Monday, June 12th 2017

Xbox One X Hardware Specs Give Gaming Desktops a Run for their Money

Microsoft Sunday dropped its mic with the most powerful game console on paper, the Xbox One X, formerly codenamed "Project Scorpio." The bottom-line of this console is that it enables 4K Ultra HD gaming at 60 Hz. Something like this requires you to spend at least $1,200 on a gaming desktop right now. Unlike a Windows 10 PC that's been put together by various pieces of hardware, the Xbox One X is built on a closed ecosystem that's tightly controlled by Microsoft, with heavily optimized software, and a lot of secret sauce the company won't talk about. The console still puts up some mighty impressive hardware specs on paper.

To begin with, at the heart of the Xbox One X is a semi-custom SoC Microsoft co-developed with AMD, built on TSMC's 16 nm FinFET node (the same one NVIDIA builds its "Pascal" GPUs on). This chip features a GPU with almost quadruple the single-precision floating point compute power as the one which drives the Xbox One. It features 40 Graphics CoreNext (GCN) compute units (2,560 stream processors) based on one of the later versions of GCN (likely "Polaris"). The GPU is clocked at 1172 MHz. The other big component of the SoC is an eight-core CPU based on an unnamed micro-architecture evolved from "Jaguar" rather than "Bulldozer" or even "Zen." The eight cores are arranged in two quad-core units of four cores, each; with 4 MB of L2 cache. The CPU is clocked at 2.30 GHz.
The third major component of the Xbox One X SoC is the 384-bit wide GDDR5 memory controller, wired to 12 GB of memory. This memory is used both as system- and graphics-memory, and is the most ideal implementation of AMD's hUMA (heterogeneous unified memory architecture), where there's no visible partition between the system and graphics memory on the physical memory, and depending on the usage scenario, any amount of memory can be used by the CPU and GPU components. Developers are still forced to build their games under the assumption that the system only has 8 GB of memory; so that the remaining 4 GB is used as a kind of "guarantee" that 4K UHD @60 Hz runs smoothly. The total memory bandwidth available is a staggering 326 GB/s.

The SoC features an integrated audio CODEC with 7.1-channel output over HDMI, with support for Dolby Atmos, and HRTF, a new audio format Microsoft developed for the Hololens, which is optimized for VR.

A 1 TB 2.5-inch SATA hard drive comes standard on the Xbox One X. You can swap this drive out for larger HDDs, or faster SATA SSDs. You can also plug in external storage devices over the console's USB 3.0 ports. The console's operating system resides on a smaller eMMC chip that isn't accessible to end-users. The 1 TB HDD is used to store games you've downloaded from your online library à la Steam.

Microsoft switched from bulky external power bricks to internal PSUs with the Xbox One S, and the trend carries forward with the Xbox One X. Powering the whole thing is a 275W internal power-supply. A large fan-heatsink cools the SoC and GDDR5 memory chips.
Source: Eurogamer.net
Add your own comment

132 Comments on Xbox One X Hardware Specs Give Gaming Desktops a Run for their Money

#101
londiste
Nihilus
Does anyone even know the IPC of Jaguar? 3ghz Kaby quad = '12 ghz'. 2.3 ghz jaguar octo = '18.4 ghz'. For the i5 to be nearly twice as fast or '36 ghz of jaguar' it would need 3x the IPC.
look for a4-5000 reviews/benchmarks and multiply the result by 2 for best case scenario.

example: cinebench r15 single-core result for a4-5000 (@ default 1500mhz) should be 30-35.

there are considerable architectural differences between (ultra)mobile soc-s with jaguar cores (or atom cores) and today's full-fat deskop cpu-s.
Posted on Reply
#102
Nihilus
rtwjunkie
He doesn't quite understand the term. :rolleyes:
Where is the confusion? I was just trying to compare IPC.
londiste
look for a4-5000 reviews/benchmarks and multiply the result by 2 for best case scenario.

example: cinebench r15 single-core result for a4-5000 (@ default 1500mhz) should be 30-35.

there are considerable architectural differences between (ultra)mobile soc-s with jaguar cores (or atom cores) and today's full-fat deskop cpu-s.
So piledriverish? It will be interesting to see if the cpu holds it back.
Posted on Reply
#103
MxPhenom 216
ASIC Engineer
Nihilus
Does anyone even know the IPC of Jaguar? 3ghz Kaby quad = '12 ghz'. 2.3 ghz jaguar octo = '18.4 ghz'. For the i5 to be nearly twice as fast or '36 ghz of jaguar' it would need 3x the IPC.
Thats not how you calculate IPC lmao
Posted on Reply
#104
cadaveca
My name is Dave
BiggieShady
Oh please, this is purely subjective, so it's totally understandable that you hate to say it :) probably you just gotta connect the PC to the same display and adjust yourself some gamma and digital vibrancy
Subjective, sure. You can also add into it that I have quite the plethora of PC hardware to play games on, and even so, I still prefer some titles on console. I prefer though... my own opinion. As all of our eyes and ears are of varying capability, you can only expect any of this PCMR stuff to be purely subjective... because many of the ideas and complaints on the subject of PC vs console are going to depends on a person's ability to perceive the environment around them. Some might infer that being so sensitive to the minor differences makes discussion on this matter a moot point.
Posted on Reply
#105
FordGT90Concept
"I go fast!1!11!1!"
It's about on par with a RX 580 8 GiB which is very respectable for a $500 console. As always, technology will move on for PC gamers while it only gets die shrinks until it is discontinued.

That said, Microsoft seriously came out swinging because of the flak they got for putting out a weaker-than-PS4 Xbox One. One could argue they over-corrected though, as reflected by the $500 price tag. I think too many companies are going too gun-ho over VR when the demand really isn't there.
Posted on Reply
#106
efikkan
Nihilus
Does anyone even know the IPC of Jaguar? 3ghz Kaby quad = '12 ghz'. 2.3 ghz jaguar octo = '18.4 ghz'. For the i5 to be nearly twice as fast or '36 ghz of jaguar' it would need 3x the IPC.
That's not how it works at all, you can't add performance of cores. Rendering is in reality more or less limited to one core, with some minor loads delegated to others, meaning the fastest core will be the main deciding factor for performance.
Kaby Lake have ~>60% better IPC, combined with significantly higher clocks.
Posted on Reply
#107
Neosabre
T4C Fantasy
In ff14 bard with controller can out dps easilly a mouse and keyboard
Oh hell no! I can't play bard in 14 on a console controller. I've been bard for years and I main it and when I tried the gamepad I threw up in my mouth!
Posted on Reply
#108
Neosabre
Hugh Mungus
Not all and probably no switching, just multiple platforms since xbox consoles are now competitive with playstations again. Now the xbox one x is a good option for nice visuals as well.
Seems like somebody doesn't know what a first party studio is.
Posted on Reply
#109
lexluthermiester
btarunr
Xbox One X Hardware Specs Give Gaming Desktops a Run for their Money
That HIGHLY depends on your perspective. If referring to raw computational power, then this statement is hilariously laughable. Now if you are referring to "bang/buck", you're getting closer to reality, but such a statement is still woefully misleading. The chipset "Scorpio" is based on is competent, but is not a match for even a mid-range gaming focused laptop, let alone a desktop. It's like saying the Nintendo Switch is on a similar performance level to the PS4. It is not, and making such a statement is as foolish as it is inaccurate. That is not to say the Switch or Scorpio are not great gaming platforms. But great hardware depends on great software. When you have great software, the hardware becomes less relevant.

This kind of statement leads to misunderstandings and misinformation which are a part of what causes so much mistrust within the industry at Microsoft, aside from their own shenanigans.
rtwjunkie
Thing is, it will give upper mid-end PC's a run for their money....for about 6-8 months. As those PC users upgrade, console gets left behind.
You can't be serious with that... My 4 year old pc is better spec-wise than the Scorpio.
rtwjunkie
As to the 4K 60fps claim, I call Bullshit. Cherry picked titles, but not across the board. A 1080Ti cannot even do that in every game. Maybe they will manage it by giving crappy visuals at that resolution.
At 4k res, turn off AA completely. It isn't needed at that resolution and will completely waste GPU resources. At 4k, I'm seeing a solid 120fps minimum on most games. AA is off and AF is on 4X[forced through the driver]. Performance is smooth and fluid. And before you argue, TRY IT!
Posted on Reply
#110
Nihilus
MxPhenom 216
Thats not how you calculate IPC lmao
I wasn't calculating IPC, I was comparing. It is easiest to do this by multiplying. E.g. An 8 core cpu running at 2.5 Ghz has the same performance as a 4.5 Ghz quad. '20 Ghz > 18 ghz, so the 8 core has lower IPC.
Posted on Reply
#111
Nihilus
lexluthermiester
...The chipset "Scorpio" is based on is competent, but is not a match for even a mid-range gaming focused laptop, let alone a desktop...

You can't be serious with that... My 4 year old pc is better spec-wise than the Scorpio.
Lots of BS here as well.
It seems you guys still can't wrap your head around anything more than 'raw specs'. Most only care about the end result.
Posted on Reply
#112
lexluthermiester
Nihilus
Lots of BS here as well.
It seems you guys still can't wrap your head around anything more than 'raw specs'. Most only care about the end result.
Hmm, let us review...
lexluthermiester
But great hardware depends on great software. When you have great software, the hardware becomes less relevant.
Right. That's what I thought I said.. Perhaps you should learn to read statements COMPLETELY before quoting them, eh? Just throwing it out there..
Posted on Reply
#113
MxPhenom 216
ASIC Engineer
Nihilus
I wasn't calculating IPC, I was comparing. It is easiest to do this by multiplying. E.g. An 8 core cpu running at 2.5 Ghz has the same performance as a 4.5 Ghz quad. '20 Ghz > 18 ghz, so the 8 core has lower IPC.
That is not how it works. All youre doing is multiplying the clock frequency by the number of cores the CPU has. There is not a single performance aspect of a chip that you would do such a calculation. That is what noobs try to do to sound like they know what they are talking about.

To compare IPC, you need to be able to calculate it correctly first and fore most, otherwise you look like a doofus.
Posted on Reply
#114
Hugh Mungus
Neosabre
Seems like somebody doesn't know what a first party studio is.
Nowhere near all ps4 exclusives are from first party studios.
Posted on Reply
#115
MxPhenom 216
ASIC Engineer
Hugh Mungus
Nowhere near all ps4 exclusives are from first party studios.
If Microsoft doesnt get any new exclusives for the One X as launch titles, its going to have a rough start. Just releasing a powerful new system isnt enough to dig themselves out of the whole the original One put them in. l have more faith in Phil Spencer than I ever did in Mattrick though so.
Posted on Reply
#116
Hugh Mungus
MxPhenom 216
If Microsoft doesnt get any new exclusives for the One X as launch titles, its going to have a rough start. Just releasing a powerful new system isnt enough to dig themselves out of the whole the original One put them in. l have more faith in Phil Spencer than I ever did in Mattrick though so.
Xbox one exclusives, not one x and apparently quite a few games ran 4k within a day and the console is essentially as powerful as a 1070/16gb ram combo.
Posted on Reply
#117
Nihilus
MxPhenom 216
To compare IPC, you need to be able to calculate it correctly first and fore most, otherwise you look like a doofus.
Ok then calculate it.
Posted on Reply
#118
Nihilus
The claim was a 3 ghz quad Kaby has twice the power of a 2.3 octo jaguar. What would the IPC difference have to be?
Posted on Reply
#119
MxPhenom 216
ASIC Engineer
Nihilus
The claim was a 3 ghz quad Kaby has twice the power of a 2.3 octo jaguar. What would the IPC difference have to be?
You can't really determine IPC from just knowing the frequency and number of cores in the CPU though. You need to know how many instructions have been executed during a specific time frame. A CPU could have different IPCs depending on the instructions its executing since not all instructions require the same amount of time to be executed in the pipeline.

IPC = Instructions per cycle. The name alone should be enough to realize you can't multiply frequency and physical core count to determine it for a given CPU.

IPC is greatly affected by the CPUs cache and memory heirarchy as well. And for that you have to know the "stalls" at each level, if the levels are unified or not (L1 cache are split into data and instruction portions) but l2 and l3 are typically unified and then you need to know the percentage of data that is pulled from each level.
Posted on Reply
#120
MxPhenom 216
ASIC Engineer
Nihilus
The claim was a 3 ghz quad Kaby has twice the power of a 2.3 octo jaguar. What would the IPC difference have to be?
The Jaguar chip would have worst IPC across the board due to lack of L3 cache.

I dont know about the twice the power claim. Those jaguar cores are weak as hell though.
Posted on Reply
#121
Hugh Mungus
MxPhenom 216
The Jaguar chip would probably have worst IPC across the board due to lack of L3 cache.

I dont know about the twice the power claim though. Those jaguar cores are weak as hell though.
Custom jaguar though and gddr5 ram. Each new xbox/ps4 console has improved cpu ipc AND faster ram. The massive ram improvement, faster speed and higher ipc combined are going to make a huge difference. Desktop ipc probably still is twice as high or more, but the point is custom jaguars aren't that terrible anymore and are comparable to at least a decent i3 of some sort.
Posted on Reply
#122
Nihilus
I was multiplying to get a common denominator. If a 2 core cpu running at 4 ghz has the same multi core performance of a 4 core cpu running at 2 ghz, wouldn't they have the same IPC?
Posted on Reply
#123
Hugh Mungus
Nihilus
I was multiplying to get a common denominator. If a 2 core cpu running at 4 ghz has the same multi core performance of a 4 core cpu running at 2 ghz, wouldn't they have the same IPC?
Yes
Posted on Reply
#124
efikkan
lexluthermiester
At 4k res, turn off AA completely. It isn't needed at that resolution and will completely waste GPU resources. At 4k, I'm seeing a solid 120fps minimum on most games. AA is off and AF is on 4X[forced through the driver]. Performance is smooth and fluid. And before you argue, TRY IT!
If you think you don't need AA in 4K, then you need to get your eyesight checked.
AF at 4X is going to give you terribly blurred textures in certain angles. Not even a Titan Xp can pull the heaviest current games in 4K at 60 Hz at the highest detail level, so a gaming console with a fraction of the performance is certainly not going to do that.
Nihilus
The claim was a 3 ghz quad Kaby has twice the power of a 2.3 octo jaguar. What would the IPC difference have to be?
IPC is performance per core. Of course more cores may provide more total theoretical performance, but in real workloads things don't scale like that. And since we're talking of a gaming console, more cores is certainly not going to mitigate the slow single core performance.

The task of rendering can only be split to a small degree, so in any game the performance of a single core is the deciding factor for real performance. The other cores mainly help to do non-rendering tasks, OS tasks etc.

To estimate IPC you take one or more (scalable) benchmarks, and divide the score by clock speed.
Posted on Reply
#125
MxPhenom 216
ASIC Engineer
Nihilus
I was multiplying to get a common denominator. If a 2 core cpu running at 4 ghz has the same multi core performance of a 4 core cpu running at 2 ghz, wouldn't they have the same IPC?
Do they are the same core architecture?
Posted on Reply
Add your own comment