Bookmark and Share


An independent chip expert has managed to examine die shots of system-on-chips that power the latest game consoles from Microsoft Corp. and Sony Computer Entertainment Division. The SoCs resemble each other in many ways since they share architectures as well as are made using the same process technology. Nonetheless, the dieshots prove that the processors are different.

Chipworks, an independent chip analytics company, has dissolved packaging of Microsoft Xbox One and Sony PlayStation 4 and managed to make two high-quality die shots of the chips under a microscope. The chips are identical in die size: the PS4 SoC is 348mm2, whereas the X1 SoC is 363mm2. Given similarities in clock-rates of the hybrid processors, they should have identical power consumption as well.

Microsoft Xbox One system-on-chip

The die shots reveal two quad-core Jaguar x86 modules with some shared elements within the each SoC.  It is hard to estimate whether Microsoft and Sony altered those modules architecturally or not, but in general the quad-core modules look similar.

Sony PlayStation 4 system-on-chip

As expected, the chips have different amounts of AMD Radeon GCN [graphics core next] compute units: the Xbox One SoC features 14 CUs (with 64 stream processors per unit), two of which are reserved for redundancy; the PlayStation 4 SoC sports 20 CUs (with 64 stream processors per unit), two of which are also reserved for redundancy.

Despite of having considerably lower amount of stream processors for graphics, physics and other highly-parallel computations, the Xbox One SoC features 32MB of SRAM cache for graphics + 15MB other caches, which takes a lot of chip space. Therefore, the cost of X1 and PS4 SoCs should be similar. It is also noteworthy that Sony PS4 SoC has more complex, distributed GDDR5 memory controller, whereas Microsoft X1 SoC features less complex distributed multi-channel DDR3 memory controller.

Both system-on-chips have massive uncore space, which suggests that there are lot of proprietary logic and interfaces inside all designed to boost overall compute performance and provide simplified access to that performance by game developers.

The first analysis of very similar chips shows that Sony PlayStation 4 has a higher-performing system-on-chip under the hood. However, the analysis does not exactly reveal all the secret sauces that Microsoft might have put into the Xbox One SoC.

Tags: Microsoft, Xbox, Xbox One, Durango, Sony, Playstation, PlayStation 4, Orbis, AMD, ATI, Radeon, GCN, Jaguar, 28nm, TSMC


Comments currently: 6
Discussion started: 11/28/13 10:49:47 AM
Latest comment: 01/07/14 08:46:31 AM
Expand all threads | Collapse all threads


ps4 has a more unified simplistic look to it on its die
2 1 [Posted by: SteelCity1981  | Date: 11/28/13 10:49:47 AM]
- collapse thread

Oh, right SteelCity...thanks so much for that.
0 2 [Posted by: Herbert Herpe  | Date: 11/28/13 09:07:24 PM]
I guess being a dick is apart of your everyday vocab huh.
1 0 [Posted by: SteelCity1981  | Date: 11/28/13 09:45:06 PM]
I noted his sarcasm added so much to the conversation .....
0 0 [Posted by: caring1  | Date: 11/29/13 04:53:07 AM]

XBO has a bigger chip? What a fail. Whoever thought of that eSRAM idea was an idiot.
They should have just used the same chip in both consoles, and the entire world would be happy.
0 0 [Posted by: Harry Lloyd  | Date: 11/29/13 06:01:34 AM]

PS4 chip has more processing power but is more depandant to RAM bandwidth and latency, while uses GDDR5.

XBO chip has less processing power but has internal SRAM that makes easier to use all their power and not let them idle waiting for data to come from DDR3 RAM.

IDK how GDDR5 is (bandwidth and latency) compared to DDR3, this is what will matter to compare their effective usable power.

We also must know if 32MB is enough to store all textures and poligons that are processed over time. In a game that has little motion and just changes some stuff in a almost-fixed scenario, XBO should be very faster. In example, a game like SC2 while camera is stopped and only units are moving and there are many units of the same type sharing same textures. But in games with lots of motions, like first person shooters with big and complex scenarios, that 32MB cache won't be so useful.
0 0 [Posted by: Hikari  | Date: 01/07/14 08:46:31 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture