When AMD launched its Ryzen 4000 H-series (portable) APUs, it generated great expectations, something perfectly understandable since we were facing a high-performance generation that combined a Zen 2 CPU with a Radeon Vega GPU, both manufactured in a 7 nm process.
The performance improvement offered by these new APUs, compared to the previous generation, is more than considerable, although it is especially noticeable at the CPU level. Overall AMD did a good job with Ryzen 4000 H-series APUs, but a recent review by IgorsLab confirms that, for some reason I can’t quite understand, the Sunnyvale company decided it was a good idea to limit them to eight PCIe Gen 3.0 lines, instead of using the classic 16-line setup.
This is a problem because, as many of our readers will know, the number of lines available determines the total bandwidth and therefore affects the gross performance that a specific graphics card can develop. So that you have a clearer vision I leave you a graph with the available bandwidth in each type of PCIE Gen3 configuration.
On paper there is a huge difference that suggests that a top of the range graphics card could be very limited when using a PCIE Gen3 configuration like the one that APUS Ryzen 4000 H series assemble, but is it really so? It makes sense, and helps us understand why most laptops that have hit the market with such APUs are limited to mounting an RTX 2060 in almost all cases. However, the truth is very different, as we will see below.
Ryzen 4000 H-series and PCIE Gen x8 APUs: is there a bottleneck?
As we can see in the attached graphics, courtesy of TechPowerUP !, there is a loss of performance, but it is so minimal that the difference between mounting an RTX 2080 Ti under PCIE Gen3 x8 and doing it in an x16 configuration barely reaches 3% on average. There is, therefore, no real bottleneck, and this is not why manufacturers have chosen not to offer laptops with Ryzen 4000 H-series APUs and more powerful graphics cards at the moment.
Why then do we have a majority of settings based on the RTX 2060? I can’t safely answer that, as I’m not on the OEMs’ minds, but it’s likely all due to an attempt to balance temperatures and costsThat is, to keep notebooks with Ryzen 4000 H series APUs as very interesting options from a price-performance perspective.
I remind you that the Ryzen 4000 desktop APUs that were recently launched do not have this problem, that is, they come with a configuration that allows the GPU to access an interface PCIE Gen3 x16. I hope this article helps you, by the way, dispel all the myths that still exist around the subject of PCIE standards, lines, and the impact of bandwidth.