PINE64

Full Version: Benchmarking the SOPINE Modules on the Clusterboard
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I did run the NPB3.4.1-MZ Benchmark on my cluster of 7 compute modules. Here are the results for SP-MZ Class=B:


[Image: attachment.php?aid=1997][Image: attachment.php?aid=1996]

[Image: attachment.php?aid=1995]
[Image: attachment.php?aid=1994]
For comparison:



My Intel Core-i5 5300U needs 72 seconds on two cores to complete the benchmark.






I plan on doing some more benchmarks, let me know what you think!
So with 20+ threads the runtime was sub 100s? A bit hard to tell due to the scale of the y-axis.

Thanks for making these benchmarks Smile
surprised at the diminishing returns of >3 modules...
That's actually cores/threads. So >1 module, but I suspect something else is at play there...
i was assuming each module is 4 threads, so >3 modules or >12 threads