Back at CES Nvidia unveiled the Tegra 4 chipset, which it dubbed the fastest mobile processor on the planet. Contentious as the claim was it integrated 4 ARM Cortex A15 core with a phantom core and had a 72-core GPU and could do HDR for video and stills simultaneously through what it called Computational Photography. However, it lack one key feature critical especially in developed markets – An integrated LTE modem. The Tegra 4i comes in and solves this problem as it becomes Nvidia’s first SoC to combine a software programmable modem. Also Read - NVIDIA RTX 3080Ti, 3070Ti graphics cards unveiled along with AMD RX 6000M series cards
Codenamed grey, the Tegra 4i should not be mistaken for an ultra high-end smartphone SoC. In terms of sheer horsepower, it is lower than the Tegra 4 in the pecking order. Instead of the 4 plus 1 Cortex A15 architecture on the Tegra 4, it subplants lesser 4 plus 1 Cortex A9 cores. The GPU core count also goes down from 72 on the Tegra 4 to 60 on the Tegra 4i. That said, the top clock speed goes up to 2.3GHz and now Nvidia calls Computational Photography Chimera. Also Read - Colorful iGame GeForce RTX 3060 bilibili E-sports Limited Edition launched: Price in India, specifications
Another difference is in the RAM type. The Tegra 4 supports 4GB DDR3L and LPDDR3 RAM, while the Tegra 4i only supports 2GB of LPDDR3 RAM. Also Read - Nvidia GeForce RTX 30 series GPUs launched for entry level gaming laptops
These variances become more obvious when we talk about resolutions that can be driven by it. While the Tegra 4 is capable of driving up to 4K resolutions, devices powered by the Tegra 4i are restricted to 1920×1200. That said, both the chipsets are made on the same 28-nanometer process.
But the Tegra 4i remains a big play for Nvidia. Partially this is because it expects within a year the chipset will power $200 smartphones. To this end, Nvidia has even released a reference platform called the Phoenix that has a 5-inch 1080p display and is only 8mm thick. While this is definitely high-end for early 2013, by 2014 this would not remain the same, and Nvidia plans on taking on the likes of Qualcomm and Intel with a powerful chipset reference design of its own.