9.30.2020- 11:00 a.m ksa
On Tuesday, British chip design company Arm announced the launch of three new automation device chips suitable for vehicles, including the high-performance ARM Cortex-A78AE CPU, the Mali-G78AE GPU and the Mali-C71AE image signal processor.
These three chips are worthy of consideration because all of them have security features built-in. This implies that the chip has additional functions to ensure that a double check essentially runs through any calculation.
Traditionally, for its CPU, Arm has supported two modes. Both cores function independently in the “split mode” and only occasionally go offline for fast tests. For applications with low to no security requirements, this is very appropriate because the kernel will run close to its full capacity.
In “locked mode”, the kernels run in pairs, and their operations are cross-checked with each other. This helps these chips meet various automotive safety requirements, but it also brings significant performance loss, because users can only use half of the core capacity.
Today, Arm has introduced a new hybrid mode for its CPUs. In high-performance use cases that require only moderate failure detection, this mode combines the best results in both areas. It allows the kernel to still run in split mode, but the shared cluster logic of the integrated kernel now runs in locked mode. This provides a safety mechanism for locked mode, and at the same time has the performance of split mode.
In Mali-G78AE GPU, Arm introduced so-called “flexible partitioning”, which makes it easier to split various GPU cores between different workloads as needed.
This means that functions such as maps can be operated in a zone independent of safety functions such as driver monitoring or running instrument clusters. Arm’s current goal is to apply it to the broader market for autonomous driving systems.
Although Arm remained silent on NVIDIA’s $40 billion acquisition transaction and is still undergoing regulatory procedures, it is worth noting that both companies have set their sights on this automation system market. NVIDIA provides its own platform for automated robots, such as Launched Jetson AGX, in addition to NVIDIA’s own GPU, it also uses Arm’s CPU.