Google Pixel 7, Pixel 7 Pro Tensor G2 SoC Built on 5nm Process, Not 4nm Claims Report

Google’s Pixel 7 and Pixel 7 Pro were launched in India earlier this month. The company’s latest Pixel-branded smartphones debuted with the next-generation Tensor G2 SoC, which was earlier reported to be built using Samsung’s 4nm fabrication process. However, according to a new report, the Tensor G2 SoC is said to utilise a 5nm fabrication process instead of 4nm. Google at its keynote announcing the Pixel 7 series with the Tensor G2 SoC, claimed that the processor on the Pixel 7 and Pixel 7 Pro would offer a 60 percent speed gain and 20 percent battery efficiency gain over its predecessor, the Tensor G1 SoC, which was seen on the Pixel 6 series. Google didn’t actually confirm if this improvement is coming from an improved fabrication process or other optimisations and tweaks.

Google in interaction with Android Authority has confirmed that the Tensor G2 SoC used in the Pixel 7 and Pixel 7 Pro is built upon a 5nm process, instead of the earlier predicted 4nm process.

The nanometer measurement on a chipset is the half gate pitch or half of the distance between the “Fins” of a transistor gate. Considering there are billions of transistor gates on a single SoC, a nanometer of difference in size would make a significant improvement in terms of power consumption, space utilisation, and performance. The decreased size of an individual transistor means that more of them could fit into the same-sized chip, making the chip more powerful.

Google, however, claims to have focused on Tensor G2 SoC’s machine learning capabilities enabled through a next-generation TPU rather than breaking heads over the fabrication process.

“We purpose-built Google Tensor G2 for real-world use cases. Our final architecture, which includes 5nm, helped us reach that goal while increasing both performance and power efficiency. This approach also allowed us to add new capabilities while taking a step forward on machine learning with our next generation TPU with G2,” said a Google spokesperson in the interaction with Android Authority.

Google, however, did not confirm who built the chip, while it is widely known that Samsung is the company’s chief partner. Samsung has two 5nm nodes in its stratagem, which include the 5LPE process used for the original Tensor G1 SoC, and a newer 5LPP process. Even so, the process used for the Tensor G2 SoC has not been confirmed by the company.

The Tensor G2 SoC was previously reported as having benchmark clock speeds that did not show a significant enough improvement for an improved fabrication process. The figures were interpreted by Android developer Kuba Wojciechowski basis a Geekbench listing that clocked two 2.85GHz ARM Cortex-X1 cores, as well as two ARM Cortex-A78 cores at a 2.35GHz clock rate, and the four ARM Cortex-A55 cores at 1.8GHz.

Google’s upcoming Tensor G2 to use the same CPU cores as the first gen Tensor, Mali-G710 GPU – details below 👇 🧵

— Kuba Wojciechowski⚡ (@Za_Raczke) September 16, 2022

Furthermore, the Tensor G2 SoC is said to be paired with a Mali-G710 GPU for improved gaming and HDR performance. The Google Tensor G2 SoC also could be integrated with a second-generation EdgeTPU AI processor. It is reportedly compatible with LPDDR5 RAM. It features 4MB of shared L3 cache and 8MB of total system cache.

The next-generation Tensor G2 SoC is said to feature an improved image signal processor (ISP), which enables it to record 4K 60fps videos on all cameras. The chipset can also reportedly handle a 108-megapixel camera without stutter or lag.

Whether the revelation made on the Tensor G2 SoC being built as a 5nm chip instead of the more advanced 4nm makes a significant difference to the competitive position and actual performance that the Pixel 7 and Pixel 7 Pro will adorn, or if it will remain a minor on-paper detail, remains to be seen.

Affiliate links may be automatically generated – see our ethics statement for details.

Source link

Source: News

Add a Comment

Your email address will not be published. Required fields are marked *