AMD Ryzen 4000 Mobile: Technical and Building Features

AMD has introduced a large number of APU models in the market, as it has never done before, indicating that it is here to stay and deal with both 14 nm for Intel and 10 nm for its competitors.

The Ryzen 4000 Mobile APUs are divided into two completely separate phases, this is the H series and the U series, with a focus on high-performance gaming computers in the former case, while the latter is specifically addressed for the travel and battery life market.

AMD Ryzen 4000 Mobile

When we look at the 11 APUs, the three HS variants AMD has on the market are amazing. This is different than the low TDP and with the exception of the Ryzen 9 4900HS running at 100 MHz under its brother, some keep all aspects of their original versions.

However, purchasing these HS processes requires direct interaction from AMD and the manufacturer, because they are specialized and refined versions of silicon.

Of all the grades and series of H there are no surprises, the larger version means not only the speed in the trees, but the greater speed in the GPU.

The U series is also compatible with this design, except that all APUs that maintain 15 watts of rigor are powered by AMD.

RAM and PCIe 3.0 speed

As usual all APUs will have dual memory controls, where one lies in support of DDR4-3200 MHz or LPDDR4X-4266 MHz up to 64 GB with 51.2 GB / s in first instance and 32 GB in -68.3 GB / s per second.

Unlike its desktop versions, AMD has decided to leave PCIe 3.0 on its new APUs. The reason is that PCIe 4.0 requires a lot of processing power and logically APUs don't take advantage of the remaining bandwidth.

Where SSD will affect fast, because even though it has 16 PCIe rows (8 for iGPU and 2 X4 SSD links or storage in general) we should prefer PCIe 3.0 hard disks instead of 4.0.

More frequency, more IPC and Vega at 7 nm

Although the number of CUs has been reduced from 11 to 8 (we would assume that we will lose more work) the fact that some upgrades to the CPU mean jumping to more + 59% return for each CU enabled even though using Vega architecture like all previous APUs.

AMD says the peak performance will remain the same 1.79 TFLOPS in FP32, which after being reduced to 8 CU will consider +27 power compared to GPU as Vega 11.

Redesigned infinity fabric

Another development involving these APUs is the fact that the IF frequency is disconnected from RAM memory. This is possible because unlike desktop Ryzen, here we have a monolithic design and not MCM chiplets, which is one of the reasons why AMD doesn't have standard APU designs yet.

Reduced means that the IF can operate where the low frequency is required, thus saving energy, and increasing the high frequency where there is sufficient workload.

In addition, saving about 75% of the energy in the IF has resulted in AMD being able to add new coding engines without incurring additional energy costs when used. So, now we have it HDR / WCG for HEVC, VP9 Gen 2, H.264, and H.265.

AMD SmartShift and STT V2

SmartShift is a new firmware-based solution that requires collaboration between the CPU and iGPU service on the Ryzen 4000 Mobile. What AMD has achieved is that when one of the two is idle, the carrier can use the remainder of the TDP to increase the frequency and efficiency.

It is a type of low-load TDP, either a CPU or an iGPU that will improve the chip performance without skipping the manufacturer's indicated TDP.

Also, SmartShift an integral part of Smart Temperature Following V2 or STTv2, which helps the system increase long-term performance due to the installation of high-speed opu control probes on the APU.

The learning of these new processes passes directly to the IF in an existing invoice controller, which determines if the system has room for its turbo modes and how long they can be.

In other words, a large set of readings and values ​​are taken to evaluate the complete thermal profile, thus ensuring that the Boost remains highly active for seconds and thus achieves additional performance.