SkylonHost (FHR) benchmark and review

First, I want to thank @FHR who was generous enough to provide me with access to one of their VPS (KVM) with 2 vCores, 2 GB memory and a SSD hosted in Prague, CZ.

This is the system data:

Machine: amd64, Arch.: x86_64, Model: Intel(R) Xeon(R) CPU E5-2630 v2 @ 2.60GHz
OS, version: Linux 4.19.0, Mem.: 1.971 GB
CPU - Cores: 2, Family/Model/Stepping: 6/62/4
Cache: 32K/32K L1d/L1i, 256K L2, 15M L3
Std. Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
pse36 cflsh mmx fxsr sse sse2 ss sse3 pclmulqdq ssse3 cx16 pcid
sse4_1 sse4_2 x2apic popcnt tsc_deadline aes xsave osxsave avx f16c
rdrnd hypervisor
Ext. Flags: fsgsbase tsc_adjust smep erms syscall nx pdpe1gb rdtscp lm lahf_lm


So the VPS comes with a reasonably decent processors and has all the usual flags available, in particular aes, sseX, avs, etc.

Please note in the following results that the benchmark was performed under Linux and hence the disk performance is probably "optimistic" (Linux tends to make disks look considerably faster than they actually are in real world use, particularly on a VPS).
Note - If you just want a quick overview: The first column always is the average result (over all benchmark runs).

First the processor:

Avg Min Max Dmin Dmax Span


Single core: 254,13 244,80 276,38 3,67 8,76 12,43
Multi core: 535,14 493,08 599,85 7,86 12,09 19,95

That's a very decent result and I'd like to stress in particular that the multi core results indeed reflects an adequate performance and also suggests that the node is not oversold. Deviation both towards min and towards max are within a tight boundary. Nice.

Now the disk:

Seq Wr: 371,95 287,80 620,89 22,62 66,93 89,55
Rnd Wr: 1.317 1.138 1.521 13,63 15,44 29,07
Seq Rd: 3.722 3.174 4.001 14,73 7,48 22,22
Rnd Rd: 2.714 2.523 2.908 7,06 7,13 14,18


Those are quite nice results and suggest that the SSD drive is really fast. One notes however that sequential writing has quite considerable deviations that indicate that I/O is limited and writing to the disk only occasionally achieves the speed the disk is capable off.
But still, all in all a really nice result set.

Finally the network:

GR_UNK: 142,72 111,00 176,50 22,23 23,67 45,89
US_SJC: 83,62 79,00 87,60 5,53 4,76 10,28
IT_MIL: 198,76 156,10 332,60 21,46 67,34 88,80
DE_FRA: 240,81 183,70 403,00 23,72 67,35 91,07
BR_SAO: 60,31 10,50 72,90 82,59 20,87 103,46
FR_PAR: 208,46 157,30 338,90 24,54 62,58 87,12
US_WDC: 117,11 41,90 161,60 64,22 37,99 102,21
UK_LON: 220,03 127,10 344,70 42,24 56,66 98,89
AU_MEL: 46,56 43,90 49,10 5,70 5,47 11,17
SG_SGP: 98,13 78,50 106,80 20,01 8,83 28,84
US_DAL: 100,69 90,80 113,60 9,82 12,82 22,64
RO_BUC: 193,68 72,50 298,20 62,57 53,97 116,53
RU_MOS: 205,38 179,80 276,30 12,45 34,53 46,99
JP_TOK: 64,58 61,10 70,20 5,39 8,71 14,09
NO_OSL: 180,56 139,70 257,60 22,63 42,67 65,30
IN_CHN: 92,79 85,50 103,10 7,86 11,11 18,97


Being hosted in CZ DE_FRA is not surprisingly the fastest target with the other major european locations following but what I really like is that even well not "problem" locations like Sao Paolo and India, Chennai have quite good results. That is also true for the worst results in the series. Nice, really nice.
If asked for constructive criticism I'd suggest SkylonHost to work towards using more of the potential both in networking and the disk.

But all in all a really nice VPS.

Side note: As often unless specifically asked to do so I did not have a closer look at the panel but it has everything one needs.

Please excuse the poor formatting. I guess I'll never become a HTML master ...

Top News