doubt about TSUBAME supercomputer's peak performance

Hi, all

Tokyo Tech said their TSUBAME supercomputer had gotten a theoretical peak performance of nearly 170 Tflops by adding 170 Tesla S1070 systems, but there should be 680 Tflops of theoretical peak performance since 1 Tesla S1070 delivers 4 Tflops of performance! What wrong with these number? Can anyone give a reasonable explanation?

Gimurk

Linpack is a double precision benchmark and the peak double precision performance of the 100 series GPUs is one eighth of the headline single precision performance most people like to quote (each multiprocessor has eight single precision stream processors but only one double precision stream processor).

And before you ask the next obvious question, read this: http://doi.acm.org/10.1145/1513895.1513901

Thank you for your answer, and let me have a look :)

is this document available somewhere else on the internet for free? I tried to get it but they ask me for a payment.

I have downloaded it

give me your mail if you want to get it for free ^_^