2 NB alternatives - Webhorizon & Greencloud benchmark & review
I'm certainly not the only one who needed to find a replacement for NexusBytes's (now largely dead, it seems) VPS. So I was looking around and got myself a couple of (purchased normally) VPS from providers on LET with whom I had not yet any concrete experience and which are roughly in the ballpark of NB.
And then, of bloody course, I immediately went to benchmark them two full days around the clock ...
From Greencloudvps / @NDTN I first got a dirt cheap ($15/yr or $40/3yrs) basic VPS with 1 vCore (E5 v4), 2 GB mem., 20 GB SSD in UK and was very pleased with it, so much so that I right away purchased a "big" Budget KVM, also in UK and with 4 vCores (E5 v4), 8 GB mem and 60 GB SSD ($45/yr). It's that VPS benchmark results I show now and review. Based on 62 benchmark runs.
First system info and processor/memory
Version 2.5.0a, (c) 2018+ jsg (->lowendtalk.com)
Machine: amd64, Arch.: amd64, Model: Intel(R) Xeon(R) CPU E5-2698 v4 @ 2.20GHz
OS, version: FreeBSD 13.1, Mem.: 7.990 GB
CPU - Cores: 4, Family/Model/Stepping: 6/79/1
Cache: 32K/32K L1d/L1i, 2M L2, 16M 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 vmx ssse3 fma cx16
pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline aes xsave
osxsave avx f16c rdrnd hypervisor
Ext. Flags: fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed
adx smap umip syscall nx pdpe1gb rdtscp lm lahf_lm lzcnt
AES? Yes
Nested Virt.? Yes
HW RNG? Yes
ProcMem SC [MB/s]: avg 126.2 - min 43.7 (34.6 %), max 218.7 (173.3 %)
ProcMem MA [MB/s]: avg 490.8 - min 406.4 (82.8 %), max 608.6 (124.0 %)
ProcMem MB [MB/s]: avg 557.1 - min 464.7 (83.4 %), max 627.4 (112.6 %)
ProcMem AES [MB/s]: avg 442.2 - min 311.5 (70.4 %), max 531.1 (120.1 %)
ProcMem RSA [kp/s]: avg 66.8 - min 43.7 (65.4 %), max 82.5 (123.5 %)
Mee, that's not a great result, that doesn't look like v4 but rather like v2 and not even a great v2 processor. Probably a mix of high occupancy and slow RAM. But then 8 GB of memory and 4 cores (with near perfect 4x single core performance, which is by no means usual, so multithreaded applications should run quite nicely on that VPS) for $45 per year still is a nice deal.
Now the disk which btw is a 60 GB SSD
--- Disk 4 KB - Buffered ---
Write seq. [MB/s]: avg 3.48 - min 3.02 (86.9%), max 4.09 (117.6%)
Write rnd. [MB/s]: avg 3.37 - min 2.92 (86.7%), max 3.94 (117.0%)
Read seq. [MB/s]: avg 9.22 - min 7.53 (81.7%), max 10.47 (113.6%)
Read rnd. [MB/s]: avg 6.74 - min 5.34 (79.2%), max 9.02 (133.7%)
--- Disk 4 KB - Sync/Direct ---
Write seq. [MB/s]: avg 3.41 - min 2.31 (67.7%), max 3.94 (115.5%)
Write rnd. [MB/s]: avg 3.32 - min 2.93 (88.2%), max 3.82 (114.9%)
Read seq. [MB/s]: avg 9.27 - min 6.98 (75.3%), max 10.93 (117.9%)
Read rnd. [MB/s]: avg 6.54 - min 2.57 (39.3%), max 9.25 (141.5%)
--- Disk 64 KB - Buffered ---
Write seq. [MB/s]: avg 31.52 - min 27.07 (85.9%), max 36.45 (115.7%)
Write rnd. [MB/s]: avg 34.92 - min 29.99 (85.9%), max 41.32 (118.3%)
Read seq. [MB/s]: avg 523.43 - min 426.21 (81.4%), max 1002.10 (191.4%)
Read rnd. [MB/s]: avg 96.98 - min 73.05 (75.3%), max 132.76 (136.9%)
--- Disk 64 KB - Sync/Direct ---
Write seq. [MB/s]: avg 5.22 - min 4.52 (86.6%), max 5.80 (111.1%)
Write rnd. [MB/s]: avg 2.74 - min 2.53 (92.4%), max 2.97 (108.4%)
Read seq. [MB/s]: avg 517.71 - min 426.44 (82.4%), max 967.28 (186.8%)
Read rnd. [MB/s]: avg 98.94 - min 72.56 (73.3%), max 126.90 (128.3%)
--- Disk 1 MB - Buffered ---
Write seq. [MB/s]: avg 45.00 - min 37.76 (83.9%), max 55.24 (122.8%)
Write rnd. [MB/s]: avg 90.21 - min 72.15 (80.0%), max 110.71 (122.7%)
Read seq. [MB/s]: avg 938.28 - min 784.31 (83.6%), max 1560.72 (166.3%)
Read rnd. [MB/s]: avg 597.42 - min 530.15 (88.7%), max 666.57 (111.6%)
--- Disk 1 MB - Sync/Direct ---
Write seq. [MB/s]: avg 14.50 - min 13.30 (91.7%), max 17.00 (117.2%)
Write rnd. [MB/s]: avg 10.38 - min 9.67 (93.2%), max 11.29 (108.8%)
Read seq. [MB/s]: avg 940.22 - min 774.72 (82.4%), max 1384.31 (147.2%)
Read rnd. [MB/s]: avg 581.84 - min 510.19 (87.7%), max 656.93 (112.9%)
--- Disk IOps (Sync/Direct) ---
Write seq. [MB/s]: avg 16.64 - min 13.49 (81.1%), max 21.74 (130.7%)
IOps : avg 4258.54 - min 3452.43 (81.1%), max 5565.12 (130.7%)
Not bad at all for a really cheap VPS. Not really great but hey, this is a dirt cheap VPS and for that it's really nice especially when considering the not at all excessive spread. Not a disk to brag about but neither one to complain about, I'm fine with it.
And finally the network
US SJC speedtest.sjc01.softlayer.com [F: 0]
DL [Mb/s]: avg 46.5 - min 32.5 (69.9%), max 71.9 (154.6%)
Ping [ms]: avg 144.9 - min 144.2 (99.5%), max 161.1 (111.2%)
Web ping [ms]: avg 145.8 - min 144.2 (98.9%), max 180.6 (123.9%)
FR PAR ipv4.paris.testdebit.info [F: 0]
DL [Mb/s]: avg 723.5 - min 395.9 (54.7%), max 899.1 (124.3%)
Ping [ms]: avg 12.5 - min 12.3 (98.4%), max 15.5 (124.0%)
Web ping [ms]: avg 13.1 - min 12.4 (94.7%), max 32.2 (245.9%)
AU MEL speedtest.c1.mel1.dediserve.com [F: 12]
DL [Mb/s]: avg 20.7 - min 0.0 (0.0%), max 37.0 (179.4%)
Ping [ms]: avg 278.2 - min 0.0 (0.0%), max 286.8 (103.1%)
Web ping [ms]: avg 279.1 - min 0.0 (0.0%), max 289.8 (103.8%)
JP TOK speedtest.tokyo2.linode.com [F: 0]
DL [Mb/s]: avg 35.3 - min 24.2 (68.4%), max 45.5 (129.0%)
Ping [ms]: avg 246.3 - min 244.1 (99.1%), max 334.6 (135.9%)
Web ping [ms]: avg 248.8 - min 244.1 (98.1%), max 334.6 (134.5%)
IT MIL speedtest.mil01.softlayer.com [F: 0]
DL [Mb/s]: avg 176.5 - min 77.7 (44.0%), max 281.4 (159.5%)
Ping [ms]: avg 24.4 - min 24.3 (99.5%), max 24.5 (100.3%)
Web ping [ms]: avg 27.3 - min 24.3 (88.9%), max 87.5 (320.2%)
US NYC nyc.speedtest.clouvider.net [F: 0]
DL [Mb/s]: avg 108.7 - min 60.9 (56.0%), max 144.0 (132.5%)
Ping [ms]: avg 73.6 - min 73.3 (99.7%), max 74.3 (101.0%)
Web ping [ms]: avg 75.0 - min 73.3 (97.7%), max 121.9 (162.6%)
IN MUM mirrors.piconets.webwerks.in [F: 0]
DL [Mb/s]: avg 66.0 - min 38.9 (58.9%), max 91.7 (138.9%)
Ping [ms]: avg 125.0 - min 120.0 (96.0%), max 140.8 (112.6%)
Web ping [ms]: avg 140.2 - min 120.1 (85.7%), max 193.4 (138.0%)
GR UNK speedtest.ftp.otenet.gr [F: 16]
DL [Mb/s]: avg 113.4 - min 0.0 (0.0%), max 223.0 (196.7%)
Ping [ms]: avg 36.6 - min 0.0 (0.0%), max 91.0 (248.8%)
Web ping [ms]: avg 37.1 - min 0.0 (0.0%), max 91.0 (245.1%)
SG SGP mirror.sg.leaseweb.net [F: 0]
DL [Mb/s]: avg 45.2 - min 28.7 (63.5%), max 56.7 (125.5%)
Ping [ms]: avg 202.1 - min 201.9 (99.9%), max 203.5 (100.7%)
Web ping [ms]: avg 202.6 - min 201.9 (99.7%), max 208.9 (103.1%)
BR SAO speedtest.sao01.softlayer.com [F: 29]
DL [Mb/s]: avg 14.7 - min 0.0 (0.0%), max 38.5 (261.7%)
Ping [ms]: avg 204.6 - min 202.3 (98.9%), max 205.8 (100.6%)
Web ping [ms]: avg 205.8 - min 202.3 (98.3%), max 225.8 (109.7%)
US WDC mirror.wdc1.us.leaseweb.net [F: 17]
DL [Mb/s]: avg 55.8 - min 0.0 (0.0%), max 137.0 (245.7%)
Ping [ms]: avg 80.9 - min 80.6 (99.6%), max 82.0 (101.3%)
Web ping [ms]: avg 89.6 - min 80.6 (90.0%), max 575.9 (642.7%)
RU MOS speedtest.hostkey.ru [F: 0]
DL [Mb/s]: avg 158.6 - min 74.0 (46.6%), max 224.5 (141.6%)
Ping [ms]: avg 52.2 - min 50.5 (96.7%), max 57.6 (110.3%)
Web ping [ms]: avg 55.0 - min 51.0 (92.6%), max 84.9 (154.2%)
US LAX la.speedtest.clouvider.net [F: 0]
DL [Mb/s]: avg 63.9 - min 34.5 (54.0%), max 82.9 (129.8%)
Ping [ms]: avg 131.3 - min 129.8 (98.9%), max 136.6 (104.0%)
Web ping [ms]: avg 133.7 - min 129.9 (97.1%), max 185.4 (138.7%)
NL AMS mirror.nl.leaseweb.net [F: 0]
DL [Mb/s]: avg 543.3 - min 197.6 (36.4%), max 955.9 (175.9%)
Ping [ms]: avg 11.6 - min 11.4 (98.2%), max 13.6 (117.2%)
Web ping [ms]: avg 11.8 - min 11.4 (96.6%), max 18.2 (154.2%)
CN HK mirror.hk.leaseweb.net [F: 0]
DL [Mb/s]: avg 46.8 - min 25.4 (54.3%), max 54.4 (116.1%)
Ping [ms]: avg 202.9 - min 202.5 (99.8%), max 203.6 (100.4%)
Web ping [ms]: avg 206.5 - min 202.7 (98.1%), max 281.1 (136.1%)
US DAL mirror.dal13.us.leaseweb.net [F: 0]
DL [Mb/s]: avg 80.8 - min 47.8 (59.1%), max 101.5 (125.7%)
Ping [ms]: avg 109.8 - min 109.4 (99.7%), max 112.9 (102.8%)
Web ping [ms]: avg 110.7 - min 109.4 (98.9%), max 122.8 (111.0%)
DE FRA fra.lg.core-backbone.com [F: 0]
DL [Mb/s]: avg 488.2 - min 241.3 (49.4%), max 746.7 (153.0%)
Ping [ms]: avg 14.5 - min 14.3 (98.5%), max 15.1 (104.1%)
Web ping [ms]: avg 15.1 - min 14.4 (95.6%), max 33.4 (221.8%)
UK LON lon.speedtest.clouvider.net [F: 0]
DL [Mb/s]: avg 1422.1 - min 556.5 (39.1%), max 2197.0 (154.5%)
Ping [ms]: avg 6.5 - min 4.1 (62.8%), max 17.1 (261.9%)
Web ping [ms]: avg 6.6 - min 4.1 (62.1%), max 17.1 (259.2%)
BANG, surprise! And a very nice one. The three major european locations about or above 500 Mb/s and London even solidly above 1 Gb/s, yay! Somewhat weird IMO with e.g. IT, Mil far behind FR, Par and DE, Fra or US, Nyc with respectable 100+ Mb/s but US, Wdc just half of that, but still, all in all very nice connectivity, especially within Europe.
Oh, and their support is really, really great. Fast (as in "often just 5 or 10 minutes"), helpful and so far never offered me canned BS but to the point help.
Comments
Now, on to Webhorizon
This one is a "Ryzen beast system" and costs $25/yr.
Again, first the sysinfo and proc/mem results based on a bit over 40 runs (side note, both VPS test runs started almost simultaneously but this VPS here got rebooted by WH, hence less runs although each one ran faster than on GC VPS).
Version 2.5.0a, (c) 2018+ jsg (->lowendtalk.com)
Machine: amd64, Arch.: amd64, Model: AMD Ryzen 9 5950X 16-Core Processor
OS, version: FreeBSD 13.1, Mem.: 1.990 GB
CPU - Cores: 1, Family/Model/Stepping: 25/33/0
Cache: 32K/32K L1d/L1i, 512K L2, 64M 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 sse3 pclmulqdq ssse3 fma cx16 sse4_1
sse4_2 popcnt aes xsave osxsave avx f16c rdrnd hypervisor
Ext. Flags: syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm lahf_lm cmp_legacy svm
cr8_legacy lzcnt sse4a misalignsse 3dnowprefetch osvw perfctr_core
AES? Yes
Nested Virt.? Yes
HW RNG? Yes
ProcMem SC [MB/s]: avg 375.4 - min 137.0 (36.5 %), max 618.0 (164.6 %)
ProcMem MA [MB/s]: avg 593.5 - min 562.6 (94.8 %), max 609.9 (102.8 %)
ProcMem MB [MB/s]: avg 588.4 - min 559.4 (95.1 %), max 610.2 (103.7 %)
ProcMem AES [MB/s]: avg 1585.9 - min 1524.4 (96.1 %), max 1628.7 (102.7 %)
ProcMem RSA [kp/s]: avg 148.3 - min 137.0 (92.4 %), max 154.8 (104.4 %)
Of bloody course that ZEN processor is fasssst, but look at the AES and SSL keypairs numbers! That's significantly faster than any VPS I've seen so far. I'm very pleased with what I get out of a single vCore (hint: more than out of the 4 vCores VPS tested above ...).
Now, the 60 GB NVMe
--- Disk 4 KB - Buffered ---
Write seq. [MB/s]: avg 18.94 - min 17.03 (89.9%), max 20.46 (108.0%)
Write rnd. [MB/s]: avg 18.53 - min 16.64 (89.8%), max 20.27 (109.4%)
Read seq. [MB/s]: avg 53.35 - min 45.38 (85.1%), max 60.81 (114.0%)
Read rnd. [MB/s]: avg 47.13 - min 41.76 (88.6%), max 55.38 (117.5%)
--- Disk 4 KB - Sync/Direct ---
Write seq. [MB/s]: avg 18.41 - min 16.78 (91.2%), max 21.29 (115.7%)
Write rnd. [MB/s]: avg 18.27 - min 16.63 (91.0%), max 19.76 (108.1%)
Read seq. [MB/s]: avg 52.71 - min 45.62 (86.6%), max 57.94 (109.9%)
Read rnd. [MB/s]: avg 46.99 - min 41.71 (88.8%), max 51.50 (109.6%)
--- Disk 64 KB - Buffered ---
Write seq. [MB/s]: avg 176.81 - min 119.59 (67.6%), max 202.90 (114.8%)
Write rnd. [MB/s]: avg 191.72 - min 173.59 (90.5%), max 212.35 (110.8%)
Read seq. [MB/s]: avg 4474.85 - min 3578.97 (80.0%), max 5484.29 (122.6%)
Read rnd. [MB/s]: avg 598.68 - min 512.88 (85.7%), max 720.36 (120.3%)
--- Disk 64 KB - Sync/Direct ---
Write seq. [MB/s]: avg 23.23 - min 19.09 (82.2%), max 26.96 (116.1%)
Write rnd. [MB/s]: avg 11.49 - min 8.55 (74.4%), max 13.03 (113.4%)
Read seq. [MB/s]: avg 4491.99 - min 3567.76 (79.4%), max 5425.88 (120.8%)
Read rnd. [MB/s]: avg 596.79 - min 519.76 (87.1%), max 673.52 (112.9%)
--- Disk 1 MB - Buffered ---
Write seq. [MB/s]: avg 274.01 - min 231.86 (84.6%), max 298.28 (108.9%)
Write rnd. [MB/s]: avg 507.34 - min 451.87 (89.1%), max 568.85 (112.1%)
Read seq. [MB/s]: avg 5906.37 - min 4654.57 (78.8%), max 7115.63 (120.5%)
Read rnd. [MB/s]: avg 2147.20 - min 1742.83 (81.2%), max 2683.48 (125.0%)
--- Disk 1 MB - Sync/Direct ---
Write seq. [MB/s]: avg 96.90 - min 74.20 (76.6%), max 110.08 (113.6%)
Write rnd. [MB/s]: avg 92.23 - min 68.71 (74.5%), max 108.41 (117.5%)
Read seq. [MB/s]: avg 5874.59 - min 4750.37 (80.9%), max 6633.30 (112.9%)
Read rnd. [MB/s]: avg 2152.14 - min 1818.68 (84.5%), max 2439.37 (113.3%)
--- Disk IOps (Sync/Direct) ---
Write seq. [MB/s]: avg 73.83 - min 65.98 (89.4%), max 81.00 (109.7%)
IOps : avg 18900.38 - min 16889.95 (89.4%), max 20734.89 (109.7%)
In case you are still looking for a definition of "screaming fast", here you go. No more questions you honour, I'm fully satisfied. YAY!
And finally the network
US SJC speedtest.sjc01.softlayer.com [F: 7]
DL [Mb/s]: avg 40.8 - min 0.0 (0.0%), max 64.4 (157.8%)
Ping [ms]: avg 164.8 - min 156.1 (94.7%), max 178.9 (108.5%)
Web ping [ms]: avg 165.2 - min 156.1 (94.5%), max 178.9 (108.3%)
FR PAR ipv4.paris.testdebit.info [F: 0]
DL [Mb/s]: avg 325.7 - min 139.8 (42.9%), max 432.0 (132.6%)
Ping [ms]: avg 21.8 - min 20.1 (92.1%), max 35.4 (162.1%)
Web ping [ms]: avg 22.5 - min 20.3 (90.3%), max 37.9 (168.5%)
AU MEL speedtest.c1.mel1.dediserve.com [F: 11]
DL [Mb/s]: avg 22.6 - min 0.0 (0.0%), max 38.7 (171.1%)
Ping [ms]: avg 279.9 - min 278.2 (99.4%), max 292.4 (104.5%)
Web ping [ms]: avg 290.9 - min 278.3 (95.7%), max 323.5 (111.2%)
JP TOK speedtest.tokyo2.linode.com [F: 17]
DL [Mb/s]: avg 19.4 - min 0.0 (0.0%), max 48.0 (247.9%)
Ping [ms]: avg 267.1 - min 265.4 (99.4%), max 279.2 (104.5%)
Web ping [ms]: avg 267.3 - min 265.4 (99.3%), max 279.2 (104.5%)
IT MIL speedtest.mil01.softlayer.com [F: 5]
DL [Mb/s]: avg 173.0 - min 0.0 (0.0%), max 248.7 (143.7%)
Ping [ms]: avg 39.1 - min 38.1 (97.4%), max 52.0 (132.9%)
Web ping [ms]: avg 39.5 - min 38.2 (96.7%), max 52.4 (132.7%)
US NYC nyc.speedtest.clouvider.net [F: 5]
DL [Mb/s]: avg 97.1 - min 0.0 (0.0%), max 132.6 (136.5%)
Ping [ms]: avg 93.5 - min 89.2 (95.4%), max 105.3 (112.6%)
Web ping [ms]: avg 93.8 - min 89.2 (95.1%), max 105.3 (112.3%)
IN MUM mirrors.piconets.webwerks.in [F: 5]
DL [Mb/s]: avg 63.0 - min 0.0 (0.0%), max 88.5 (140.6%)
Ping [ms]: avg 134.9 - min 127.3 (94.4%), max 142.0 (105.3%)
Web ping [ms]: avg 148.8 - min 127.4 (85.6%), max 192.9 (129.6%)
GR UNK speedtest.ftp.otenet.gr [F: 17]
DL [Mb/s]: avg 111.7 - min 0.0 (0.0%), max 229.7 (205.6%)
Ping [ms]: avg 33.5 - min 0.0 (0.0%), max 58.5 (174.6%)
Web ping [ms]: avg 41.1 - min 0.0 (0.0%), max 358.5 (871.3%)
SG SGP mirror.sg.leaseweb.net [F: 6]
DL [Mb/s]: avg 50.8 - min 0.0 (0.0%), max 67.9 (133.6%)
Ping [ms]: avg 169.4 - min 167.9 (99.1%), max 182.0 (107.5%)
Web ping [ms]: avg 170.8 - min 167.9 (98.3%), max 182.0 (106.5%)
BR SAO speedtest.sao01.softlayer.com [F: 8]
DL [Mb/s]: avg 30.2 - min 0.0 (0.0%), max 46.4 (153.8%)
Ping [ms]: avg 217.4 - min 200.0 (92.0%), max 231.8 (106.6%)
Web ping [ms]: avg 219.9 - min 200.1 (91.0%), max 237.3 (107.9%)
US WDC mirror.wdc1.us.leaseweb.net [F: 6]
DL [Mb/s]: avg 78.6 - min 0.0 (0.0%), max 113.2 (144.1%)
Ping [ms]: avg 105.2 - min 102.0 (96.9%), max 199.6 (189.7%)
Web ping [ms]: avg 106.5 - min 102.1 (95.9%), max 231.1 (217.1%)
RU MOS speedtest.hostkey.ru [F: 4]
DL [Mb/s]: avg 160.8 - min 0.0 (0.0%), max 225.7 (140.4%)
Ping [ms]: avg 53.5 - min 51.5 (96.2%), max 62.3 (116.4%)
Web ping [ms]: avg 56.9 - min 51.5 (90.5%), max 72.1 (126.7%)
US LAX la.speedtest.clouvider.net [F: 9]
DL [Mb/s]: avg 45.8 - min 0.0 (0.0%), max 70.9 (154.6%)
Ping [ms]: avg 163.2 - min 151.3 (92.7%), max 192.4 (117.9%)
Web ping [ms]: avg 165.6 - min 151.6 (91.6%), max 192.4 (116.2%)
NL AMS mirror.nl.leaseweb.net [F: 1]
DL [Mb/s]: avg 272.0 - min 0.0 (0.0%), max 421.0 (154.8%)
Ping [ms]: avg 12.4 - min 11.0 (88.5%), max 35.9 (289.0%)
Web ping [ms]: avg 13.2 - min 11.1 (84.2%), max 35.9 (272.2%)
CN HK mirror.hk.leaseweb.net [F: 17]
DL [Mb/s]: avg 18.1 - min 0.0 (0.0%), max 50.4 (279.0%)
Ping [ms]: avg 220.9 - min 219.2 (99.2%), max 231.0 (104.6%)
Web ping [ms]: avg 221.1 - min 219.3 (99.2%), max 231.0 (104.5%)
US DAL mirror.dal13.us.leaseweb.net [F: 5]
DL [Mb/s]: avg 70.1 - min 0.0 (0.0%), max 93.0 (132.7%)
Ping [ms]: avg 125.5 - min 124.1 (98.9%), max 134.1 (106.9%)
Web ping [ms]: avg 126.4 - min 124.1 (98.2%), max 134.1 (106.1%)
DE FRA fra.lg.core-backbone.com [F: 2]
DL [Mb/s]: avg 272.6 - min 0.0 (0.0%), max 423.0 (155.2%)
Ping [ms]: avg 18.4 - min 16.9 (91.9%), max 32.3 (175.6%)
Web ping [ms]: avg 18.5 - min 17.1 (92.3%), max 32.3 (174.3%)
UK LON lon.speedtest.clouvider.net [F: 1]
DL [Mb/s]: avg 253.9 - min 0.0 (0.0%), max 444.4 (175.0%)
Ping [ms]: avg 18.1 - min 17.3 (95.7%), max 25.2 (139.5%)
Web ping [ms]: avg 19.8 - min 17.4 (87.9%), max 36.7 (185.4%)
Quite nice but a bit weird, e.g. FR, Par is faster than NL, Ams with a VPS located in NL?? Weird. But careful, while this VPS's connectivity might look poor at first sight when compared to the above GC VPS, I'm actually not that sure. Yes, the nice "porn spikes" (like 700+ Mb/s) are missing, but this VPS IMO actually looks more consistent and all around the world. If you need decent global connectivity, you'll probably be better off with this VPS.
WH's support is good and relatively fast (I'd put it in the 1 - 6 hrs bucket which is quite good in the LE world)
TL;DR, Summary of both:
If your priority is a seriously fast and sexy processor, a screaming fast NVMe, decent (but not great) connectivity, the Webhorizon VPS clearly is the way to go.
I personally however will gladly keep my WH VPS but prefer the GC ones, Two main reasons: (a) I'm almost exclusively focused on Europe (connectivity) and in case I need something in the Americas or in Asia I do have some (1 or 2) VPS there (mainly Contabos, I'm still really happy with them), and (b) and more importantly for me, in my field of work (mainly crypto and server centered) I strongly prefer more cores (even if mediocre ones) and in particular more RAM over a very fast core with relatively little memory. So, for me personally 4 E5 v4 cores (although somewhat disappointing) plus 8GB memory trump the fast Ryzen with 2 GB memory. Depending on what type it might be different if I ran web+app servers though.
Side note: Regrettably I only know GC's UK location and would be interested in seeing some other locations, too, e.g. 1 in each, Asia and the Americas. So, @NDTN if that isn't asking too much I'd really like to get access to those for a couple of days or a week ... * wink wink nudge nudge
So, I think actually both VPSs offer good to very good bang per buck. Bonus points for Greencloud for their exceptionally good and incredibly fast support.
@jsg don't forget about those beasts (with Double RAM + Bandwidth or Double CPU + Bandwidth for triennial payments) https://greencloudvps.com/billing/store/premium-kvm-sale
In the Netherlands they give full /64 instead of /112 IPv6.
NL Ryzen & NVMe RAID-1 Hard drive
DE E5v4 & SSD RAID-10 Hard drive
Florida E5v4 & SSD RAID-10 Hard drive
My GreenCloud VPS Ryzen results (superfast IMO):
Location, we need location!
(my guess is San Jose considering speed test)
Correct!
So it seems this thread became GreenCloud YABS showcase?
Well, then there's also my YABS!
GreenCloud 6666NL ($66/2year, limited stock on black friday sale)
and orderlink!
@jsg "Now, on to Webhorizon - This one is a "Ryzen beast system" and costs $25/yr." - link?
It was probably a mistake for us to post YABS, as this is a review thread
greencloud is rocksolid!
but avoro is atm the best for me.
I found it in a current Webhorizon / @Abd thread here.
Thank you for the review and benchmark @jsg. We do have other Premium plans with EPYC Milan & Ryzen CPUs in APAC/EU/USA. Let me know if you need access to VMs on busy nodes for testing.
In the past, they were tunneling with the free IPv4s from ovh. Now that ovh doesn't offer free IPv4s anymore, they are probably doing the same thing but with their own subnets. It might explain the weird connectivity here.
No, it's really not. His last paragraph does not make it a review.
is there any alternative to Nexusbyte email hosting? for the same price range
AFAIK all these are MXroute resellers, the same as NB/Smallweb
https://my.onepoundemail.co.uk/order/main/packages/email-hosting/?group_id=2
https://my.webhorizon.net/order/main/packages/email/?group_id=11
https://yuzhen.cloud/order.php?step=1&productGroup=8
And again GreenCloudVPS / @NDTN amazed me, this time by providing me with access to five more locations (in addition to the E5v4 VPS I purchased in UK), so that now I can offer benchmarks in Asia, Europe, and NA, each for a Ryzen and a E5v4 VPS. The Ryzens are all equally equipped and so are the E5v4. Even better, I was provided both types in 2 locations in each region, so we can compare their connectivity (and there are some differences).
I'll start with the Ryzens, all of which have 2 vCores, 4 GB memory, and 35 GB NVMe. First the one in Montana ("MO" is Montana, right?):
All benchmarks in all locations are based on over 100 runs (avg: ca. 120 runs or so)
Here's the system and processor info:
Now the processor/memory benchmark results
Yep, that's clearly Ryzen territory and with decent speed RAM. Nice.
Now, the disk
Pardon me? I think Greencloud are engaged in a conspiracy against me; just recently I lauded the other providers NVMe as screaming fast and now Greencloud smirks and lets me know "screaming fast? We'll show you what's screaming fast". OK, guys, you got me, I'll admit that this NVMe in your Ryzen system is the new definition of "not from this planet screaming fast". Congrats, your NVMe is very convincing.
Let's have a look at connectivity (note that I added some targets for this benchmark)
First it should be noted that except for otenet.gr (which is known to be weird and problematic) there have been no or almost no failures. Not surprisingly the US targets are all quite nice (with NYC ca. 375 Mb/s), but the major european targets have respectable results too (about 100 Mb/s) and the asian targets are OK too (e.g. HK and SG ca. 50 Mb/s) and even the african targets achieve usable connectivity. All in all nice.
Now on to the Netherlands (in or not far from AMS I guess) Ryzen VPS
Sysinfo
Note that this is a Ryzen with more cores.
Processor/memory results
Mee, a bit lower than the first system, but then this node has 50% more cores, so of course it's slower (per vCore). But those results still are very decent.
Now the disk
Oops there you see the difference between a (supposedly not fully occupied) 16 core node and a (supposedly fully occupied) 24 core node. Not bad, not bad at all, but by far not as excellent as the first VPS we looked at.
On to the connectivity
First a grin: San Jose and some DC in Iraq pretty much the same results, haha.
Then with a tear in the eye: here you see why I_m switching away from the @terrahost, NO OSL target; sad because I really like terrahost a lot
Note the NL AMS target with almost 5 Gb/s and even the worst results was close to 1 Gb/s. Quite good results generally all over Europe and even Moscow is about 250 Mb/s (and even the Siberia target is about 100 Mb/s). Very cool. Cross Atlantic is good too with NVC solidly over 100 (and even close to 159) Mb/s. All in all really nice.
Part 2
Now on to the HongKong Ryzen VPS
First the Sysinfo
Same processor as the Montana Ryzen VPS. Again the full load of flags. Nice
Performance:
No surprise, very good processor/memory performance.
On to the disk
I'm afraid we're more looking into occupancy rate here. I guess GreenCloudVPS uses the same (very good and very fast) NVMe in all Ryzens, no matter the location. This node seems to be fully occupied but has the faster 16-core Ryzen and hence the performance numbers are in between 16-core Ryzen with low occupancy and 24-core Ryzen with full occupancy. Anyway a very nice NVMe, nothing to complain indeed.
Finally the connectivity
Lacking experience with that region I limit myself to noting that this VPS seems to offer really decent (very good it seems to me) connectivity to the whole region from India to Japan and Australia.
Now, on to the 3 Xeon E5 v4 systems. First my own one in the UK.
Sysinfo (see above in the OP)
Processor and memory
Somewhat mee and 100+ benchmark runs later I still stick to "meee, that v2 territory" but with 2 big BUTs, (a) that thingy is dirt cheap, I payed just $40 for three years, and (b) if I look as if it was a normal, not ultra-cheap E5v4 with just 2 vCores I'm happy again except for the crypto results.
On to the disk
Hmm, I'm actually amazed by the SSD performance. Hey, I've seen NVMes that were slower. Considering the price tag that is a really nice SSD disk. Yay.
Finally connectivity
I still find it somewhat weird that the main european targets (granted, all in the north-western part) show very nice results > 500 Mb/s bit e.g. Milano less than half, hmmm.
Anyway, pretty much all european targets show at least 100 Mb/s (and usually more) and the NA east coast about 100 Mb/s but again with a riddle; usually I see WDC being a bit slower than NYC but here WDC shows just about half the speed of NYC (even Dallas is significantly faster than WDC). Strange. On the other hand the speeds to most asian and even african targets isn't bad at all.
I'm very Europe-centric so for me that connectivity is totally OK, even quite good, especially when considering the price. But YMMV.
Part 3
Now let's look at the NYC E5v4 VPS.
Sysinfo
and processor/memory results
Yikes, why do the get a much faster E5v4 VPS then we get in the UK?! Or in other words: nice! That's more like what I expect from an e5v4 system.
disk
Pfff, they get a faster SSD too. In fact, that disk is in NVMe territory. And me idiot just said that I'm happy with my UK SSD. Injustice! Where is my hot chocolate, sniff.
Let's look at the connectivity
Hehe, finally I found something to console me; my UK VPS clearly has better connectivity. Pfff, that thing doesn't even reach 1 Gb/s within its own city.
But, oh well, even with not great connectivity the fast processor and memory as well as the NVMe territory SSD clearly make this a good deal anyway for the low price.
Finally on to the last VPS, a E5v4 in Singapore (damn they really have the tasty asian locations)
Sysinfo
and processor/memory results
Take that New Yorkers! This processor and memory is even faster than yours. Really nice numbers for the price.
Disk
See, I told you that the Greencloud guys run a conspiracy against me! And again, hardly after I praised an excellent result they grin and make the next VPS (this one) even much better. Grrrr. I call foul, Greencloud! It's not fair to sell really cheap VPSs with SSDs that are better than more than a few NVMes I've seen.
Let's look at connectivity
I want to note a miracle. AFAIK this is the first time that I've seen any VPS achieve > 100 Mb/s to Ozziland. That is the remarkable point here, not the fact that both India and Hongkong are about (or even above) 200 Mb/s. As for the rest I won't comment, again, as I don't know that region well but my gut tells me that the Greencloudvps people do know the region very well, obviously.
Summary / TL;DR
It was a joy to deal with GreenCloudVPS / @NDTN. Just like with their support everything went very smooth, professional, and quick. I left it to them and just asked for a Ryzen and a E5v4 VPS in Asia, EU, and NA, if possible. And possible it was and they chose well. Not only shiny pearls but bog standard VPS and in 6 locations.
The (premium) Ryzen systems are, well, bloody fast Ryzens, and come with fast NVMe and, somewhat depending on location (duh) with decent to very decent connectivity.
The E5v4 (budget) systems come with a very low price, actually really fast SSDs, and very good connectivity for a budget system.
I won't talk around the point. The point was that I originally purchased a VPS to replace a NexusBytes (on their dying bed, it seems) VPS that is, I wanted a really decent VPS for roughly a NB price. And it seems I've found what I was looking for. I not only recommend GreenCloudVPS but are on my way to become fan.
What convinced me the most? 3 points, basically
Bonus points: They actually listen to customers. When I asked for a FreeBSD ISO they took care of it really quickly. And when I noticed a small weak point early on during benchmarking the improved it within an hour or so. Incredible, those guys.
Spoiler tags should be used on his paragraphs (ignore) and code tags for the results (show results properly). Using spoiler tag for the actual part people want to look at makes zero sense.