# Description

Implementation of the [STREAM benchmark](https://www.cs.virginia.edu/stream/ref.html). 
STREAM-static uses static array allocation and STREAM-dynamic uses dynamic array allocation.
If running STREAM-static on a machine with limited memory, you may wish to change the 
size of the allocated arrays.

# How to use

## create a execution file
### On the K computer or FX10
 $ make compiler=K

### Use Intel Compiler
 $ make compiler=intel

### Other
 $ make

## run
 $ mpirun -np 4 ./STREAM-dynamic 1024

 The argument (1024) is a vector size.

# MEMO
## Performance on HPC Challenge Class 1
 3,857 TB/s on 82,944 nodes (46.5 GB/s/node, 72.7 % of Theoretical peak)
 http://icl.cs.utk.edu/hpcc/hpcc_record.cgi?id=492


 
 
 
