAF who has posted this:

XBOX720 has:

• stronger CPU
• slightly weaker GPU
• more RAM (8GB?), but slower?
• stronger CPU and RAM may be more towards being a Windows 8 box?

PS4 has:

• weaker CPU
• slightly stronger GPU
• less RAM (2GB, maybe 4GB at max), but faster?

According to bgassassin it´s a good early summary.


I´m to starting to be doubtful about the rumors tha claimed the Xbox was gonna be more powerful than the Sony system. All the latest sign are pointing to an Xbiix, a much more casual game console than the 360.

even if the Xbox Next is only 4X as power as the Xbox 360 with the newer tech devs can push out games that look so good normal everyday people wouldn't even care that it's only 4X as powerful. just look at Last of Us & Beyond if that's being done on a console that's just a little bit more powerful than the Xbox 360 thanks to the Cell helping with the GPU work just think what can be done on a new console with a GPGPU with even more power & ram.

I even think the Wii U is going to amaze people once devs get the hang of it.

 

[aaronspink]

Ok, well that sounds good, but how's random access going to be on a slow-clocked, 1k bits wide interface...?

same as any other DRAM technology but a hair bit faster since you have less latency on the bus, (AKA your prefetch-8 goes across the bus at the same time).

I'm also curious what kind of area penalty that has to be paid to drill 1000+ holes in the die... It would also need power and ground as well of course, control signals etc. Would the cost of those things be much cheaper than traditional high-speed bus transcievers?

The penalties (if they exist) would be in the memory dies. And yes, high-speed tranceivers/pad logic is pretty hefty.

 

[liolio]

same as any other DRAM technology but a hair bit faster since you have less latency on the bus, (AKA your prefetch-8 goes across the bus at the same time).



The penalties (if they exist) would be in the memory dies. And yes, high-speed tranceivers/pad logic is pretty hefty.

Looking at a launch in 2013 waht do you think is the most likely and the less risky wrt mass producing a system?
I mean stacking, wide IO are really promising and game changer but could they make it in a system launching next year?

Am I right to assume that the less risky approach (production pov, so you don't miss your time lines, etc.) would be a big pool of DDR4 and a chip akin to Xenos daughter die?

I've another question, You said earlier that you would want in fact two big cores and 6-8 tinier cores. Looking at what AMD jas now would you see say a Streamtoller module (~2 cores) and 6 bobcat/jaguar cores as a sane proposal?
Or you meant more specialized cores than Bobcat/Jaguar which are just low perfs/low power cores?

 

[Xenus]

I think stacking and wide/IO are 2014 tech rather then next year. Which in turn it would make sense for MS and Sony to hold off on their consoles despite everyone craving something new to not be on the wrong end of that jump.

 

[wco81]

I think stacking and wide/IO are 2014 tech rather then next year. Which in turn it would make sense for MS and Sony to hold off on their consoles despite everyone craving something new to not be on the wrong end of that jump.

Problem is sales are cratering now.

 

[Xenus]

Nothing a significant pricedrop can't fix. Plus sales in the consoles first year generally aren't anything mind blowing and for Sony and MS usually are a net loss anyways. If the technology is within a year or two or possibly less depending on plans and disruptive enough to make a large difference like Wide/IO appears to be in my mind it's worth the risk.

 

[Rangers]

Nothing a significant pricedrop can't fix. Plus sales in the consoles first year generally aren't anything mind blowing and for Sony and MS usually are a net loss anyways. If the technology is within a year or two or possibly less depending on plans and disruptive enough to make a large difference like Wide/IO appears to be in my mind it's worth the risk.

Make a large difference to what? It doesnt appear Sony and MS are all that worried about maxing the power anyway. Using mid to low end PC GPU's already shows theyre not stretching for every last ounce of power.

In other words, they probably dont care. If anything they'd probably be more interested in any cost savings abilities such tech might have.

Everybody wants this newfangled stuff but, I bet it's not going to happen. People always too aggressive...

 

[anexanhume]

Problem is sales are cratering now.

Nothing a significant pricedrop can't fix. Plus sales in the consoles first year generally aren't anything mind blowing and for Sony and MS usually are a net loss anyways. If the technology is within a year or two or possibly less depending on plans and disruptive enough to make a large difference like Wide/IO appears to be in my mind it's worth the risk.

They have a price drop in their repertoire and a good line of AAA games over the next year. Sony isn't going to be in a hurry to put a console out there that loses money since they are bleeding badly. I imagine they will react to MS and try to launch within a month or so.

MS on the other hand has the best selling console now, kinect is doing well and the whole business turns a profit. If they don't feel their hand forced by the WiiU, then only Sony can force it. That's why I don't see a 2013 launch as a foregone conclusion.

 

[Grall]

That's why I don't see a 2013 launch as a foregone conclusion.

Is that a double negative in disguise? I honestly can't tell.

Anyhow, they're gonna launch next gen in '13, there's no way it's going to take longer than that. Haven't they pretty much said so already btw.

 

[Rangers]

They have a price drop in their repertoire and a good line of AAA games over the next year. Sony isn't going to be in a hurry to put a console out there that loses money since they are bleeding badly. I imagine they will react to MS and try to launch within a month or so.

MS on the other hand has the best selling console now, kinect is doing well and the whole business turns a profit. If they don't feel their hand forced by the WiiU, then only Sony can force it. That's why I don't see a 2013 launch as a foregone conclusion.

Hardware and software sales both are plummeting though, I suspect there's a little nervousness at both MS and Sony HQ.

This is June 2012, we're looking to Nov 2013, so that's a lot more slow months to come. I dont see Nov 13 as too early.

Everybody talks price drops, which is true, but I wonder at this point if "profit center" MS isn't waiting for the major Oban/361/whatever redesign to do anything. Sony lost how many billions early in the gen? I think they're in profit mode too.

Plus, you're underestimating the lead time involved. Sony cant just up and decide to launch a month after MS if they haven't already been planning it for ages, doesn't work that way. Hardware redesigns are measured in months and years. The only way Sony can decide to launch a month after Xbox on short notice is if the console is done, millions are already manufactured and sitting in warehouses in the USA just waiting for the go ahead. Boat freight from China alone takes 8+ weeks (alternative is air shipping, much more expensive)

Frankly I imagine the hardware is getting somewhat locked down right now, for a late 13 launch.

 

[Xenus]

They have a price drop in their repertoire and a good line of AAA games over the next year. Sony isn't going to be in a hurry to put a console out there that loses money since they are bleeding badly. I imagine they will react to MS and try to launch within a month or so.

MS on the other hand has the best selling console now, kinect is doing well and the whole business turns a profit. If they don't feel their hand forced by the WiiU, then only Sony can force it. That's why I don't see a 2013 launch as a foregone conclusion.

I hope not that's what appears to have got them in so much trouble with the PS3. They need to choose there own timetable and stick with it. Not Try to fit their time scale to their competitors. Forcing something that's not ready to come out earlier or changing at the last minute cause what you had in mind won't be ready for the new timetable almost always result in bad things.

 

[ninzel]

Hardware and software sales both are plummeting though, .

But what about profits? Wouldn't that be the thing they will look at.