RK3328 is the running king, game king, ultra clear king, RK3288 set a variety of advantages in one, is the world's first ARM new architecture core chip, or the world's first new Mali-T76x series GPU chip, and the world's first 4Kx2K hard chip for H.265 video.
OTT TV BOX - RK3328
- Quad-core Cortex-A53 up to 1.5GHz
- Mali-450MP2 GPU
- 4K UHD H265/H264/VP9
- H265/H264 encoder
- TS in/CSA 2.0
- HDMI 2.0a with HDCP 2.2
- FE PHY/Audio DAC/CVBS/RGMII
||• Quad-Core Cortex-A53,up to 1.5GHz
||• Mali-450MP2, Support OpenGL ES1.1/2.0
||• 32bit DDR3-1866/DDR3L-1866/LPDDR3-1866/DDR4-2133
|• Support eMMC 4.51,SDCard, SPI Flash
||• 4K VP9 and 4K 10bits H265/H264 video decode, up to 60fps
|• 1080P other video decoders (VC-1, MPEG-1/2/4, VP8)
|• 1080P video encoder for H.264 and H.265
|• Video post processor: de-interlace, de-noise, enhancement for edge/detail/color
|• Support HDR10 ,HLG HDR , Support conversion between SDR and HDR
||• HDMI 2.0a for 4K@60Hz with HDCP 1.4/2.2
|• Support conversion between Rec.2020 and Rec.709
||• ARM TrustZone (TEE), Secure Video Path, Cipher Engine, Secure boot
|• 8 channels I2S/PDM interface, supports 8 channels Mic array
|• Embedded CVBS、HDMI、Ethernet MAC and PHY、S/PDIF、Audio DAC
|• TS in/CSA2.0 ,support DTV function
||• BGA316 14X14, 0.65mm pitch
||• MP Now
Android OTT tv box RK3328 CPU features：
1: Faster Booting in 10 Seconds
2: Fast Loading Apps Without Delay
3: Multi-Window Function
4: Webpage Speedy Loading
5: 2560x1600 Screen @60fps Supported
6: Perfect Play 4Kx2K H.265 Video
8: Security Systems Supported
9: Dolby Sound Effect Supported
10: 13M pixel Camera Supported
OTT tv box RK3328 Technical features:
1: ARM Frame Buffer Compression Technique
2: Adaptive Scalable Texture Compression Technique
3: Transaction Elimination Technique
4: Stereoscopic 3D Game Driver
5: POP-Star Searching Entertainment Engine.
OTT TV box rk3328 Detailed introduction:
Run King: The world's first Cortex-A17 chip
The Rokchip RK3288 processor uses the Cortex-A17. Its biggest feature is the complete out-of-order execution architecture like the Cortex-A15. Its performance improvement has an immediate effect. The special architecture design of dual instruction decoding, dual address generator (AGU), multi-issue (multi issue) makes the Cortex-A17 have a considerable improvement in instruction execution efficiency. It is based on the above improvements, in the same At the frequency, the single-thread performance of the Cortex-A17 is 40% higher than that of the Cortex-A9, while the RK3288 with a frequency of up to 1.8GHz has a 50% performance improvement over the previous generation RK3188.
Game King: Mali-T764 series GPU chip
In the previous RK3188 processor, the GPU Mali-400 used by Rokchip has been the weakness of the RK3188. In the latest RK3288 processor, the GPU uses the Mali T76X series GPU. The biggest feature of the Mali T764 core is the third-generation Midgard architecture. Compared to the second-generation MIDgard architecture used in the T62x/65x series GPUs, the third-generation Midgard. The architecture has been redesigned on the internal architecture, specifically changing the way the shader core is configured; the number of shaders is up to 4 shader cores, twice the previous T654; 2400M/s pixel fill rate and 300M The triangle generation rate of /s has been doubled compared to the past.
Ultra Qing Wang: Support 4Kx2K real-time hard solution H.265, HDMI2.0
The Rokchip RK3288 is a chip that supports 4Kx2K and hard-release H.265 video, giving users clear visual enjoyment as a more efficient HD compression format. The H.265 standard has become a lifeline for video sites. Compared with H.264, H.265 uses a size-floating coding unit and a more advanced intra prediction mode. With the same picture quality, H.265 video can save half the bandwidth compared to H.264. Rokchip RK3288 is the industry's first ARM core chip to support 4K H.265 real-time hardware decoding. It has accumulated its accumulated advantages in the audio and video field in video support, and truly supports the output and playback of 4K video.
RK3288 latest GPU: Mali -T764 detailed
The third generation midgard architecture, more firepower
The biggest feature of the mali-t764 is the use of the third-generation midgard architecture, which has been redesigned on the internal architecture, dramatically changing the way the shader core is configured. The memory synchronization between the shaders and the scu between the core groups are still maintained. However, independent l2 caches and snoop units are still used between the core groups, and the connection between them is realized through the amba4 bus. Therefore, it can be considered that the task assignment can be optimized as needed.
At the same time, Mali-t764 also supports gpgpu (Universal Computing Graphics Processor) to accelerate complex and computationally intensive algorithms or operations. The rk3288 will also shine on wearable devices. The rk3288 can identify moving objects on the camera screen and render it in real time with efficient performance, resulting in a better experience. Applications such as image stitching recognition, face recognition, smile recognition, landmark recognition and even frown recognition can be easily implemented. .
Mali-t764 further enhanced support for different APIs, and officially added support for OpenGL es3.0, opencl 1.1, directx 11.1 and renderscript. With these new apis, the integrated mali-t764gpu processor can display more detailed graphics, and has the opportunity to run on the traditional PC platform, making it easier to migrate PC software and mobile software.
ARM frame buffer compression, low memory bandwidth for high quality
Supporting the ARM framebuffer compression format is also a highlight of the Mali-T764, which allows it to ease the pressure of insufficient memory bandwidth when running high-definition games. Many tablets that use the Mali-400MP4+ retina screen will appear very stuck in the highest resolution mode, and to a large extent is the manifestation of insufficient memory bandwidth. To address the memory bandwidth challenges of video, ARM developed the ARM framebuffer compression format, which provides fast, real-time lossless compression and decompression. This reduces the memory bandwidth requirements of the entire system and reduces the corresponding power consumption by up to 50%.
In addition, the Mali-T764 also supports the ASTC texture compression format when it supports OpenGL ES 3.0. In addition to reducing the transmission size of the texture material, this also allows the high-quality games of the handheld device to completely reduce the bandwidth consumption on the texture material and maintain the game. Picture quality can also reduce bandwidth. Rokchip RK3288 and Apple A7 performance comparison.
Apple's new A7 processor uses the PowerVR 6 core architecture. The PowerVR 6 architecture is a unified superscalar rendering engine, shader array design, with four shaders designed for a SIMD array. The A7 uses the IMG PowerVR G6430 with four SIMD array designs.
The Mali-T764 support features used in the Rokchip RK3288 are not inferior to the A7's PowerVR G6430. The Mali-T764 uses ARM's latest third-generation MIDgard architecture, which is also a unified rendering architecture using the same GPU on the current PC. It features improved graphics processing while maintaining sufficient graphics processing efficiency, including 64-bit. Double-precision floating-point operations and 64-bit integer operations, 64-bit addressing, memory synchronization between shader units. However, compared to the SIMD architecture of the PowerVR 6430 core using the decomposition vector engine and the scalar engine, the Mali-T764's computational pipeline is designed with a 128-bit vector engine and a flexible pipeline configuration. In particular, as the highest model of the Mali-T700 series, the Mali-T764 has a single scheduling mechanism, and each core group is divided into the same processing tasks, and performance optimization can be achieved as needed. It can be said that if the two are evenly matched from the hard configuration, the Mali-T764 is superior in the design of the computing pipeline.
Graphics processing performance, Mali-T764 and G6230 roughly equal
The performance of the A7's PowerVR G6430 is similar to that of the A6X processor PowerVR 554MP4: the same-frequency floating-point performance is the same as the SGX554MP4 on the iPad4, and the triangle fill rate and texture performance are twice that of the SGX554MP4.
The Mali-T764 is equipped with four ALUs (arithmetic logic units), one LSU (local storage unit), and one texture unit per shader. Moreover, the four shader groups of Mali-T764 have independent core groups sharing L2 cache, and the four core groups pass through a single dispatcher, and each core group is divided into the same processing task. More importantly, each shader has independent thread scheduling, which can allocate program instructions from each thread to their respective pipelines, so that each pipeline can be flexibly used to execute instructions from their respective threads, so if there are 128 pipelines, then It can run 128 threads at the same time, and the Mali-T764's shader unit can control more than 800 threads. This flexible thread scheduling method makes the Mali-T764's shader unit more efficient and even has the same graphics performance as a desktop PC.
The Mali-T760 has a peak performance of 326.4 GFLOPS, a 1066M triangle generation rate and a 9600M pixel fill rate with 16 cores at 600MHz, which is better than the A7's PowerVR G6430. Since the Mali-T764 is only equipped with 4 cores, the performance is only one quarter of the Mali-T760mp16. In addition, the Mali series calculates the floating point as FP16, while other manufacturers' GPUs are calculated under FP32, so the Mali-T760 floating point performance. In fact, only the floating point performance of about half of the publicity value, the floating point of the Mali T764FP32 is 81.6 GFLOPS, the actual floating point of the FP16 is only 40.8 GFLOPS, and the G6230 is 76.8GFLOPS of the FP16 at 600mhz, so the performance of the T764 is not comparable to the G6230, but the difference is not Too big.