About the APE project	Schifano S. Fabio - INFN

• APE: Array Processor Elements
• born at INFN in the middle of 1980

The need for high performance computing convinced a bunch of physicists to build at home dedicated machine optimized for QCD computing.

The APE project had several generation:

1. APE (1985 - 1990):
   • 16 nodes
   • 1 Gflops peak performance
   • 1 + 1/2 machine built

2. APE100 (1992 - 1997):
   • 128 - 1024 nodes
   • 6 - 50 Gflops peak performance
   • O(10) machines built and still running

3. APE1000 (1996 - 2000)
   • 128 - 1024 nodes
   • 60 - 500 Gflops peak performance
   • integrated 1,4 Tflops under construction

4. apeNEXT ... under definition