Amd fenghuang: new higher lcm chip than kaby lake

AMD is working on an MCM chip called Fenghuang, which was once seen in the SiSoftware Sandra database last year, but since then, no other information had come out about this chip until today.

AMD Fenghuang - Faster than Kaby Lake-G

The MCM chip is similar to the Kaby Lake-G that Intel and AMD jointly developed (Intel CPU + GPU Radeon), but this one will apparently be more advanced. The AMD Fenghuang 15FF is an internal code name for graphics, while the chip it was tested with is currently called DG02SRTBP4MFA. Different code names have been detected since the chip was first leaked, but it should be clear that this is largely an early engineering sample.

As we can see in the 3DMark results

Looking at the specifications reported in 3DMark, the processor itself appears to be a Zen + based 4 core 8 core APU piece. The chip is registering base frequencies of 3.00 GHz while the frequency in Turbo remains a mystery. The GPU has 2 GB of HBM2 memory, which registers a speed of 1200 MHz (2.4 GHz effective). This may be the fastest speed for an AMD implemented HBM2 memory. The core clock for the chip is clocked at 300 MHz, which must be the frequency when it is not working, remember that 3DMark is not too accurate for reporting clocks on engineering samples.

We know that the GPU has about 28 CUs and 1, 792 SPs. AMD supplied the Kaby Lake-G chips with RX Vega GPUs with 24 CUs and 1, 536 SPs and 4 GB of HBM2 at 945 MHz. With this data, it is normal for AMD Fenghuang to have smashed the Intel chip. What's more, the graphics performance of this MCM chip would be comparable to that of a GTX 1060.

We hope to see this chip at Computex.

Wccftech font