Xbox One X (Scorpio) SoC Discussion

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

What's inside Scorpio's SoC?

  • Jaguar CPU Cores + Polaris Based (GFX8) GPU

    Votes: 42 30.2%
  • Jaguar CPU Cores + Vega Based (GFX9) GPU

    Votes: 43 30.9%
  • Zen CPU Cores + Polaris Based (GFX8) GPU

    Votes: 16 11.5%
  • Zen CPU Cores + Vega Based (GFX9) GPU

    Votes: 38 27.3%

  • Total voters
    139
  • Poll closed .

Sweepr

Diamond Member
May 12, 2006
5,148
1,143
136
DigitalFoundry said:
"Typically for CPU, the top two items are frequency and memory latency. If the CPU has data, the faster it can process it, the quicker the result, but it also means that if it doesn't have the data, it sits there idle, so latency is a big component. On frequency, we pushed it up to 2.3GHz" explains Nick Baker "On the latency, a couple of the areas we tackled, one was all the queues coming back from the memory interface, we sped those up as well. Specifically, within the core, because we're running a virtualised OS environment, we wanted to optimise how memory translation operations happen so there are some key changes inside the core to speed those things up. The end result is that not only does the CPU run faster, it also runs more efficiently meaning more power for you at the end."

...According to Goossen, some performance optimisations from the upcoming AMD Vega architecture factor into the Scorpio Engine's design, but other features that made it into PS4 Pro - for example, double-rate FP16 processing - do not. However, customisation was extensive elsewhere. Microsoft's GPU command processor implementation of DX12 has provided big wins for Xbox One developers, and it's set for expansion in Scorpio.
 
  • Like
Reactions: Olikan
May 11, 2008
22,223
1,413
126

I have not found where the qoute comes from, but this if true is interesting. Because it mitigates the use of the jaguar and any complaints people might have over using jaguar:


edit: Found it:

http://www.eurogamer.net/articles/digitalfoundry-2017-project-scorpio-tech-revealed
"We essentially moved Direct3D 12," says Goossen. "We built that into the command processor of the GPU and what that means is that, for all the high frequency API invocations that the games do, they'll all natively implemented in the logic of the command processor - and what this means is that our communication from the game to the GPU is super-efficient."

Processing draw calls - effectively telling the graphics hardware what to draw - is one of the most important tasks the CPU carries out. It can suck up a lot of processor resources, a pipeline that traditionally takes thousands - perhaps hundreds of thousands - of CPU instructions. With Scorpio's hardware offload, any draw call can be executed with just 11 instructions, and just nine for a state change.

"It's a massive win for us and for the developers who've adopted D3D12 on Xbox, they've told us they've been able to cut their CPU rendering overhead by half, which is pretty amazing because now the driver portion of that is such a tiny fraction," adds Goossen.


[UPDATE 7/4/17 20:44: Microsoft's Andrew Goossen has been in touch to clarify that D3D12 support at the hardware level is actually a part of the existing Xbox One and Xbox One S too. "Scorpio builds on the Command Processor capability present in the original Xbox One," we're told. "Our implementation of D3D12 supports all Xbox Ones, and games have already shipped that use it. When a game using D3D12 starts up, we reprogram the GPU's Command Processor front-end. The 50 per cent CPU rendering overhead improvement was reported by shipping games. The amount of win is dependent on the game engine and content, and not all games will see that size of improvement. Scorpio's Command Processor provides additional capability and programmability beyond what Xbox One/Xbox One S can do. We plan to take advantage of this in the future."]

To be clear, then: Project Scorpio doesn't feature Ryzen cores, but the Xbox team are not so concerned about this. "On the CPU side of things, we could still meet our design goals with the custom changes we made," Kevin Gammill points out. "At the end of the day we are still a consumer product. We want to hit the price-points where consumers want to purchase this. It's about balancing the two."
 

SpaceBeer

Senior member
Apr 2, 2016
307
100
116
Wow, thank you for this response :)

But at one point, MS and Sony will have to move to new (x86) architecture and break compatibility with current consoles. Just like before. I know it was important to keep XB1 and PS4 alive since users would be very angry if those were disconitnued only 3-4 years after release. But when their EOL come, PS4 Pro and Scorpio will still be quite new. And since consoles use semi-custom chips, couldn't MS and Sony ask AMD to make modified/custom Zen core with all the same instructions as Jaguar?

Zen has much better IPC and it is also very efficient bellow 3 GHz. So even if some Jaguar optimized code runs slower than Zen optimized would, it should still be faster on Zen. Or at least as much as fast as on Jaguar?
 
May 11, 2008
22,223
1,413
126
Wow, thank you for this response :)

But at one point, MS and Sony will have to move to new (x86) architecture and break compatibility with current consoles. Just like before. I know it was important to keep XB1 and PS4 alive since users would be very angry if those were disconitnued only 3-4 years after release. But when their EOL come, PS4 Pro and Scorpio will still be quite new. And since consoles use semi-custom chips, couldn't MS and Sony ask AMD to make modified/custom Zen core with all the same instructions as Jaguar?

Zen has much better IPC and it is also very efficient bellow 3 GHz. So even if some Jaguar optimized code runs slower than Zen optimized would, it should still be faster on Zen. Or at least as much as fast as on Jaguar?

Well, in all honesty , the game developers normally use profiling tools to see where the bottlenecks are. These tools automate the searching for bottlenecks in software. Slowdowns or low fps numbers can be often be solved that way.
But once in a while, programmers who are more like geniuses and extremely hardware savvy than others and who love to dig into the hardware find solutions to speed up code. This does not happen often but it does happen.


Why not using zen ?
I will just leave this quote from the article :
To be clear, then: Project Scorpio doesn't feature Ryzen cores, but the Xbox team are not so concerned about this. "On the CPU side of things, we could still meet our design goals with the custom changes we made," Kevin Gammill points out. "At the end of the day we are still a consumer product. We want to hit the price-points where consumers want to purchase this. It's about balancing the two."

In the future, zen or derivatives from zen will highly likely end up in the next console revisions.
I myself am waiting for a hbm2 based console. I expect a possible future where microsoft will also promote the gameconsole as a pc system where windows 10 software can run on.

Also, backwards compatibility for games can also be solved by recompiling and profiling the game engine to be able to run on the new system. Usually, the graphic data like textures can be reused and the sound files as well. This does require efforts from the game developer so, there must be a financial reimburse for the developers to do so. But if that is the case, the user would not know, since the games store server can query what console revision is requesting the game and provide the required files that are needed for that given console revision. This would solve any compatibility issues and hide all the headaches from the end user. The user just downloads the needed files and can start playing.I do not know if this is actual the case already but i would not be surprised if it would become reality.

edit;
added some text.
 

AMDisTheBEST

Senior member
Dec 17, 2015
682
90
61
scorpio supports 4k! if it costs just $500, doesn't this make all the gaming pcs in the sub $1000 range irrelevant?
 
May 11, 2008
22,223
1,413
126
I should note, if you would take a high end pc like
scorpio supports 4k! if it costs just $500, doesn't this make all the gaming pcs in the sub $1000 range irrelevant?

Well, it is a kind of trickery.

$1000 dollar pc within a few months can actually render all the pixels on a 4k screen. Consoles use a trick.
https://en.wikipedia.org/wiki/Checkerboard_rendering

https://www.extremetech.com/gaming/235274-how-sonys-playstation-4-pro-does-4k-without-delivering-4k

I assume scorpio will also use some sort of checkerboarding.
 

AMDisTheBEST

Senior member
Dec 17, 2015
682
90
61
I should note, if you would take a high end pc like


Well, it is a kind of trickery.

$1000 dollar pc within a few months can actually render all the pixels on a 4k screen. Consoles use a trick.
https://en.wikipedia.org/wiki/Checkerboard_rendering

https://www.extremetech.com/gaming/235274-how-sonys-playstation-4-pro-does-4k-without-delivering-4k

I assume scorpio will also use some sort of checkerboarding.
Oh no, it's native 4k.
https://www.google.com/amp/s/www.fo...pio-native-4k-60-fps-with-power-to-spare/amp/
This console pretty much murders all the flagship GPUs and $1000+ pc of two gen ago(starting at 780ti and older) and will compete with the performance of $500-1000 gaming pc for years to come.
 

Olikan

Platinum Member
Sep 23, 2011
2,023
275
126
So... they did some changes in the jaguar core, but the claims are so evasive...
"Less memory latency" could just mean a new IMC.
 
May 11, 2008
22,223
1,413
126
Oh no, it's native 4k.
https://www.google.com/amp/s/www.fo...pio-native-4k-60-fps-with-power-to-spare/amp/
This console pretty much murders all the flagship GPUs and $1000+ pc of two gen ago(starting at 780ti and older) and will compete with the performance of $500-1000 gaming pc for years to come.

Well, i am interested.
If i take a generic RX480 and take a well optimized game such as doom with vulcan (which should be similar to directx 12 in capabilities i think) it does 40fps at 4k.
If 4 extra CU units (2560 shaders total) combined with all optimizations and improved command processors could do 60fps. That is amazing.
 
Last edited:
May 11, 2008
22,223
1,413
126
So... they did some changes in the jaguar core, but the claims are so evasive...
"Less memory latency" could just mean a new IMC.

Yeah, from ddr3 to gddr5.
Or deeper queues. Stuff like more cache storage for TLB seems to help a lot too.

https://en.wikipedia.org/wiki/Translation_lookaside_buffer

A Translation lookaside buffer (TLB) is a memory cache that is used to reduce the time taken to access a user memory location.[1][2] It is a part of the chip’s memory-management unit (MMU). The TLB stores the recent translations of virtual memory to physical memory and can be called an address-translation cache. A TLB may reside between the CPU and the CPU cache, between CPU cache and the main memory or between the different levels of the multi-level cache. The majority of desktop, laptop, and server processors include one or more TLBs in the memory management hardware, and it is nearly always present in any processor that utilizes paged or segmented virtual memory.

The TLB is sometimes implemented as content-addressable memory (CAM). The CAM search key is the virtual address and the search result is a physical address. If the requested address is present in the TLB, the CAM search yields a match quickly and the retrieved physical address can be used to access memory. This is called a TLB hit. If the requested address is not in the TLB, it is a miss, and the translation proceeds by looking up the page table in a process called a page walk. The page walk is time consuming when compared to the processor speed, as it involves reading the contents of multiple memory locations and using them to compute the physical address. After the physical address is determined by the page walk, the virtual address to physical address mapping is entered into the TLB. The PowerPC 604, for example, has a two-way set-associative TLB for data loads and stores.[3] Some processors have different instruction and data address TLBs.
 

Sweepr

Diamond Member
May 12, 2006
5,148
1,143
136
Now that we know RX 580 is a glorified 185W Polaris card, I'm more impressed that Microsoft managed to pack 44 EUs (40 EUs active) at relatively high clocks in their Scorpio Engine. 245W power supply for the entire system, including 12GB of GDDR5.
 

Topweasel

Diamond Member
Oct 19, 2000
5,437
1,659
136
Now that we know RX 580 is a glorified 185W Polaris card, I'm more impressed that Microsoft managed to pack 44 EUs (40 EUs active) at relatively high clocks in their Scorpio Engine. 245W power supply for the entire system, including 12GB of GDDR5.
Some of that is also efficiency on the CPU end. The new Jaguars are clocked higher but I bet also use less power.
 

imported_bman

Senior member
Jul 29, 2007
262
54
101
The RX480 gets a lot out of memory OCing, a 15% boost to memory clocks will result in around a 9% increase in performance. So I think the additional bandwidth (around 1.2x available to GPU over the RX480) + additional CUs while accounting of the frequency (1.03x Tflops) would make the Scorpio around 1.17x faster than a RX480. This would be about 5% faster than a 1060 6GB, maybe it can get to around the performance of a 980Ti (which is 1.32x faster than a 480) with optimizations. The 1070 is 1.50x faster than a 480, so I think it is out of reach even with console optimizations.
 
Last edited:

dogen1

Senior member
Oct 14, 2014
739
40
91
We don't know what it's closest to. Something like a 480, but it could have some vega features which might put it at a higher performance level.
 

imported_bman

Senior member
Jul 29, 2007
262
54
101
I also wonder about the L2 cache. We know that the L2 cache in Polaris got an overhaul over GCN 1.2 and that the Scorpio's has 4x the L2 GPU cache over the One. Is the 4x increase in L2 GPU cache just result of moving to Polaris or is it an additional customization?
 

NTMBK

Lifer
Nov 14, 2011
10,411
5,677
136
I also wonder about the L2 cache. We know that the L2 cache in Polaris got an overhaul over GCN 1.2 and that the Scorpio's has 4x the L2 GPU cache over the One. Is the 4x increase in L2 GPU cache just result of moving to Polaris or is it an additional customization?

I would guess it's a deliberately requested change. Render targets are getting 4X bigger for 4K, makes sense to quadruple the cache.
 

CatMerc

Golden Member
Jul 16, 2016
1,114
1,153
136
Some games will be native 4k, some will use checkerboarding or other techniques.
It is built to run anything that runs on Xbox One, only at 4K.
Microsoft is very clear in marketing it as a "true 4K" console.
 

french toast

Senior member
Feb 22, 2017
988
825
136
So what is the most comparable GPU in the scorpio? 1080? 1070?
The scorpio gpu is more advanced and powerful than a Polaris 10, it has more cache, double the rops, probably 2x fp 16? a better command processor, higher flops vs rx 480, much more bandwidth (polaris is very bandwidth sensitive) and likely many other small improvements.
Its not really comparable to a straight pc gpu, if it were put in a pc to run un optimised pc games it would probably be gtx 1070 level.

As scorpio is optimised at an unprecedented level, the whole chip its self is probably equal to a gtx 1080 + i7 6700k combo in actual gaming output.
They had forza engine running at native 4k 60fps at pc ultra settings, the gpu was topping 88% usage, this was up in a very short time with no specific scorpio architecture optimisations, pretty incredible.

Edit, Scorpio does not have 2x fp 16 like neo, but does have some vega features, also 60 changes to the rendering pipeline over xbox one or polaris? Improvements to cpu-gpu coherency, reduction of draw calls by 50% to the cpu, extra function blocks to the jaguar cores to improve latency as well as other latency reducing improvements in the memory pipeline.
 
Last edited:

dogen1

Senior member
Oct 14, 2014
739
40
91
It is built to run anything that runs on Xbox One, only at 4K.
Microsoft is very clear in marketing it as a "true 4K" console.

Anything that was 900p and up at 4k.

Even then I wouldn't be surprised if 900p and 1080p xb1 games used checkerboard or similar techs in order to increase visuals in other areas, while still having near native 4k quality.
 
Last edited:

imported_jjj

Senior member
Feb 14, 2009
660
430
136
https://twitter.com/XboxQwik/status/869985791832870912
DBLPdsVU0AAo2lZ.jpg:large