Samsung’s latest-generation Galaxy S23 series models will use a custom Snapdragon 8 Gen 2 for Galaxy processor, and there will be no Exynos version. So how does this custom Snapdragon 8 Gen 2 for Galaxy processor perform?
First look at the public parameters. The Snapdragon 8 Gen 2 for Galaxy processor is equipped with 1 Cortex-X3+2 Cortex-A715+2 Cortex-A710 and 3 Cortex-A510 cores. Among them, the core clock frequency of Cortex-X3 is upgraded from 3.2GHz to 3.36GHz, and the frequency of Adreno 740 GPU is increased from 680 MHz to 719 MHz.
Recently, foreign media Gsmarena conducted a benchmark test on the Galaxy S23 Ultra. The Geekbench 5 CPU test results show that the Galaxy S23 Ultra lags behind the iPhone 14 Pro Max with the A16 processor in both multi-core and single-core scores. Compared with the Android lineup, the single-core test of the Galaxy S23 Ultra is ahead of the Red Magic 8 Pro, but the multi-core test lags behind the latter. According to Gsmarena’s analysis of running scores, the Galaxy S23 Ultra did not use the Cortex-X3 core in the multi-core test, which resulted in unsatisfactory multi-core running scores.
In the AnTuTu benchmark, the Galaxy S23 Ultra lags behind the Red Magic 8 Pro (8 Gen 2). This is because the Red Magic 8 Pro is a game console that actively cools down, and the screen is only 1080p+, which requires less graphics computing power. But fortunately, the Snapdragon 8 Gen 2 for Galaxy is much more powerful than the Snapdragon 8+ Gen 1 processor, and it even has an advantage over the A16 Bionic of the iPhone 14 Pro Max.
In scenarios involving GPU-intensive tasks, Snapdragon 8 Gen 2 for Galaxy performed well, almost leading other phones in the Offscreen test, but was slightly inferior in the Onscreen test due to the high screen resolution affecting the running score.
Combining the two benchmark tests, it is known that the CPU of the Galaxy S23 Ultra is 30-35% faster than the Galaxy S22 Ultra, and the AnTuTu 9 comprehensive benchmark test is regarding 30% faster. GFX Bench and 3DMark GPU tests can provide up to 2 times the performance improvement, and there will still be 80-90% improvement in most other scenarios.
