Home > News content

Measure the area of AMD sharp dragon 4000: 150 square mm into eight cores and eight units

via:驱动之家     time:2020/1/14 23:30:44     readed:110

Amd recently officially releasedReel 4000U, Reel 4000H series mobile APU (code name Renoir) It adopts 7Nm technology, zen2 architecture, up to eight CPU cores, eight GPU computing units (512sp), and thermal design power consumption is 15 / 45W.

After the launch, AnandTech visited AMD's stand and was surprised to find a naked piece of the Reelong 4000 APU, along with a naked piece of the 64 core 128 thread tearer 3990X, so they measured their core area.


The three-generation Reejon and three-generation rippers were designed using a chiplet chip with eight cores per CPU Die and a measured area of 10.32

The measured length and width of the core of Ruilong 4000 APU are 13.59mm, 10.98mm square and 149.27mm square.

Under such an area, combined with the defect index of 0.09/cm2 of TSMC 7Nm process,The yield rate of the sharp dragon 4000 APU is about 90%Although this does not consider the manufacturing yield, it is quite considerable.

In addition, before the release of the sharp dragon 4000 APU, there was speculation about how amd would design, whether four CPU cores plus more GPU cores, or eight CPU cores plus less GPU cores.

As a result, Amd not only plugs in eight CPU cores and 15W thermal design power consumption to accelerate the frequency to 4.3ghz, but also puts in eight GPU computing units and claims that the performance is 56% higher than that of the previous generation, which is very delicate and not easy.



By the way, the area of Ruilong 3000 APU manufactured by GF 12NM process is 209.78 square mm, with four CPU cores and 11 GPU units.

The 10-nm process has an Intel Tiger Lake area of 146.10 square millimeters with four CPU cores and 96 GPU units (of course, AMD/InterGPU units are not directly comparable).


China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments