The IDM929 GPU employs Zhihui Micro's exclusive IDMV architecture, instruction set, and compiler. The new GPU boasts 2.5 TFLOPS of computing power, a fill rate of 76.8 GTexel/s, and a pixel fill rate of 19.2 GPixels/s.

In contrast to conventional devices produced on the 14nm CMOS process, the IDM929 GPU achieves a clock frequency of only 1.2GHz. However, Zhihui Micro remains committed to ensuring the new GPU's actual performance will be 8 times that of the original design. Additionally, with the 14nm manufacturing process, there are speculations that the new GPU will be produced by Semiconductor Manufacturing International Corporation (SMIC).

As per the information shared by the company, the IDM929 GPU will possess three key advantages: computational power, flexibility, and high energy efficiency. Moreover, the product supports up to 4 independent displays with a resolution of up to 4K through HDMI, VGA, and DVI protocols. It also integrates hardware encoding standards in H264, MPEG2, MPEG4, VC-1, DivX, and VP6 formats.
The company has completed internal tests on the new GPU when combined with domestic CPUs and operating systems. The product is expected to find widespread applications in tasks such as office computers, high-performance workstations, geographical information systems, and multitasking machines.
According to press releases, Hangzhou Guoxin Technology, the company manufacturing SoCs for smart TVs, is now emerging as the primary investment brand for Zhihui Microelectronics. Leaked specifications indicate that the IDM929 GPU will deliver performance equivalent to the GTX 1650, featuring 3.0 TFLOPS of computational power and produced on the 12nm process from TSMC.

The IDM929 GPU is set to be the successor to the IDM919 GPU series introduced by the company in 2020. Additionally, Zhihui Microelectronics includes the IDM939 GPU in the list of developing products, with no other related information currently available. It's intriguing to witness China's self-produced GPU series, and there's hopeful anticipation for further innovations in the next generation of GPUs.
