Disk Benchmarks for Compute Blade RC2
2 min read

Disk Benchmarks for Compute Blade RC2

Some SSD benchmarks for the compute blade rc2.
Disk Benchmarks for Compute Blade RC2

I have a RC2 compute blade and it's great!

As for the moment it's mostly idle, I figured it'd be a good idea to run some benchmarks.

These are my SSD benchmark, with a Crucial P5 500GB.

Methodology

I used the pibenchmarks script:

sudo curl https://raw.githubusercontent.com/TheRemote/PiBenchmarks/master/Storage.sh | sudo bash

And executed it three times (so I get an average). You can follow

Results

First run produced the following results:

Category Test Result
HDParm Disk Read 350.72 MB/s
HDParm Cached Disk Read 350.41 MB/s
DD Disk Write 117 MB/s
FIO 4k random read 40715 IOPS (162862 KB/s)
FIO 4k random write 11689 IOPS (46757 KB/s)
IOZone 4k read 77453 KB/s
IOZone 4k write 47675 KB/s
IOZone 4k random read 42182 KB/s
IOZone 4k random write 66383 KB/s

Score: 13572

Second run:

Category Test Result
HDParm Disk Read 350.15 MB/s
HDParm Cached Disk Read 346.75 MB/s
DD Disk Write 117 MB/s
FIO 4k random read 40554 IOPS (162217 KB/s)
FIO 4k random write 12089 IOPS (48358 KB/s)
IOZone 4k read 71972 KB/s
IOZone 4k write 47891 KB/s
IOZone 4k random read 40283 KB/s
IOZone 4k random write 65640 KB/s

Score: 13403

Third run was:

Category Test Result
HDParm Disk Read 350.08 MB/s
HDParm Cached Disk Read 347.87 MB/s
DD Disk Write 117 MB/s
FIO 4k random read 40878 IOPS (163512 KB/s)
FIO 4k random write 12018 IOPS (48075 KB/s)
IOZone 4k read 82516 KB/s
IOZone 4k write 47672 KB/s
IOZone 4k random read 41660 KB/s
IOZone 4k random write 67022 KB/s

Score: 13705

Conclusion

The benchmark score is around 13600. At the moment of writing this I can see on the site, the following results:

User SSD brand Score Notes
merocle Samsung 960 Pro 1G 20,600 Creator of the blade
stormtrooper Toshiba XG5 512G 17,638
geerlingguy Kioxia SSD 15,546 How I found out about the blade :)
jchambers Samsung 960 EVO 14,559 note: Pine64 Blade + SOQuartz
me Crucial P5 500G 13.600

This is about 3 times faster than eMMC :)