Machine w/ parallel/vector processor

http://www.adapteva.com/announcements/epiphany-v-a-1024-core-64-bit-risc-processor/
Actual vector processing seems dead unless you want to call wide SIMD devices like KNL or GPUs vector processors. Maybe the Epiphany - or if poorfag the Parallela - are for you?

>>58587254
Use a CPU if possible, if not use a GPU, if not use FPGA, if not use ASIC.

For general purpose applications? Use a CPU. Most cores are fast enough. Heck our CPUs are fast enough for most gaymes to run at 60 fps or higher on a single core.

Otherwise I would suggest you to just use a lot of smaller downclocked cores but that requires more competent programmers.

Raspberry Pi cluster.

File: r32V794.gif (615KB, 300x190px) Image search: [Google]

615KB, 300x190px

>>58587254
I've sat on one of those, the cushions are comfy

File: Linköping_University_CRAY_X-MP.jpg (101KB, 1024x640px) Image search: [Google]

101KB, 1024x640px

>>58590470
me too desu

>>58587254
>How
Use loads of bit slice chips. Many great computers were made that way.

Or learn VHDL or Verilog and make your own in FPGA. Go to Opencores and build on freely available processor designs you can find there.

>>58587254
Just use a GPU???!? or a DSC maybe