All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Why? VPS I/O [Count 16k=1.1 GB/s] [Count 512=6.2 MB/s]
I have two VPS. Tested the VPS I/O as below.
(1) Which below VPS is the best?
(2) Why VPS 2 [Count 16k=1.1 GB/s] [Count 512=6.2 MB/s]? 6.2MB/s which my USB Flash Drive may faster than it.
=============
VPS 1
=============
(1) Count=16k
cd /
rm -rf test
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
1073741824 bytes (1.1 GB) copied, 1.97455 s, 544 MB/s
1073741824 bytes (1.1 GB) copied, 1.81736 s, 591 MB/s
1073741824 bytes (1.1 GB) copied, 1.70579 s, 629 MB/s
(2) Count=512
cd /
rm -rf test
dd if=/dev/zero of=test bs=64k count=512 oflag=dsync
33554432 bytes (34 MB) copied, 0.806547 s, 41.6 MB/s
33554432 bytes (34 MB) copied, 1.02883 s, 32.6 MB/s
33554432 bytes (34 MB) copied, 0.797054 s, 42.1 MB/s
=============
VPS 2
=============
(1) Count=16k
cd /
rm -rf test
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
1073741824 bytes (1.1 GB) copied, 0.95544 s, 1.1 GB/s
1073741824 bytes (1.1 GB) copied, 0.936963 s, 1.1 GB/s
1073741824 bytes (1.1 GB) copied, 0.955365 s, 1.1 GB/s
(2) Count=512
cd /
rm -rf test
dd if=/dev/zero of=test bs=64k count=512 oflag=dsync
33554432 bytes (34 MB) copied, 5.4428 s, 6.2 MB/s
33554432 bytes (34 MB) copied, 5.58766 s, 6.0 MB/s
33554432 bytes (34 MB) copied, 5.37726 s, 6.2 MB/s
Comments
No.
what? I do not understand the question you are asking.
The main difference here is not even the count, but conv=fdatasync vs oflag=dsync. Do not use the latter, it will give you an unrealistic extremely low result. See my page for more details.
My guess would be that VPS 2 likely has a SAN or similar network-based storage behind it; giving it increased latency on individual requests (which "dsync" triggers) while providing more throughput overall.
In general though, dd (or any throughput-based result) is a poor benchmark because most people are not going to be writing out gigabytes files very often. Use FIO and measure the iops.
DD doesn't exactly show whats best really, it all depends on real word usage.
Example: One provider could have 200MB/s DD test with great uptime, but the second provider could have a 1GB/s DD test but have bad uptime. Which would be better? I'd go with the provider with the 200MB/s DD test, not the second host with horrible uptime.
My 2 cents.
Thanks all !! One more question.
VPS1= 2 core and VPS2 = 4 core which affect the I/O performance or not?
No, not really, while iowait will affect CPU, the CPU will not really affect iowait. That being said, it might be slightly better to have more cores so when some are bogged down in waiting states, others are doing something, but that is only in theory because usually iowait means the data to process is not available, so all cores will probably wait.
In short, not really.
You honestly look like a knowldeable guy rather than one of those DD porn addicts around here.
DD is a very good test to use for most stuff, for example I can tell which VPS is best for writing my 1.1gb sequential files
Are you trolling?
No, it's a good test
So you are trolling.