It looks like Samsung is starting to get serious with the development of its own GPU (dubbed the SGPU). According to sources, the company hired Dr. Chien-Ping Lu – and industry veteran in the realm of GPU – late last year to spearhead the company’s project.
Dr. Chien-Ping Lu, who is also referred to as CP by people close to him, formerly worked at NVIDIA in his early years. Specifically, the man was responsible for developing NVIDIA’s nForce integrated graphics processor (IGP), a GPU module built into both AMD and Intel motherboards, before GPU were ultimately built together with CPUs.
In later years, CP moved over to MediaTek, where he helped the company develop its own mobile GPU. However, just four years into its development, MediaTek canceled the project, for reasons known only to them.
At current, CP is leading Samsung’s engineering teams based in Austin and San Jose, who are dedicated to “hardware, software, test, verification, and implementation”.
While there still isn’t a lot of details regarding its hardware, Samsung claims that in simulations, the SGPU delivers performance-per-watt numbers that were better than expected. The simulations also indicated that the SGPU showed promise in the reduction of latency in motion-to-photon path, which could lead to “a form of low-power consumption VR with fast recovery, very high resolution dynamic display capability”.
No official timeline has been given for the SGPU’s release, but it’s possible that Samsung could debut its GPU module in future Exynos chips as early as next year.