[fChk]

Hello there!.. @rudra

Sorry for the late comment. I did saw this review back when it was posted, but given its richness information-wise, I just postponed this post to when I have other data to compare it with.

I noticed that most of your benchmarks were directed to testing the performance of your disk I/O, which is understandable, given that's where bottlenecks always are. The data you came up with also served me to try a comparison between VPS4 HDD, VPS16 SSD, @'Hidden Refuge' 's NanoKVM box's NVMe, my NanoKVM box SSD and my VPS9 SSD.

The idea was to make sense of how reliable those bench scores are in comparing disk I/O performance even in the limited way of sequential read/write--given that's all we have published here.

So, to complete the picture, I'm still missing few data-point from you :

> I've noticed that, in your dd-based disk seq. write test, you didn't use the standard values (ie the values used in Hidden's bench script) and the fdatasync flag. Running instead the following command:

Code: (Select All)

sync; dd if=/dev/zero of=tempfile bs=1M count=1024; sync

Which isn't the same as running:

Code: (Select All)

dd if=/dev/zero of=tempfile bs=1M count=1024 conv=fdatasync;

So, can we please have the output of this command run 3 times (mimicking HR's I/O script.)

Code: (Select All)

dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync && rm -f test

When comparing data, we need to stick to the same script for the comparison to make any sense. Granted, that the test doesn't have much weight in itself, but it's just a data-point for comparison purposes.

> Also, I'm curious about your I/O controllers: storage and network. I'm expecting it to be of the Virtio type. So please give us a feedback on those 2 commands:

Code: (Select All)

lshw -class disk

Code: (Select All)

lshw -class network

Apart from the above, your review is EPIC, on par with Hidden's reviews of VPS-16 and his NanoBox. Kudos for both of you!

And +1Rep for this review.