Cuda learning from the CPU architecture

Recently to learn GPU programming, go to the NVIDIA network download Cuda, the first problem encountered is the choice of architecture

So the first step I learned was to learn about the CPU architecture, x86-64 abbreviated x64, a 64-bit version of the x86 instruction set, forward-compatible with the 16-bit version and the 32-bit version of the x86 architecture. x64 was originally designed by AMD in 1999, and AMD first exposes 64-bit sets to x86, called "AMD64". It was later adopted by Intel and was referred to by Intel as "Intel 64".

So what does ppc64le mean, referring to the two paragraphs on the wiki below?

PPC64 is commonly used in the Linux and GCC open source software community, which points to the target architecture of the 64-bit PowerPC and power architecture processors, and the identifiers of the applications that are optimized for this purpose.

Ppc64le is a technology that has launched a pure small-end model, POWER8 as the primary goal of the OpenPOWER Foundation, trying to make the porting of x86-based Linux software work with minimal effort. ”

Cuda learning from the CPU architecture