ARM expands its Mali GPU range with two new offerings. The ARM Mali-G52 and ARM Mali-G31. The former one offers High area efficiency while the latter falls under Ultra-Low power GPU category.
Starting with ARM Mali-G31, its ARM’s very first Ultra-Efficient GPU which is based on Bifrost architecture. It not only offers significant energy and area saving but also supports very latest API’s. Statistically speaking, the G31 is 20% smaller that usual mainstream GPU but offers 20% improved performance density for complex workloads. Company says that it has 12% improvement for the growing complexity of UI. Moreover, it is ARM’s smallest GPU to support Open GI ES 3.2 and Vulkan and also first to integrate AFBC.
ARM Mali-G52 is ARM’s second Bifrost-based mainstream GPU. It not only offers premium visual experiences but also comes with heightened Machine Learning capabilities. It offers 30% more performance density, 15% better energy efficiency and 3 X ML Capabilities. It also supports UI complexities for next Gen Smart TV’s and bring premium content like AR, Gaming and ML to Mainstream.
ARM Mali-G31 Features
- Technology: in 28nm HPM
- Frequency: 650 MHz
- Pixel / Texturing throughput: 1.3Gpix/s
- Anti-Aliasing: 4x MSAA, 8x MSAA, 16x MSAA
- API Support: OpenGL ES1.1/2.0/3.1/3.2, Vulkan 1.0*, OpenCL 1.1/1.2/2.0 Full Profile, RenderScript
- Bus Interface: AMBA 4, ACE-LITE
- L2 Cache: Configurable 32kB – 512kB
- Memory System: Virtual memory
- Multi-Core Scaling: 1 uni-pixel core or dual pixel core
- Adaptive Scalable Texture Compression (ASTC): Low Dynamic Range (LDR) and High Dynamic Range (HDR), Supports both 2D / 3D images
- ARM Frame Buffer Compress (AFBC): Version 1.2, 4×4 Pixel Block size
- Transaction Elimination: 16 x 16 pixel block size
- Smart Composition: 16 x 16 Pixel block size
ARM Mali-G52 Features
- Technology: in 16nm
- Frequency: 850 MHz
- Pixel / Texturing throughput: 6.8Gpix/s
- Anti-Aliasing: 4x MSAA, 8x MSAA, 16x MSAA
- API Support: OpenGL ES1.1/2.0/3.1/3.2, Vulkan 1.0*, OpenCL 1.1/1.2/2.0 Full Profile, RenderScript
- Bus Interface: AMBA 4, ACE-LITE
- L2 Cache: Configurable 32kB – 512kB
- Memory System: Virtual memory
- Multi-Core Scaling: 1 to 4 dual-pixel cores
- Adaptive Scalable Texture Compression (ASTC): Low Dynamic Range (LDR) and High Dynamic Range (HDR), Supports both 2D / 3D images
- ARM Frame Buffer Compress (AFBC): Version 1.2, 4×4 Pixel Block size
- Transaction Elimination: 16 x 16 pixel block size
- Smart Composition: 16 x 16 Pixel block size