The three new chipsets from Qualcomm promise support for dual cameras, better performance, pin compatibility, and dual VoLTE.
Adding to the existing 600 and 400 series of Snapdragon chipsets, Qualcomm has introduced three new mid-tier mobile chipsets, Snapdragon 632, Snapdragon 439, and Snapdragon 429. The three chipsets find a place for themselves between the less common Snapdragon 200 series and the flagship-level Snapdragon 800 series.
The Snapdragon 632 chipset is powered by an octa-core Qualcomm Kryo 250 CPU, an Adreno 506 GPU, and a Snapdragon X9 LTE modem. It is capable of supporting either a single 24-megapixel camera or a dual 13-megapixel camera, and an FHD+ display. The chipset is pin-compatible with its sister models, Snapdragon 626, Snapdragon 625, and Snapdragon 450.
The Snapdragon 439 chipset is powered by octa-core ARM Cortex-A53 CPU, an Adreno 505 GPU, and a Snapdragon X6 LTE modem. It is capable of supporting either a single 21-megapixel camera or a dual 8-megapixel camera, and an FHD+ display.
The Snapdragon 429 chipset is powered by a quad-core ARM Cortex-A53 CPU and an Adreno 504 GPU. It borrows the Snapdragon X6 LTE modem from the Snapdragon 439. It is capable of supporting either a single 16-megapixel camera or a dual 8-megapixel camera, and an HD+ display. Both the Snapdragon 439 and Snapdragon 429 are pin compatible with each other.
The Kryo 250 CPU and the Cortex-A53 CPU doing duty on the Snapdragon 632 chipset and the Snapdragon 439 chipset respectively have four performance cores and four efficiency cores. Kryo 250 sees an increase of up to 40 percent in performance, while Cortex-A53 sees an increase of up to 25 percent. The GPUs Adreno 506, Adreno 505, and Adreno 504 see a performance increase of up to 10 percent, 20 percent, and 50 percent respectively.
The new Snapdragon chipsets are designed to support dual SIM slots with both SIMs supporting VoLTE.According to Qualcomm, they also have support for Qualcomm Neural Processing SDK and Android NN. Frameworks supported are Caffe/Caffe 2, TensorFlow/TensorFlow Lite, and ONNX.