what are differences between - Andreno , Power VR and Mali GPUs?
Answers
Answer:
There's no way to accurately answer this with just public information, and it's compounded by the fact that one of the embedded GPUs in your list isn't available as licensable IP (Qualcomm Adreno). That means that even those privy to the detailed attributes of the other GPUs will have a hard time accurately comparing them to Adreno.
The private measures take on a number of forms, some based on the public peak performance attributes of the IP (pixel and texturing fillrates, setup rate, instruction throughputs), but many based on information that's never published and rarely measured in public (area and power in particular).
The reason it's impossible to accurately answer in public, and still very hard to do in private even if you're in the market for the IP to integrate, is that much of the final data depends on the choices made at the physical design level.
Your foundry partner, tooling, prowess and resources at back-end physical design, cell libraries, wafer cost, yield profile, available package type and how much money and time you have to combine it all to develop your chip, never mind the rest of the IP in the mix, all have an influence on the final, power, performance and area of the GPU.
I have a great view of those things (it's my job after all, I lead competitive analysis for PowerVR), but there's no way to publicly answer accurately. All I can say is that you can engineer SoC designs and given cost models where all licensable GPU IPs will look better than their immediate peers.
i think he is our new hommie
