That is what I got from those links myself. Super Pi is programmed to make use of the x87 math co-processor instruction set and both companies need to implement it anyway for the sake of compatibility. If Intel has kept its x87 instruction execution efficient through its architecture changes, why is Super Pi being blamed for being biased. The guy is himself claiming that the application is very old, In fact its from an era when AMD was making licensed clones of Intel 386/486 CPU's. Nobody stopped AMD from implementing the instruction set just as efficiently in its future CPU's. I understand the thread authors desire for benchmarks that test SSE and newer instruction set performance, but as long as its running the same instructions on both CPU's I don't see any thing that negates Super Pi as a fair benchmark for that particular area of the CPU.thebanik said:You clearly didnt follow the discussion then,The difference is that its not as if the benchmark is optimised for Intel (or Intel has optimised their CPU for the benchmarks), but Superpi runs better on Intel because of in simple words Intel architecture is better at running the instruction set used in Superpi.
Take a hypothetical case for example. IF Nvidia cards have more/faster shaders than ATI cards they will run Shader heavy games/scenes better than ATI. In this case its not that the game is optimised for Nvidia but the game relies on a particular feature which is better in Nvidia.
I hope you understand the difference now, (@singh), and why I called the guy a fanboi and though a logical person would get the gist out of those threads but beginners may not understand and take a different picture home.
