AMD Radeon 780M “RDNA 3” Integrated GPU Benchmarks Show 26% Boost Over RDNA 2, Comparable To NVIDIA RTX 2050

The first performance benchmarks of AMD's upcoming RDNA 3 integrated GPU, the Radeon 780M, have been reported by Golden Pig Upgrade.

The AMD Radeon 780M will be the fastest integrated GPU when it launches, leveraging the brand-new RDNA 3 GPU architecture. The RDNA 3 iGPU lineup will initially include two parts, the Radeon 780M and the Radeon 760M. The Radeon 780M will feature 12 Compute Units for a total of 768 SPs and a clock speed of up to 3 GHz while the Radeon 760M will feature 8 Compute Units for a total of 512 SPs and a clock speed of up to 2.8 GHz.

These iGPUs will ship with AMD's Ryzen 7040 "Phoenix" APUs which will carry a TDP rating of 35-45 Watts. Additionally, these chips are expected to retain all of their Radeon-exclusive features such as raytracing, FSR, RSR, and other tech support when the laptops hit shelves in March this year.

Golden Pig Upgrade, a well-known tech creator at Bilibili has posted the first performance numbers of the AMD Radeon 780M integrated graphics. The graphics benchmarks were run within 3DMark Time Spy and two different configurations were used. Both configurations featured LPDDR5X but the first one used LPDDR5X-7500 and the second DDR5-5600 memory. Considering that APUs have been vastly dependent on memory bandwidth, the LPDDR5X solution undoubtedly delivers the best performance results.

With the LPDDR5X memory, the AMD Radeon 780M RDNA 3 GPU scored around 3000 points while the DDR5 solution scored around 2750 points in 3DMark Time Spy graphics. This marks an increase of 26% over its predecessor, the Radeon 680M RDNA 2 GPU. The GPU is easily able to beat last-gen popular laptop GPUs such as the GTX 1050 Ti and even comes close to the RTX 2050 laptop GPU.

Since the launch is still a few months away, we can expect AMD to optimize its iGPU driver future to deliver even better performance. We expect that by launch, the Radeon 780M with faster memory solutions and overclocks will be easily able to outperform the RTX 2050 from NVIDIA. This is also why NVIDIA decided to finish its MX line of entry-level chips and focus on discrete GPUs since the iGPU landscape is getting more competitive with each generation.