96MB Low End VPS Review Part XXXIII – SSD Nodes
Matt from SSD Nodes has dropped me a message on Webhostingtalk over a month ago asking for a review on their product and although he has encouraged me to try out a VPS using SSD on the root partition, I have decided to keep up with the spirit of choosing selecting a standard LEB offer where possible and ended up choosing the small plan that is almost the same as what is advertised on LEB, with a slight modification (but at the same cost), and we’ll take a look at how this box performs.
General Information and Set Up
As per the advertisement on LEB, here is what you will be getting for 6.99 per month:
However, while reading through the comments section, I have noticed that Matt has posted a comment that we can use SSD as a secondary storage at no additional cost, and that is what I have decided to give it a shot.
The word “cloud” and “SSD” are definitely some of the key words on their webpage:
Once the START button is clicked, the default price and configuration are displayed on the right hand side of the order page:
There are quite a few configurable options available on WHMCS, starting with the management option (which I think most of the LEB fans would not even want that :)). You can then add more RAMS to the package if desired. The RAM used here is pretty high end, which, of course, comes with a pretty high price as well:
There are very limited options available for the initial OS installations, however I was being told that all templates are built from scratch and they can pretty much install any Linux distributions. Although, I have to say I was pretty amazed when I did not see CentOS distributions available, how could CPanel be installed in that case?
There are two options available for primary storage, either 8GB of SATA II (default) or 8GB of SATA III SSD (at a much higher cost though):
As per Matt’s comment on LEB, they have recently made 1GB SSD free as a secondary storage device, if you are willing to give up 40GB of SATA II hard drive for secondary storage. For each GB of storage, it cost approximately 2.50USD. However, considering it is a SATA III SSD, it could be worthwhile:
There is only 100GB of bandwidth included, which I have to say is not exactly that generous. Extra bandwidth cost 10USD per 100GB:
The highlight in the pricing is really their cPanel license, which cost only 9.99USD, a lot lower than the “standard” CPanel for VPS would cost (around 15USD per month):
There were actually some payment issue that I have had while trying to pay for the invoice, for some reason the invoice has only Subscribe option available (compared to both One Time Payment and Subscribe options that you will normally see in the invoices) and even when I tried to subscribe to the payment plan, the payment would not go through for some reason, instead, I ended up seeing an error message that says: You must add funds to your PayPal account before sending more money. I had to raise a ticket to SSD nodes for the payment issue, which was responded within 5 minutes (ticket raised on a Friday at 12:49PM, responded at 12:54PM).
I have received my invoice payment confirmation email at the same time with my VPS information. However I do not think it is instant provision. The welcome email is pretty unique:
As you can see, there is no link to the control panels but instead a link to SSL VPN, which made me wondering for a while why that is the case before I hit the last few lines (there is also very good documentation in their knowledgebase as well).
Basically, the SSH in VPS, by default, only listens to the internal address. Therefore, you can not access the VPS from external network unless you have the VPN connection first. Although I can definitely see it as a security benefit, for some reason my VPN connection using Softlayer basically gets disconnected once every 2 minutes if I am running it from Firefox (although it runs fine when I am using Internet Explorer), which did gave me a few moments of headaches. SSD Nodes has also mentioned you can get this setting changed by submitting a ticket, which is great.
There is currently no control panel available for them and if you log in to the WHMCS system, you do not see anything that can be used to control the VPS:
Therefore, you will need to send in a ticket for everything from changing root password to reload OS templates. From what James in SSD Nodes have told me, the infrastructure is “highly customized” and they will develop an entirely in-house control panel, which Matt has promised me it would be a top priority. However, until then, everything is handled via tickets and James believes it also added the “human” element of the services.
Test on the VPS
As mentioned above, the VPS that I have received for testing has 128MB of RAM, 8GB of SATA II hard drive as primary storage and 1GB of SATA III as secondary storage. The VPS is in Softlayer’s data centre in Dallas, Texas, and initially I have ordered Debian 6 64 bit, and actually submitted a ticket later on to get Debian 6 32 bit loaded into the VPS so that I can use the free version of GeekBench. To my pleasant surprise, instead of bringing my VPS offline for hours while they install the OS, he has actually offered to install the OS on a brand new VPS and then reassign my IP address so there won’t be any down time for the VPS. Although I do not actually need that kind of availability, I truly appreciate the thoughts that they have put in.
Initially, the 64 bit OS was installed:
uname -a Linux ****** 2.6.32-5-amd64 #1 SMP Wed May 18 23:13:22 UTC 2011 x86_64 GNU/Linux
With this OS installed, about 26MB of RAM was used after the fresh installation:
free -m
total used free shared buffers cached
Mem: 119 94 24 0 3 63
-/+ buffers/cache: 26 92
Swap: 382 0 382
The top output showing the running processes:
top - 09:47:50 up 8:28, 4 users, load average: 0.01, 0.01, 0.00
Tasks: 57 total, 1 running, 56 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 121868k total, 97100k used, 24768k free, 4056k buffers
Swap: 392184k total, 324k used, 391860k free, 65320k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20387 root 20 0 20564 3328 1504 S 0.0 2.7 0:00.10 bash
19970 root 20 0 70580 3324 2520 S 0.0 2.7 0:00.03 sshd
20385 root 20 0 70488 3236 2532 S 0.0 2.7 0:00.04 sshd
20112 root 20 0 70488 3228 2520 S 0.0 2.6 0:00.02 sshd
20207 root 20 0 70488 3224 2520 S 0.0 2.6 0:00.02 sshd
20210 root 20 0 20492 3164 1412 S 0.0 2.6 0:00.10 bash
20115 root 20 0 20492 3152 1404 S 0.0 2.6 0:00.09 bash
19973 root 20 0 20492 3148 1400 S 0.0 2.6 0:00.10 bash
20844 root 20 0 19032 1284 1000 R 0.3 1.1 0:00.02 top
915 root 20 0 9096 1268 1052 S 0.0 1.0 0:00.29 xe-daemon
600 root 20 0 54156 1216 744 S 0.0 1.0 0:00.01 rsyslogd
20347 root 20 0 49168 1140 584 S 0.0 0.9 0:00.00 sshd
994 Debian-e 20 0 44144 1040 656 S 0.0 0.9 0:00.00 exim4
649 root 20 0 22392 860 652 S 0.0 0.7 0:00.02 cron
498 statd 20 0 14376 736 588 S 0.0 0.6 0:00.00 rpc.statd
1 root 20 0 8352 684 628 S 0.0 0.6 0:00.29 init
150 root 16 -4 16736 648 340 S 0.0 0.5 0:00.02 udevd
And close to 700MB of hard drive space was used:
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda1 7.6G 689M 6.5G 10% / tmpfs 60M 0 60M 0% /lib/init/rw udev 47M 76K 47M 1% /dev tmpfs 60M 0 60M 0% /dev/shm /dev/xvdb 1008M 34M 924M 4% /disk1
After the LNMP stack was installed, 44MB of RAM was used:
free -m
total used free shared buffers cached
Mem: 119 115 3 0 4 66
-/+ buffers/cache: 44 74
Swap: 382 14 368
And again, the corresponding top output:
top - 10:19:17 up 9:00, 3 users, load average: 0.02, 0.49, 0.60
Tasks: 65 total, 1 running, 64 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni, 89.9%id, 10.1%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 121868k total, 119220k used, 2648k free, 3964k buffers
Swap: 392184k total, 14940k used, 377244k free, 68688k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
26482 www 20 0 47020 19m 436 S 0.0 16.7 0:00.01 nginx
26470 root 20 0 106m 5608 1308 S 0.0 4.6 0:00.03 php-cgi
26472 www 20 0 106m 5168 868 S 0.0 4.2 0:00.00 php-cgi
26473 www 20 0 106m 5168 868 S 0.0 4.2 0:00.00 php-cgi
26474 www 20 0 106m 5168 868 S 0.0 4.2 0:00.00 php-cgi
26476 www 20 0 106m 5168 868 S 0.0 4.2 0:00.00 php-cgi
26477 www 20 0 106m 5168 868 S 0.0 4.2 0:00.00 php-cgi
26662 root 20 0 19040 1292 1000 R 0.0 1.1 0:00.00 top
26480 root 20 0 27808 948 268 S 0.0 0.8 0:00.00 nginx
20387 root 20 0 20564 928 576 S 0.0 0.8 0:00.11 bash
600 root 20 0 54156 660 424 S 0.0 0.5 0:00.01 rsyslogd
915 root 20 0 9096 528 388 S 0.0 0.4 0:00.31 xe-daemon
26658 root 20 0 3868 504 420 S 0.0 0.4 0:00.00 sleep
20385 root 20 0 70488 336 204 S 0.0 0.3 0:00.63 sshd
649 root 20 0 22392 180 116 S 0.0 0.1 0:00.03 cron
994 Debian-e 20 0 44144 124 60 S 0.0 0.1 0:00.00 exim4
1 root 20 0 8352 88 56 S 0.0 0.1 0:00.32 init
As well as the htop output for the htop fans:
And 2.1GB of hard drive space was used:
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda1 7.6G 2.1G 5.1G 30% / tmpfs 60M 0 60M 0% /lib/init/rw udev 47M 76K 47M 1% /dev tmpfs 60M 0 60M 0% /dev/shm /dev/xvdb 1008M 34M 924M 4% /disk1
Note that here, you can also see the 1GB SSD space was mounted as /disk1 on the VPS.
Inodes are set to pretty low level, therefore it might not be suitable if you have too many small files:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/xvda1 499712 74122 425590 15% / tmpfs 15233 5 15228 1% /lib/init/rw udev 11889 454 11435 4% /dev tmpfs 15233 1 15232 1% /dev/shm /dev/xvdb 65536 11 65525 1% /disk1
VMStat output taken right after the LNMP stack installation shows some usage on the CPU:
vmstat procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 17932 25504 5084 49280 5 6 37 53 13 25 1 0 98 1
As well as the uptime output:
uptime 10:22:02 up 9:02, 3 users, load average: 0.01, 0.30, 0.50
However, I believe it is just caused by the high CPU usage during the installation.
There is one single CPU core available and does not appear to be throttled:
cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 42 model name : Intel(R) Xeon(R) CPU E31270 @ 3.40GHz stepping : 7 cpu MHz : 3392.294 cache size : 8192 KB fpu : yes fpu_exception : yes cpuid level : 13 wp : yes flags : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc up rep_good nonstop_tsc pni pclmulqdq ssse3 cx16 sse4_1 sse4_2 x2apic popcnt avx hypervisor lahf_lm bogomips : 6784.58 clflush size : 64 cache_alignment : 64 address sizes : 36 bits physical, 48 bits virtual power management:
And the results of meminfo:
cat /proc/meminfo MemTotal: 121868 kB MemFree: 63188 kB Buffers: 3556 kB Cached: 13612 kB SwapCached: 3268 kB Active: 17520 kB Inactive: 25420 kB Active(anon): 1008 kB Inactive(anon): 24768 kB Active(file): 16512 kB Inactive(file): 652 kB Unevictable: 0 kB Mlocked: 0 kB SwapTotal: 392184 kB SwapFree: 373104 kB Dirty: 0 kB Writeback: 0 kB AnonPages: 23076 kB Mapped: 2136 kB Shmem: 4 kB Slab: 8840 kB SReclaimable: 5088 kB SUnreclaim: 3752 kB KernelStack: 536 kB PageTables: 3196 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 453116 kB Committed_AS: 130608 kB VmallocTotal: 34359738367 kB VmallocUsed: 804 kB VmallocChunk: 34359737512 kB HardwareCorrupted: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 2048 kB DirectMap4k: 131072 kB DirectMap2M: 0 kB
As well as time sync:
time sync real 0m0.024s user 0m0.004s sys 0m0.000s
Matt was really keen to ask me to test out the disk I/O on SSD, which, obviously, is a big selling point to them. So here is goes:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync dd: writing `test': No space left on device 15595+0 records in 15594+0 records out 1021976576 bytes (1.0 GB) copied, 4.94472 s, 207 MB/s
Test again showed very impressive results again as well:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync dd: writing `test': No space left on device 15558+0 records in 15557+0 records out 1019564032 bytes (1.0 GB) copied, 3.98608 s, 256 MB/s
As you can see, the SSD disk I/O easily beat most of the disk I/O speed. I have also done several iopings just to see how things goes:
ioping -c 10 /disk1/test 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=1 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=2 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=3 time=0.3 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=4 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=5 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=6 time=0.4 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=7 time=0.4 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=8 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=9 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=10 time=0.5 ms --- /disk1/test (ext4 /dev/xvdb) ioping statistics --- 10 requests completed in 9033.9 ms, 2163 iops, 8.4 mb/s min/avg/max/mdev = 0.3/0.5/0.5/0.1 ms
And of course, test again:
ioping -c 10 /disk1/test 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=1 time=0.4 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=2 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=3 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=4 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=5 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=6 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=7 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=8 time=0.4 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=9 time=0.5 ms 4096 bytes from /disk1/test (ext4 /dev/xvdb): request=10 time=0.5 ms --- /disk1/test (ext4 /dev/xvdb) ioping statistics --- 10 requests completed in 9006.0 ms, 2049 iops, 8.0 mb/s min/avg/max/mdev = 0.4/0.5/0.5/0.0 ms
However, since majority of the low end VPS users would opt for the SATA III drive simply due to cost reasons, I think it is only fair to test the disk I/O on the SATA III drives as well:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 14.939 s, 71.9 MB/s
And test again:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 18.4207 s, 58.3 MB/s
As you can see, the I/O of SATA III drives are a lot less impressive compare to the SSD drive. However, I can definitely think of some use for a hybrid box like this though. Say for example, if you have a WordPress blog, you can always put the MySQL database, which requires heavy disk I/O, on the SSD drive, and the less I/O intensive tasks, such as PHP children and Nginx workers, on the SATA III drives.
And here are the ioping results for the SATA III drives as well:
ioping -c 10 100mb.test 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=1 time=14.1 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=2 time=8.5 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=3 time=8.8 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=4 time=6.1 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=5 time=7.0 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=6 time=14.0 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=7 time=4.3 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=8 time=6.1 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=9 time=5.1 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=10 time=0.4 ms --- 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320) ioping statistics --- 10 requests completed in 9075.5 ms, 134 iops, 0.5 mb/s min/avg/max/mdev = 0.4/7.4/14.1/4.0 ms
And test again:
ioping -c 10 100mb.test 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=1 time=14.2 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=2 time=0.4 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=3 time=7.5 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=4 time=7.3 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=5 time=0.4 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=6 time=8.2 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=7 time=5.3 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=8 time=8.8 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=9 time=3.2 ms 4096 bytes from 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320): request=10 time=5.0 ms --- 100mb.test (ext3 /dev/disk/by-uuid/898f0159-0540-4ccf-95dc-97aff400b320) ioping statistics --- 10 requests completed in 9061.6 ms, 165 iops, 0.6 mb/s min/avg/max/mdev = 0.4/6.0/14.2/4.0 ms
As you can see, the ioping results are generally in line with the output of dd as well, although I do see (particularly in the second ioping test) some of the loops is a lot faster than the SSD drive, although much less stable.
The network speed is what actually puzzled me, innitially I was testing the download speed without actually have it to write to the disk, hoping to get a better results since disk I/O could take time, and the results is pretty bad:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2011-10-08 10:37:14-- http://cachefly.cachefly.net/100mb.test Resolving cachefly.cachefly.net... 205.234.175.175 Connecting to cachefly.cachefly.net|205.234.175.175|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: a€?/dev/nulla€ 100%[======================================>] 104,857,600 4.17M/s in 24s 2011-10-08 10:37:43 (4.10 MB/s) - a€?/dev/nulla€
I then waited for a bit, hoping there won’t be any caching, before running the wget test again, but this time actually writing to the disk:
wget cachefly.cachefly.net/100mb.test --2011-10-08 10:40:20-- http://cachefly.cachefly.net/100mb.test Resolving cachefly.cachefly.net... 205.234.175.175 Connecting to cachefly.cachefly.net|205.234.175.175|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: a€?100mb.testa€ 100%[======================================>] 104,857,600 40.4M/s in 2.5s 2011-10-08 10:40:28 (40.4 MB/s) - a€?100mb.testa€
As you can see, although I was expecting the wget results to be worse than what I have before, it is actually the other way round.
Test a few seconds later, again writing to the disk, actually showed me even better results:
wget cachefly.cachefly.net/100mb.test --2011-10-08 10:40:50-- http://cachefly.cachefly.net/100mb.test Resolving cachefly.cachefly.net... 205.234.175.175 Connecting to cachefly.cachefly.net|205.234.175.175|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: a€?100mb.test.1a€ 100%[======================================>] 104,857,600 60.4M/s in 1.7s 2011-10-08 10:40:57 (60.4 MB/s) - a€?100mb.test.1a€
I can certainly see the improvement from the second wget to the third one was due to caching, but why the big difference between first and second wget tests really puzzled me.
Upload speed is pretty low, which was to my surprise, especially considering the VPS is supposed to be on a 1Gbps port. First is the results from my BuffaloVPS in Chicago, IL:
wget 173.193.165.236/100mb.test -O /dev/null --2011-10-07 22:43:17-- http://173.193.165.236/100mb.test Connecting to 173.193.165.236:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 4.01M/s in 28s 2011-10-07 22:43:45 (3.62 MB/s) - `/dev/null' saved [104857600/104857600]
Next from BuyVM VPS in San Jose, CA:
wget 173.193.165.236/100mb.test -O /dev/null --2011-10-07 22:44:30-- http://173.193.165.236/100mb.test Connecting to 173.193.165.236:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 3.01M/s in 42s 2011-10-07 22:45:12 (2.38 MB/s) - `/dev/null' saved [104857600/104857600]
And finally from my Quickweb VPS from London, UK:
wget 173.193.165.236/100mb.test -O /dev/null --2011-10-07 22:45:58-- http://173.193.165.236/100mb.test Connecting to 173.193.165.236:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 3.22M/s in 65s 2011-10-07 22:47:04 (1.53 MB/s) - `/dev/null' saved [104857600/104857600]
As you can see, the speed at the east coast is the best, followed by west coast and then Europe.
Finally, the benchmark time. Interestingly, the GeekBench test did not complain about my license, but instead I was caught with another error:
dist/Geekbench-2.2.0-Linux/geekbench_x86_64
Geekbench 2.2.0 : http://www.primatelabs.ca/geekbench/
Geekbench is in tryout mode
Geekbench is limited to running 32-bit benchmarks while in tryout mode.
Please purchase Geekbench to remove all limitations found in tryout mode.
If you would like to purchase Geekbench you can do so online:
http://store.primatelabs.ca/
If you have already purchased Geekbench, enter your email address and license
key from your email receipt with the following command line:
dist/Geekbench-2.2.0-Linux/geekbench_x86_64 -r
Run Geekbench as root to gather accurate system information.
Floating point exception
I am not sure if I have ever get Geekbench 2.2 to work, for some reason it is definitely not as VPS-friendly as its predecessor.
The UnixBench, as usual, managed to run properly though:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: : GNU/Linux
OS: GNU/Linux -- 2.6.32-5-amd64 -- #1 SMP Wed May 18 23:13:22 UTC 2011
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz (6784.6 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
22:58:26 up 9:39, 3 users, load average: 0.00, 0.01, 0.07; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Fri Oct 07 2011 22:58:26 - 23:26:32
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 36690330.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4040.7 MWIPS (9.9 s, 7 samples)
Execl Throughput 2142.7 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 303524.1 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 77960.9 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1001254.4 KBps (30.0 s, 2 samples)
Pipe Throughput 453192.7 lps (10.0 s, 7 samples)
Pipe-based Context Switching 91498.0 lps (10.0 s, 7 samples)
Process Creation 4305.4 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4552.3 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 597.6 lpm (60.0 s, 2 samples)
System Call Overhead 523649.3 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 36690330.4 3144.0
Double-Precision Whetstone 55.0 4040.7 734.7
Execl Throughput 43.0 2142.7 498.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 303524.1 766.5
File Copy 256 bufsize 500 maxblocks 1655.0 77960.9 471.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 1001254.4 1726.3
Pipe Throughput 12440.0 453192.7 364.3
Pipe-based Context Switching 4000.0 91498.0 228.7
Process Creation 126.0 4305.4 341.7
Shell Scripts (1 concurrent) 42.4 4552.3 1073.6
Shell Scripts (8 concurrent) 6.0 597.6 996.0
System Call Overhead 15000.0 523649.3 349.1
========
System Benchmarks Index Score 666.1
Given the fact that it just have a single core, the UnixBench actually ran really fast (comparing to those 24 core monsters) and I did another run the next morning just to see how things goes:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: : GNU/Linux
OS: GNU/Linux -- 2.6.32-5-amd64 -- #1 SMP Wed May 18 23:13:22 UTC 2011
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz (6784.6 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
10:06:42 up 20:47, 1 user, load average: 0.16, 0.42, 0.24; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Sat Oct 08 2011 10:06:42 - 10:34:48
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 36644939.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4048.3 MWIPS (9.9 s, 7 samples)
Execl Throughput 2125.2 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 303831.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 78273.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 992818.0 KBps (30.0 s, 2 samples)
Pipe Throughput 454269.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 91776.8 lps (10.0 s, 7 samples)
Process Creation 4345.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4533.3 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 595.1 lpm (60.0 s, 2 samples)
System Call Overhead 524091.9 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 36644939.5 3140.1
Double-Precision Whetstone 55.0 4048.3 736.0
Execl Throughput 43.0 2125.2 494.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 303831.9 767.3
File Copy 256 bufsize 500 maxblocks 1655.0 78273.1 472.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 992818.0 1711.8
Pipe Throughput 12440.0 454269.2 365.2
Pipe-based Context Switching 4000.0 91776.8 229.4
Process Creation 126.0 4345.3 344.9
Shell Scripts (1 concurrent) 42.4 4533.3 1069.2
Shell Scripts (8 concurrent) 6.0 595.1 991.8
System Call Overhead 15000.0 524091.9 349.4
========
System Benchmarks Index Score 665.9
For a single core CPU VPS, I think the benchmark score is actually pretty good.
After I have the UnixBench tests done, I submitted a ticket to get the OS changed to Debian 6 32 bit version and Matt has basically provisioned another VPS for me:
uname -a Linux 2.6.32-5-686-bigmem #1 SMP Mon Oct 3 05:03:32 UTC 2011 i686 GNU/Linux
The 32 bit version, as you can guess, is a lot more memory efficient compare to its 64 bit sibling, using only 15MB of memory after a fresh install:
free -m
total used free shared buffers cached
Mem: 121 116 4 0 26 74
-/+ buffers/cache: 15 105
Swap: 382 0 382
And the corresponding top output showing all the processes running:
top - 17:00:25 up 1 day, 4:05, 1 user, load average: 0.00, 0.00, 0.00
Tasks: 51 total, 1 running, 50 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 124336k total, 120636k used, 3700k free, 27592k buffers
Swap: 392184k total, 0k used, 392184k free, 76720k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
893 root 20 0 8476 2940 2320 S 0.0 2.4 0:00.07 sshd
896 root 20 0 5512 2800 1452 S 0.0 2.3 0:00.10 bash
597 root 20 0 27320 1456 1024 S 0.0 1.2 0:00.00 rsyslogd
938 root 20 0 2664 1196 1008 S 0.0 1.0 0:00.95 xe-daemon
1009 root 20 0 2340 1120 900 R 0.0 0.9 0:00.00 top
1004 root 20 0 5500 980 592 S 0.0 0.8 0:00.00 sshd
909 Debian-e 20 0 6524 944 624 S 0.0 0.8 0:00.02 exim4
695 root 20 0 3792 780 604 S 0.0 0.6 0:00.08 cron
488 statd 20 0 1944 776 640 S 0.0 0.6 0:00.00 rpc.statd
150 root 16 -4 2264 748 416 S 0.0 0.6 0:00.01 udevd
1 root 20 0 2040 724 628 S 0.0 0.6 0:00.79 init
205 root 18 -2 2260 636 308 S 0.0 0.5 0:00.00 udevd
212 root 18 -2 2260 636 308 S 0.0 0.5 0:00.00 udevd
1029 root 20 0 1712 560 480 S 0.0 0.5 0:00.00 getty
1031 root 20 0 1712 556 476 S 0.0 0.4 0:00.00 getty
1033 root 20 0 1712 556 476 S 0.0 0.4 0:00.00 getty
1034 root 20 0 1712 556 476 S 0.0 0.4 0:00.00 getty
And htop output for the htop fans:
The installation used about 650MB of hard drive space.
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda1 7.6G 648M 6.5G 9% / tmpfs 61M 0 61M 0% /lib/init/rw udev 51M 68K 51M 1% /dev tmpfs 61M 0 61M 0% /dev/shm /dev/xvdb 1008M 34M 924M 4% /disk1
With the full LNMP stack installed, only 28MB of RAM was used, which is pretty impressive:
free -m
total used free shared buffers cached
Mem: 121 118 2 0 5 84
-/+ buffers/cache: 28 92
Swap: 382 6 376
And the top output showing the processes running:
top - 05:33:56 up 1 day, 4:38, 1 user, load average: 0.05, 0.55, 0.57
Tasks: 61 total, 1 running, 60 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.3%us, 0.3%sy, 0.0%ni, 90.6%id, 8.7%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 124336k total, 121472k used, 2864k free, 5656k buffers
Swap: 392184k total, 6928k used, 385256k free, 86008k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
6648 www 20 0 14508 10m 412 S 0.0 8.6 0:00.00 nginx
6638 root 20 0 24208 4572 1408 S 0.0 3.7 0:00.02 php-cgi
6639 www 20 0 24208 4172 1008 S 0.0 3.4 0:00.00 php-cgi
6640 www 20 0 24208 4172 1008 S 0.0 3.4 0:00.00 php-cgi
6642 www 20 0 24208 4172 1008 S 0.0 3.4 0:00.00 php-cgi
6643 www 20 0 24208 4172 1008 S 0.0 3.4 0:00.00 php-cgi
6644 www 20 0 24208 4172 1008 S 0.0 3.4 0:00.00 php-cgi
1301 root 20 0 5512 1164 860 S 0.0 0.9 0:00.10 bash
6747 root 20 0 2340 1124 900 R 0.0 0.9 0:00.00 top
597 root 20 0 27320 928 732 S 0.0 0.7 0:00.01 rsyslogd
20003 mysql 20 0 34816 896 548 S 0.0 0.7 0:00.00 mysqld
938 root 20 0 2664 848 712 S 0.3 0.7 0:00.98 xe-daemon
1298 root 20 0 8404 784 668 S 0.0 0.6 0:00.65 sshd
6646 root 20 0 4696 700 264 S 0.0 0.6 0:00.00 nginx
1 root 20 0 2040 512 484 S 0.0 0.4 0:00.80 init
19901 root 20 0 1756 468 464 S 0.0 0.4 0:00.00 mysqld_safe
695 root 20 0 3792 452 408 S 0.0 0.4 0:00.08 cron
2GB of hard drive space was used, which is slightly less than the 64 bit version.
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda1 7.6G 2.0G 5.2G 28% / tmpfs 61M 0 61M 0% /lib/init/rw udev 51M 68K 51M 1% /dev tmpfs 61M 0 61M 0% /dev/shm /dev/xvdb 1008M 34M 924M 4% /disk1
The inodes limits are pretty low as well:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/xvda1 499712 74086 425626 15% / tmpfs 15542 5 15537 1% /lib/init/rw udev 12892 448 12444 4% /dev tmpfs 15542 1 15541 1% /dev/shm /dev/xvdb 65536 11 65525 1% /disk1
I have actually done the dd tests again on this VPS, but for the SATA III hard drives only, to see how it performs:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 16.2198 s, 66.2 MB/s
And test again:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 14.1243 s, 76.0 MB/s
As you can see, this time the disk I/O is a lot more stable compare to the tests I did on the 64 bit OS.
I tried, yet again, to get GeekBench 2.2 running on this VPS, and obviously, again, it failed:
Geekbench 2.2.0 : http://www.primatelabs.ca/geekbench/
Geekbench is in tryout mode
Geekbench is limited to running 32-bit benchmarks while in tryout mode.
Please purchase Geekbench to remove all limitations found in tryout mode.
If you would like to purchase Geekbench you can do so online:
http://store.primatelabs.ca/
If you have already purchased Geekbench, enter your email address and license
key from your email receipt with the following command line:
dist/Geekbench-2.2.0-Linux/geekbench_x86_32 -r
Run Geekbench as root to gather accurate system information.
Floating point exception
However, rolling back to GeekBench 2.1 did give me some pretty impressive results though:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 2.6.32-5-686-bigmem i686
Model: Linux PC (Intel(R) Xeon(R) CPU E31270 @ 3.40GHz)
Motherboard: Unknown Motherboard
Processor: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz
Processor ID: GenuineIntel Family 6 Model 42 Stepping 7
Logical Processors: 1
Physical Processors: 1
Processor Frequency: 3.39 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 256 KB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 121 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 2196 ||||||||
multi-threaded scalar 2354 |||||||||
Text Compress
single-threaded scalar 2963 |||||||||||
multi-threaded scalar 2901 |||||||||||
Text Decompress
single-threaded scalar 3204 ||||||||||||
multi-threaded scalar 3227 ||||||||||||
Image Compress
single-threaded scalar 2320 |||||||||
multi-threaded scalar 2304 |||||||||
Image Decompress
single-threaded scalar 2131 ||||||||
multi-threaded scalar 2309 |||||||||
Lua
single-threaded scalar 3856 |||||||||||||||
multi-threaded scalar 3753 |||||||||||||||
Floating Point
Mandelbrot
single-threaded scalar 2817 |||||||||||
multi-threaded scalar 2874 |||||||||||
Dot Product
single-threaded scalar 4544 ||||||||||||||||||
multi-threaded scalar 4805 |||||||||||||||||||
single-threaded vector 5495 |||||||||||||||||||||
multi-threaded vector 5992 |||||||||||||||||||||||
LU Decomposition
single-threaded scalar 3535 ||||||||||||||
multi-threaded scalar 3627 ||||||||||||||
Primality Test
single-threaded scalar 4893 |||||||||||||||||||
multi-threaded scalar 3920 |||||||||||||||
Sharpen Image
single-threaded scalar 10769 |||||||||||||||||||||||||||||||||||||||||||
multi-threaded scalar 10391 |||||||||||||||||||||||||||||||||||||||||
Blur Image
single-threaded scalar 8761 |||||||||||||||||||||||||||||||||||
multi-threaded scalar 8374 |||||||||||||||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 7763 |||||||||||||||||||||||||||||||
Write Sequential
single-threaded scalar 12008 ||||||||||||||||||||||||||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 4872 |||||||||||||||||||
Stdlib Write
single-threaded scalar 8315 |||||||||||||||||||||||||||||||||
Stdlib Copy
single-threaded scalar 15246 ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 6150 ||||||||||||||||||||||||
single-threaded vector 7232 ||||||||||||||||||||||||||||
Stream Scale
single-threaded scalar 6645 ||||||||||||||||||||||||||
single-threaded vector 7025 ||||||||||||||||||||||||||||
Stream Add
single-threaded scalar 6051 ||||||||||||||||||||||||
single-threaded vector 6905 |||||||||||||||||||||||||||
Stream Triad
single-threaded scalar 6306 |||||||||||||||||||||||||
single-threaded vector 5077 ||||||||||||||||||||
Integer Score: 2793 |||||||||||
Floating Point Score: 5771 |||||||||||||||||||||||
Memory Score: 9640 ||||||||||||||||||||||||||||||||||||||
Stream Score: 6423 |||||||||||||||||||||||||
Overall Geekbench Score: 5567 ||||||||||||||||||||||
And test again showed similar results:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 2.6.32-5-686-bigmem i686
Model: Linux PC (Intel(R) Xeon(R) CPU E31270 @ 3.40GHz)
Motherboard: Unknown Motherboard
Processor: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz
Processor ID: GenuineIntel Family 6 Model 42 Stepping 7
Logical Processors: 1
Physical Processors: 1
Processor Frequency: 3.39 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 256 KB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 121 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 2190 ||||||||
multi-threaded scalar 2346 |||||||||
Text Compress
single-threaded scalar 2931 |||||||||||
multi-threaded scalar 2901 |||||||||||
Text Decompress
single-threaded scalar 3210 ||||||||||||
multi-threaded scalar 3297 |||||||||||||
Image Compress
single-threaded scalar 2492 |||||||||
multi-threaded scalar 2444 |||||||||
Image Decompress
single-threaded scalar 2286 |||||||||
multi-threaded scalar 2356 |||||||||
Lua
single-threaded scalar 4197 ||||||||||||||||
multi-threaded scalar 4097 ||||||||||||||||
Floating Point
Mandelbrot
single-threaded scalar 2835 |||||||||||
multi-threaded scalar 2880 |||||||||||
Dot Product
single-threaded scalar 4627 ||||||||||||||||||
multi-threaded scalar 4938 |||||||||||||||||||
single-threaded vector 5527 ||||||||||||||||||||||
multi-threaded vector 6389 |||||||||||||||||||||||||
LU Decomposition
single-threaded scalar 3555 ||||||||||||||
multi-threaded scalar 3627 ||||||||||||||
Primality Test
single-threaded scalar 4990 |||||||||||||||||||
multi-threaded scalar 4004 ||||||||||||||||
Sharpen Image
single-threaded scalar 10778 |||||||||||||||||||||||||||||||||||||||||||
multi-threaded scalar 10653 ||||||||||||||||||||||||||||||||||||||||||
Blur Image
single-threaded scalar 8635 ||||||||||||||||||||||||||||||||||
multi-threaded scalar 8668 ||||||||||||||||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 7799 |||||||||||||||||||||||||||||||
Write Sequential
single-threaded scalar 12222 ||||||||||||||||||||||||||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 4975 |||||||||||||||||||
Stdlib Write
single-threaded scalar 8608 ||||||||||||||||||||||||||||||||||
Stdlib Copy
single-threaded scalar 16870 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 6499 |||||||||||||||||||||||||
single-threaded vector 7845 |||||||||||||||||||||||||||||||
Stream Scale
single-threaded scalar 6757 |||||||||||||||||||||||||||
single-threaded vector 7489 |||||||||||||||||||||||||||||
Stream Add
single-threaded scalar 6291 |||||||||||||||||||||||||
single-threaded vector 6986 |||||||||||||||||||||||||||
Stream Triad
single-threaded scalar 6780 |||||||||||||||||||||||||||
single-threaded vector 5161 ||||||||||||||||||||
Integer Score: 2895 |||||||||||
Floating Point Score: 5864 |||||||||||||||||||||||
Memory Score: 10094 ||||||||||||||||||||||||||||||||||||||||
Stream Score: 6726 ||||||||||||||||||||||||||
Overall Geekbench Score: 5757 |||||||||||||||||||||||
The UnixBench results is also better on this VPS than on the 64 bit one:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: : GNU/Linux
OS: GNU/Linux -- 2.6.32-5-686-bigmem -- #1 SMP Mon Oct 3 05:03:32 UTC 2011
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU E31270 @ 3.40GHz (6784.6 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
05:40:40 up 1 day, 4:45, 1 user, load average: 0.01, 0.20, 0.40; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Mon Oct 10 2011 05:40:40 - 06:08:50
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 20601074.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3332.3 MWIPS (10.0 s, 7 samples)
Execl Throughput 2730.6 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 510215.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 130466.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1538331.9 KBps (30.0 s, 2 samples)
Pipe Throughput 776675.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 91051.1 lps (10.0 s, 7 samples)
Process Creation 4931.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5025.1 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 661.5 lpm (60.0 s, 2 samples)
System Call Overhead 838003.6 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 20601074.5 1765.3
Double-Precision Whetstone 55.0 3332.3 605.9
Execl Throughput 43.0 2730.6 635.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 510215.0 1288.4
File Copy 256 bufsize 500 maxblocks 1655.0 130466.5 788.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 1538331.9 2652.3
Pipe Throughput 12440.0 776675.0 624.3
Pipe-based Context Switching 4000.0 91051.1 227.6
Process Creation 126.0 4931.7 391.4
Shell Scripts (1 concurrent) 42.4 5025.1 1185.2
Shell Scripts (8 concurrent) 6.0 661.5 1102.4
System Call Overhead 15000.0 838003.6 558.7
========
System Benchmarks Index Score 805.2
Overall, I think the performance of the VPS is above average. The disk I/O on the SSD is definitely impressive, however the network speed, particularly upload speed, is what that may needs some improvements along the way.
Customer Service and Support
Since at the moment, there is no control panel available, and everything has to be done by human via tickets (including reinstalling of OS, which could take quite a bit of time), I was thinking SSD Nodes must be overwhelmed with tickets and therefore the response could be really slow. To my pleasant surprise, it is actually not the case. For the simplest technical questions, I have received a reply barely 3 minutes after I have submitted the ticket (ticket submitted at 7:12PM, received reply at 7:15PM, on a Sunday as well), and even my little complain to them for lack of control panel, which was replied with a very detailed explanation, was answered in 14 minutes on Friday evening. Even billing (which is typically the “slower” department) was able to get back to me in 5 minutes when I had trouble paying invoice and Matt actually sent me an email, even before I got a chance to submit my billing ticket, to ask me if I had any trouble paying the invoice, which was definitely the only one I have seen among all the VPS I have reviewed so far.
Conclusion
SSD Nodes has definitely offered us something pretty unique: a hybrid VPS with both SSD and SATA III drives, which makes it really economical, but also delivers the performance if you really need a good disk I/O. There are definitely a few areas for improvements, for example, getting a control panel up and running, getting stable network speed and higher inodes. However, with the impressive disk I/O of the SSD drives and fast customer support, SSD Nodes is definitely worth consideration if you need a VPS that partially requires high disk I/O without costing too much.
Pingback: Thank you! | | 96MB.com96MB.com
Pingback: 96MB Low End VPS Review Part XXXXI – VPSDeploy SSD | | 96MB.com96MB.com
Pingback: SSD Nodes up to 128GB RAM! Premium Network w/ 2×1-10Gbps Public/Private Interfaces » Dedicated Offer - gathering the best hosting deals for you.
Pingback: SSD Web Hosting | 8x Tier 1 & 2,000Gbps | Supercomputers at your Fingertips
Pingback: SSD Web Hosting 8x Tier 1 & 2,000Gbps Supercomputers at your Fingertips
Pingback: SSD Nodes Innovators, Meet Our Platform. High Performance Starting at $799/mo | CheapHostingOffer.com
Pingback: SSD Nodes | 50% off NEW Cloud | 512MB RAM for $4.95/mo | 7 Day Refund! | Canada | Host Offers Only
Pingback: CheapVPSDeals | HPC SSD Nodes up to 240GB RAM | 10Gbps NICs | Canada | Starting at $0.04/hr & $10/mo
Pingback: The Fastest HPC SSD Nodes | Dedicated I/O + 10Gbps NICs | Canada | Starting at $20/mo | Lost in the forest
Pingback: CheapVPSDeals | WHT Outlet: $9.99/mo for 2GB RAM + Enterprise SSDs | Supercharge Your Business Today!
Pingback: CheapVPSDeals | The Fastest HPC SSD Cloud – Dedicated Performance + 10Gbps NICs
Pingback: SSD Nodes – Dedicated High Performance and 10Gbps NICs – Limited time WHT Deal! | TheHostMag
Pingback: SSD Nodes - Dedicated High Performance and 10Gbps NICs - Limited time WHT Deal! - CheapVPSDeals
Pingback: SSD Nodes – 10Gbps NICs / Dedicated Performance – Your Business on Steroids | TheHostMag
Pingback: SSD Nodes - 10Gbps NICs / Dedicated Performance - Your Business on Steroids - CheapVPSDeals
Pingback: SSD Nodes – Your Business on Steroids – Get started for $10/mo! (Limited WHT offer) | TheHostMag
Pingback: SSD Nodes | 100x Faster than EC2 + 8x Backbone Providers | Annual = 2 Months FREE! - CheapVPSDeals
Pingback: █ SSD Nodes █ 100x Faster than EC2, 8x Tier-1 Providers █ Annual = 2 Months FREE! | CheapHostingOffer.com
Pingback: █ SSD Nodes █ IBM 8x Tier-1 ISPs █ Direct Engineer Support - CheapVPSDeals
Pingback: New Enterprise Offsite Storage | US/EU/HK Locations | Petabyte Scale | CheapHostingOffer.com
Pingback: $4/mo for 4GB RAM in Dallas and Canada! Fastest RAID 10 SSDs: 1GB/s IO & 200K IOPS. Get Started Now! - CheapVPSDeals