Microsoft entered the console market between 2001 and 2002 with the first Xbox, a console well above its competitors in the technical part and inheriting the PC hardware. Almost twenty years later, every desktop console on the market follows the same philosophy as the original Xbox, but each has its own unique design personality.
The Xbox Series X SoC.
The Xbox 360 SoC is a custom chip created by AMD for Microsoft that was custom made and is therefore exclusive to the Microsoft console. It is a 360mm2 monolithic processor, which has the following components inside:
- Processor based on AMD Zen 2 architecture.
- Custom GPU based on AMD RDNA 2 architecture
- Accelerators specializing in specific hardware tasks
We’ll go through the different parts of the Xbox Series X SoC point by point throughout this article. because it is the main unit which includes all the hardware except the system RAM and its SSD storage.
Xbox Series X CPU
The Xbox Series X processor is the exact same processor AMD built into the SoCs they released for laptops and PCs in 2020 under the Ryzen 4000 brand name and codenamed Renoir. These are “scaled-down” versions of the Ryzen 3000 desktop cores. The reason we put it in quotes is that, except for the L3 cache which was cut for space reasons, the rest of the chip and its functionality remains the same, unchanged.
Unlike its counterpart on PlayStation 5, with the same architecture, it can work in two different ways:
- In the first mode, it works with the active SMT and can therefore execute 2 threads per core. In this mode, it operates at 3.6 GHz, which makes it 0.1 GHz faster than its counterpart on the PlayStation 5.
- In the second mode, the SMT is inactive, but in return each core operates at 3.8 GHz. This mode is specially designed for Xbox One (X) games running on 8 AMD Jaguar cores without multhilo support.
Using the Zen 2 architecture represents a big leap, the new processor not only doubles the clock speed of standard Xbox One models, but also has more than double the power per clock cycle, assuming the one of the biggest leaps between generations of consoles. .
The Xbox Series X GPU
|Basic architecture||AMD RDNA 2 custom|
|Base clock speed||1825 MHz|
|Increase clock speed||He does not have|
|ALU FP32 by CU||64|
|Power in FP32||12 TFLOPS|
|Power in FP16||24 TFLOPS|
|Power in Int 8||32 TOPS|
|Texturing rate||379.6 G pixels / s|
|Filling rate||116.8 GPíxeles / s|
|L2 cache||5 MB|
|Primitive Mesh / Shaders||And|
|Variable rate shading||And|
|Sampler feedback flow||And|
Although they share the same basic architecture, Microsoft took a completely different path than SONY when it comes to designing its console and choosing the GPU. While in PlayStation 5 a design with fewer computing units in the GPU was chosen, compensated by a higher clock speed, in Microsoft’s case a lower clock speed was chosen but a lower clock speed was chosen. greater number of calculation units. According to our knowledge, Microsoft’s choice is better, as consumption increases exponentially with clock speed.
Being uniquely designed, its compute unit configuration does not match any PC model, as it has 52 active compute units running at 1825 MHz, giving it around 12 TFLOPS of power continuously and without fluctuations. Like PlayStation 5, it is a design based on RDNA 2 but it has more of this architecture than its direct rival, which only has RDNA 2, the unit in charge of calculating intersections in Ray Tracing.
The technology that AMD and Microsoft have not integrated is the Infinity Cache, which is to be expected given the large size of the SoC. We can’t forget that we’re looking at AMD’s larger SoC under TSMC’s 7nm node, as the PlayStation 5 is smaller than the Xbox Series X.
Backend + y VRS rendering
ROPS units, which have the function of writing the frame buffer, are called render backends. In RDNA 2 for PC and Xbox, they gained a + at the end of the name due to the fact that with the rasterization unit they were changed to perform variable rate shading. Which consists of grouping the repeated Pixel Shader lines into one to calculate them as one instead of repeating their calculation.
The difference between the RB + PC and the Xbox Series X is that the console is a somewhat reduced unit, because while the PC can write up to 128 32-bit pixels per clock cycle, 128 ROPS, on the console version, the thing stays at 64. Otherwise, the functionality of Variable Rate Shading is intact and is one of the differentials that Microsoft sticks its chest out to when it talks about its new console.
RAM on Xbox Series X
Xbox consoles have always used a UMA system, so we have all the memory in the same well, this time made up of 10 GDDR6 memory chips with asymmetric capacities, because although its bus goes from 320 bits to 320 bits at 14 Gbps, 560 GB / s of bandwidth, we do not find a configuration of 10 GB or 20 GB, which would be usual if we were talking about PC hardware, but 16 GB, because it combines different chips of 1 and 2 GB.
Although we are outside the GPU area, we must clarify that in the case of RDNA 2 architectures, the number of L2 cache partitions is directly proportional to the number of GDDR6 channels in the system. Since GDDR6 is 2 channels per memory chip and we have 10 chips, we are talking about 20 channels and cache partitions on the GPU.
In order to avoid contentions, or conflicts of access to the RAM between the CPU and its clients and the GPU and its clients, the capacity of 16 GB is divided into two different blocks,
The 16 GB of memory are divided into two different blocks, one of 6 GB and the other of 10 GB, which can be accessed under the following conditions:
|Channel 1||Channel 2||Total|
|Zone 1: 10 GB||280 Gb / s GPU||280 Gb / s GPU||560 GB / s|
|Zone 2: 6 GB||56 Gb / s processor||280 Gb / s GPU||336 GB / s|
The reason the processor only accesses 56 GB / s is related to how in Zen architectures the unit called UMC accesses RAM, it always does so at the speed of memclk * 32 bytes / cycle. Since GDDR6’s memclk is 1/8 Gbps, that means it operates at 1.75 GHz, resulting in 56 GB / s bandwidth when the processor through the UMC is using this channel.
Xbox Series X SoC Accelerators
An accelerator is a unit designed to do a specific job and free a more complex processor from that task, there are two types of accelerators:
- Fixed function: these are the ones that are not programmable, they always perform the function in the same way. Due to their simplicity, they take up the least space and consume the least, but their lack of versatility means that their performance is static.
- Specialized processors: these are small processors designed to run firmware, they are usually microcontrollers inside the processor with their own instruction memory and therefore with a Harvard type design some may have access to the memory.
In the Xbox Series X we have a series of accelerators, which can access either of the two RAM sinks described above or both at the same time.
|DMA # 0||Both|
|DMA # 1 + LZ encoding||Both|
|DMA # 2 ´LZ decoding||Both|
|DMA # 3||Ambox|
The organization of these is the same as that of the Xbox One, but the most important of them is in charge of managing access to the SSD, colloquially called by Microsoft Velocity Engine, which implements DirectStorage technology. from Microsoft.
SSD and Velocity Engine
Microsoft took a different path for the SSD, instead of designing their own unit, they decided to install an M.2 2230 type drive, which uses a 2-lane PCI Express 4.0 bus for access. Which gives you 2.4Gb / s of bandwidth, but using the data compressor and decompressor of the SoC, you can achieve 4.8Gb / s of bandwidth. The unit, as you can see in the photo, unlike the PlayStation 5 SSD, is not soldered to the motherboard and although it is slower, it has more storage capacity bringing in 1TB.
To expand storage, Microsoft opted not for M.2 drives but instead for modified CFExpress memory cards, which contain a 1TB capacity M.2 drive. These cards use the other two affected PCI Express lanes. SSD and their access capacities are therefore the same as those of the main unit. The other point of view? Being a single unit Microsoft will control the price very strongly and since storage is required we believe Microsoft could capitalize on its console in this sense.
Velocity architecture and sampler feedback
The unit in charge of managing access to the SSD is the Velocity Engine, because without it you would have to use several CPU cores to decompress the data. If on your PC the CPU is choking while decompressing at a rate of a few hundred megabytes, imagine with gig rates. This is why all new generation consoles have special units dedicated to this task within the main SoC of each of them.
In the case of the Velocity Engine, it is compatible with the DirectStorage API and is designed to apply Sampler Feedback Streaming, which consists of developers when loading a texture which versions will be loaded and which will not. , In order to avoid loading superfluous and therefore unnecessary data into the RAM.