We’ve talked a lot about this GPU and its dual-matrix variant, which could even compete with a full RTX 3060 (in theory). What is leaked today leaves so much unknown or perhaps more than just before it was known of the existence of the video revealed, since the Fenghua Type-A does not appear to be what was promised.
GFXBench is the first test for this GPU
Recall that there are two aspects which are nothing more than a single-core GPU and a multi-die, where both include the same advantages per PowerVR chip and architecture. Type A will have 5 TFLOPS on FP32 and should match RTX 2060, while version Type-B
Both support PCIe 4.0 and really incredible power consumption when you take into account that they are made for 12 nm, since it is ensured that version A would only consume 50 watts, while the double dice would reach the 100 watts only, so if performance materializes, we will talk about the best GPUs in terms of energy efficiency per FPS.
Moreover, today we know that although they will arrive in several variants of VRAM (4 GB, 8 GB and 16 GB respectively with the 32 GB version for the dual die) they will do so with GDDR6X at 19 Gbps with a bandwidth of 304 GB / s, oddly 10% below what an RTX 2060 achieves, which is strange given that current GPUs are heavily dependent on this parameter
Fenghua 1 and reference
Innosilicon graphics card. pic.twitter.com/lyMpQAo5Km
– Stewie (@Stewrandall) 23 December 2021
What is seen in this video only leaves more doubts about the performance, although it is also true that there could be several explanations that could somehow justify what was seen. First of all, it should be said that the benchmark does not have an FPS counter and there is no software running behind that overlaps a value.
The only thing on the screen is a feeling of fluidity that leaves a lot to be desired. Keep in mind that this benchmark is designed for mobile and that a GPU like Type-A (version used for testing) should achieve higher values 300 FPS
The feeling we get from watching the video is that the smoothness is not what it should for such a high framerate, but here more doubts are raised. Is it really a monitor problem? Is this an effect created by the difference in FPS that the camera can capture? It is not known and it is not at all clear because the final or partial data or values are not offered.
Until then, we won’t be able to place this new GPU on the world stage. What we can say is that many companies are already receiving engineering samples to assess not only the performance but also the potential of these GPUs, so it is a palpable reality that before it was simply data on paper.