Original title: [High-Q direct] YxVM downstream HKG Pulse Mini Hong Kong 163pp/4837/CMI 1Gbps at $33/month DetailedNetwork Test
I bought one myself to test; no AFF and no link, for reference only. 
| Plan | 🇭🇰 GoMami HKG Pulse Mini |
|---|---|
| Location | Hong Kong, China |
| CPU | 2 x vCPU EPYC™ 7763 |
| RAM | 4GB |
| Disk | 40GB |
| Transfer | 1000GB |
| Bandwidth | 1Gbps |
| Network | 1 x IPv4 |
| Route | 163 paid plus / 4837 / CMI |
| Price | monthly USD$ 33.12 |
💻 Basic Information
Performance is quite strong.
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2025-04-20 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Wed Aug 13 02:29:26 PM CST 2025
Basic System Information:
---------------------------------
Uptime : 0 days, 8 hours, 54 minutes
Processor : AMD EPYC 7763 64-Core Processor
CPU cores : 2 @ 2449.986 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 3.8 GiB
Swap : 0.0 KiB
Disk : 39.1 GiB
Distro : Debian GNU/Linux 13 (trixie)
Kernel : 6.12.38+deb13-cloud-amd64
VM Type : KVM
IPv4/IPv6 : ✔ Online / ❌ Offline
IPv4 Network Information:
---------------------------------
ISP : Next Hop LLC
ASN : AS36002 Next Hop LLC
Host : Telescope Communications
Location : Hong Kong, Kowloon ()
Country : Hong Kong
fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sdb1):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 329.87 MB/s (82.4k) | 1.12 GB/s (17.5k)
Write | 330.74 MB/s (82.6k) | 1.12 GB/s (17.6k)
Total | 660.62 MB/s (165.1k) | 2.25 GB/s (35.1k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 1.14 GB/s (2.2k) | 1.22 GB/s (1.1k)
Write | 1.20 GB/s (2.3k) | 1.30 GB/s (1.2k)
Total | 2.35 GB/s (4.5k) | 2.53 GB/s (2.4k)
Geekbench 4 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 5439
Multi Core | 9583
Full Test | https://browser.geekbench.com/v4/cpu/18940456
Geekbench 5 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 1161
Multi Core | 2303
Full Test | https://browser.geekbench.com/v5/cpu/23723975
Geekbench 6 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 1534
Multi Core | 2826
Full Test | https://browser.geekbench.com/v6/cpu/13295253
YABS completed in 12 min 4 sec
🌐 Network Test
- IP Quality & Network Quality
| IP Quality | Network Quality |
|---|---|
-
Three-network TCP Ping latency (itdog.cn )
-
Three-network TCP Ping return-route test (Zstatic CDN Dispatch Center )
| China Telecom | China Unicom | China Mobile |
|---|---|---|
| 4134 | 10099->4837 | 58453 |
notAs expected from a highly guaranteed route. This is the first time I have seen a route-optimized VPS with such low latency across all three carriers, for example Guangdong.
| North China | East China | Central China | South China | Northeast China | Northwest China | Southwest China |
|---|---|---|---|---|---|---|
| Beijing | Shanghai | Henan | Guangdong | Liaoning | Shaanxi | Chongqing |
| Tianjin | Jiangsu | Hubei | Guangxi | Jilin | Gansu | Sichuan |
| Hebei | Zhejiang | Hunan | Hainan | Heilongjiang | Qinghai | Guizhou |
| Shanxi | Anhui | Ningxia | Yunnan | |||
| Inner Mongolia | Fujian | Xinjiang | Tibet | |||
| Jiangxi | ||||||
| Shandong |
- Three-network outbound route test (itdog.cn )
| China Telecom | China Unicom | China Mobile |
|---|---|---|
| 4134 | 4837->10099 | 58453 |
| Beijing | Shanghai | Guangzhou | Wuhan | Chengdu |
|---|---|---|---|---|
| Mobile | Mobile | Mobile | Mobile | Mobile |
| Unicom | Unicom | Unicom | Unicom | Unicom |
| Telecom | Telecom | Telecom | Telecom | Telecom |
Mainland China IDC network test (non-residential, for reference only)
- ICMP Ping test(unit: ms, samples: 10,000, 2025-08-13 18:00-21:00)
latencyandpacket lossLatency and packet-loss performance are quite good. Cloud-provider return routes surprisingly go via CMI around Shanghai instead of 163. Liqun’s packet loss being that high seems a bit abnormal.
| Peer | Return Route | Average RTT | Minimum RTT | Maximum RTT | RTT standard deviation | Packet loss |
|---|---|---|---|---|---|---|
| GuangzhouBGP 🇨🇳 Alibaba Cloud Lightweight | CMI | 36.742 | 35.786 | 83.501 | 0.511 | 0.04% |
| GuangzhouBGP 🇨🇳 Tencent Cloud Ruichi | CMI | 36.690 | 35.585 | 40.675 | 0.115 | 0.35% |
| GuangdongMobile 🇨🇳 Yunxi Huanjing | CMI | 19.799 | 19.384 | 34.055 | 0.681 | 0.21% |
| ShenzhenTelecom 🇨🇳 YiAn Cloud | 163 | 12.392 | 11.297 | 48.427 | 1.211 | 0.47% |
| ShanghaiTelecomCN2 🇨🇳 Liqun Host | CN2 | 32.746 | 31.629 | 95.756 | 2.844 | 1.24% |
| WuhanTelecom 🇨🇳 YiAn Cloud | 163 | 42.119 | 39.222 | 183.315 | 6.787 | 1.78% |
| NingboTelecom 🇨🇳 Wuyuyun | 163 | 39.207 | 38.978 | 95.119 | 0.721 | 0.06% |
| NingboTelecom 🇨🇳 Haiou Cloud | 163 | 32.861 | 32.600 | 116.810 | 1.129 | 0.02% |
| ShandongUnicom 🇨🇳 Yunxi Huanjing | 4837 | 45.428 | 43.505 | 69.647 | 2.135 | 0.19% |
| BeijingUnicom9929 🇨🇳 Kuxueyun | 9929 | 43.532 | 42.191 | 67.782 | 1.492 | 0.21% |
- iPerf3 TCP single-thread test (single 30-second run, 2025-08-13 22:00)
single-threadSingle-thread performance is good; CMI also fills the pipe during evening peak. Ordinary CMI via Lumen, pulled from this Yunxi Huanjing Guangzhou node, usually only reaches around 500Mbps. Retransmissions are also good
| Peer | Return RouteRate | TransferTransfer | Retransmissions | Outbound speed | TransferTransfer | Retransmissions |
|---|---|---|---|---|---|---|
| GuangdongMobile 🇨🇳 Yunxi Huanjing @1Gbps | 966Mbps | 3.37GB | 3282 | 777Mbps | 2.71GB | 16248 |
| ShenzhenTelecom 🇨🇳 YiAn Cloud @1Gbps | 884Mbps | 3.09GB | 0 | 115Mbps | 0.41GB | 31987 |
| WuhanTelecom 🇨🇳 YiAn Cloud @1Gbps | 548Mbps | 1.92GB | 1 | 2.95Mbps | 0.11GB | 1563 |
| WuhanMobile 🇨🇳 Yunxi Huanjing @1Gbps | 557Mbps | 1.95GB | 1 | 489Mbps | 1.71GB | 497 |
| NingboTelecom 🇨🇳 Wuyuyun @500Mbps | 402Mbps | 1.40GB | 11627 | - | - | - |
| NingboTelecom 🇨🇳 Haiou Cloud @800Mbps | 643Mbps | 2.25GB | 151 | - | - | - |
| ShandongUnicom 🇨🇳 Yunxi Huanjing @1Gbps | 558Mbps | 1.95GB | 79 | 489Mbps | 1.71GB | 133 |
| BeijingUnicom9929 🇨🇳 Kuxueyun @1Gbps | 538Mbps | 1.88GB | 0 | 247Mbps | 0.88GB | 36752 |
internationalNetwork Test
- internationalspeed test (nws.sh)
---------------------------------- nws.sh ---------------------------------
A simple script to bench network performance using speedtest-cli
---------------------------------------------------------------------------
Version : v2025.07.08
Global Speedtest : wget -qO- nws.sh | bash
Region Speedtest : wget -qO- nws.sh | bash -s -- -r <region>
Ping & Routing : wget -qO- nws.sh | bash -s -- -rt <region>
---------------------------------------------------------------------------
Speedtest.net (Region: GLOBAL)
---------------------------------------------------------------------------
Location Latency Loss DL Speed UP Speed Server
ISP: NEXTHOP
Nearest 1.56 ms 0.0% 957.06 Mbps 999.26 Mbps SoSIM - Hong Kong
Bangalore, IN 69.49 ms 0.0% 962.82 Mbps 1108.71 Mbps Bharti Airtel Ltd - Bangalore
Chennai, IN 64.03 ms 0.0% 954.26 Mbps 1090.67 Mbps Jio - Chennai
Mumbai, IN 181.18 ms 0.0% 1019.20 Mbps 456.84 Mbps Melbicom - Mumbai
Seattle, US 161.18 ms N/A 902.96 Mbps 520.08 Mbps Comcast - Seattle, WA
Los Angeles, US 182.17 ms 0.0% 814.49 Mbps 322.36 Mbps ReliableSite Hosting - Los Angeles, CA
Dallas, US 231.69 ms 0.0% 891.49 Mbps 361.20 Mbps Hivelocity - Dallas, TX
Miami, US 284.79 ms 0.0% 665.31 Mbps 284.90 Mbps Telxius - Miami, FL
New York, US 233.91 ms 0.0% 990.73 Mbps 355.22 Mbps GSL Networks - New York, NY
Toronto, CA 218.92 ms 0.0% 787.73 Mbps 374.97 Mbps Rogers - Toronto, ON
Mexico City, MX 191.87 ms N/A 818.78 Mbps 431.21 Mbps INFINITUM - Ciudad de México
London, UK 183.50 ms 0.0% 998.48 Mbps 454.78 Mbps VeloxServ Communications - London
Amsterdam, NL 175.82 ms 0.0% 1002.34 Mbps 480.09 Mbps 31173 Services AB - Amsterdam
Paris, FR FAILED
Frankfurt, DE 184.87 ms 0.7% 646.78 Mbps 445.62 Mbps Clouvider Ltd - Frankfurt am Main
Warsaw, PL 168.76 ms 0.0% 997.58 Mbps 465.51 Mbps Play - Warszawa
Bucharest, RO 209.19 ms 0.0% 1010.47 Mbps 418.09 Mbps Vodafone Romania Mobile - Bucharest - Bucharest
Moscow, RU FAILED
Jeddah, SA 244.82 ms 0.0% 966.34 Mbps 344.03 Mbps Saudi Telecom Company
Dubai, AE 105.49 ms 0.0% 969.55 Mbps 105.17 Mbps e& UAE - Dubai
Istanbul, TR 292.60 ms 0.0% 732.95 Mbps 283.13 Mbps Turkcell - Istanbul
Tehran, IR 254.09 ms 0.0% 775.89 Mbps 307.43 Mbps Irancell - Tehran
Cairo, EG 196.53 ms 0.0% 867.09 Mbps 402.68 Mbps Telecom Egypt - Cairo
Tokyo, JP FAILED
Chengdu, CM-CN FAILED
Hong Kong, CN 1.49 ms 0.0% 956.94 Mbps 698.27 Mbps Misaka Network, Inc. - Hong Kong
Singapore, SG FAILED
Jakarta, ID 52.24 ms 0.0% 957.80 Mbps 1041.24 Mbps PT Solnet Indonesia - Jakarta
Sydney, AU 131.42 ms 0.0% 1005.65 Mbps 645.74 Mbps Aussie Broadband - Sydney
---------------------------------------------------------------------------
Avg DL Speed : 902.19 Mbps
Avg UL Speed : 516.55 Mbps
Total DL Data : 29.74 GB
Total UL Data : 16.79 GB
Total Data : 46.53 GB
---------------------------------------------------------------------------
Duration : 12 min 29 sec
System Time : 13/08/2025 - 14:56:03 CST
Total Script Runs : 120567
---------------------------------------------------------------------------
Result : https://result.nws.sh/r/1755067556_NZT6JP_GLOBAL.txt
---------------------------------------------------------------------------
- ICMP Ping test(unit: ms, samples: 10,000, 2025-08-13 18:00-21:00)
RFC CTCRFC CTC has less latency jitter than JINX. Latency to Hinet is good. To RFC T1 JP it goes through RETN-SoftBank, with low latency but a small amount of packet loss. Seattle has the lowest latency to the US
| Peer | Average RTT | Minimum RTT | Maximum RTT | RTT standard deviation | Packet loss |
|---|---|---|---|---|---|
| Hong Kong 🇭🇰 RFCHost Jinx | 1.747 | 1.534 | 127.522 | 1.456 | 0% |
| Hong Kong 🇭🇰 RFCHost CTC | 1.629 | 1.526 | 5.713 | 0.060 | 0.02% |
| Hong Kong 🇭🇰 DMIT T1 | 1.622 | 1.474 | 90.846 | 0.912 | 0% |
| Hong Kong 🇭🇰 Bage HKS | 1.582 | 1.424 | 78.737 | 1.060 | 0% |
| Hong Kong 🇭🇰 CNFaster HKL | 2.064 | 1.848 | 14.758 | 0.569 | 0% |
| Hong Kong 🇭🇰 Hytron HKBGP | 1.579 | 1.453 | 125.047 | 1.245 | 0% |
| Hong Kong 🇭🇰 Servps SP | 2.233 | 2.079 | 56.033 | 0.904 | 0.01% |
| Hong Kong 🇭🇰 Boil HKBN | 3.765 | 2.825 | 611.038 | 8.666 | 0.01% |
| Tokyo 🇯🇵 RFCHost T1 | 54.470 | 47.734 | 93.658 | 14.151 | 2.13% |
| Tokyo 🇯🇵 DDPS | 72.257 | 71.978 | 163.643 | 3.445 | 0.28% |
| Singapore 🇸🇬 RFCHost T1 | 31.275 | 31.110 | 34.614 | 0.069 | 0% |
| Singapore 🇸🇬 Legend SG-Lite | - | - | - | - | - |
| Taichung 🇹🇼 Boil Hinet | 28.523 | 27.957 | 37.839 | 0.336 | 0% |
| Los Angeles 🇺🇸 DMIT T1 | 162.828 | 162.663 | 199.381 | 0.702 | 0.01% |
| Los Angeles 🇺🇸 RFCHost T1 | 182.142 | 181.743 | 230.196 | 1.247 | 0.01% |
| Los Angeles 🇺🇸 Legend LAX | 161.697 | 161.572 | 165.849 | 0.070 | 0.02% |
| Los Angeles 🇺🇸 RackNerd DC2 | 184.359 | 177.140 | 223.313 | 9.783 | 0.55% |
| Seattle 🇺🇸 RackNerd SEA | 139.996 | 139.739 | 169.646 | 0.627 | 0.04% |
| New York 🇺🇸 Legend NYC | 228.929 | 226.761 | 294.182 | 3.745 | 1.87% |
| New York 🇺🇸 Netcup NYC | 221.416 | 221.276 | 221.982 | 0.511 | 0% |
| Nuremberg 🇩🇪 Netcup NUE | 177.060 | 176.851 | 184.701 | 0.335 | 0% |
| Strasbourg 🇫🇷 RackNerd | 181.031 | 180.697 | 190.110 | 0.409 | 0.06% |
| Mumbai 🇮🇳 HostDZire | 185.054 | 182.985 | 281.047 | 4.816 | 0.11% |
- iPerf3 TCP single-thread test (single 30-second run, 2025-08-13 23:00)
Within Hong Kong, it can easily pull various local landing nodes, but downloads from peers seem capped at 943Mbps, just short of saturating a gigabit port. US/Europe results are lower because TCP tuning was not applied; tuning is generally recommended
| Peer | Download from peer | TransferTransfer | Retransmissions | Upload to peer | TransferTransfer | Retransmissions |
|---|---|---|---|---|---|---|
| Hong Kong 🇭🇰 RFCHost Jinx | 938Mbps | 3.28GB | 0 | 1.03Gbps | 3.61GB | 204448 |
| Hong Kong 🇭🇰 RFCHost CTC | 943Mbps | 3.29GB | 0 | 1.03Gbps | 3.61GB | 261280 |
| Hong Kong 🇭🇰 DMIT T1 @4Gbps | 943Mbps | 3.30GB | 0 | 1.03 | 3.60GB | 215737 |
| Hong Kong 🇭🇰 Bage HKS @1Gbps | 933Mbps | 3.26GB | 10790 | 968Mbps | 3.38GB | 0 |
| Hong Kong 🇭🇰 CNFaster HKL @2Gbps | 943Mbps | 3.29GB | 0 | 1.03Gbps | 3.61GB | 218278 |
| Hong Kong 🇭🇰 Hytron HKBGP @1Gbps | 943Mbps | 3.30GB | 0 | 1.03Gbps | 3.61GB | 131708 |
| Hong Kong 🇭🇰 Servps SP @1Gbps | 943Mbps | 3.29GB | 517 | 925Mbps | 3.23GB | 0 |
| Hong Kong 🇭🇰 Boil HKBN @2.5Gbps | 940Mbps | 3.28GB | 0 | 1.03Gbps | 3.61GB | 201624 |
| Tokyo 🇯🇵 RFCHost T1 | 396Mbps | 1.39GB | 0 | 269Mbps | 0.97GB | 161 |
| Tokyo 🇯🇵 DDPS @10Gbps | 394Mbps | 1.38GB | 0 | 269Mbps | 0.97Gb | 0 |
| Singapore 🇸🇬 RFCHost T1 | 597Mbps | 2.09GB | 0 | 591Mbps | 2.07GB | 0 |
| Singapore 🇸🇬 Legend SG-Lite @1Gbps | 586Mbps | 2.05GB | 1934 | 435Mbps | 1.52GB | 0 |
| Taichung 🇹🇼 Boil Hinet @777Mbps | 738Mbps | 2.58GB | 482 | 713Mbps | 2.49GB | 5444 |
| Los Angeles 🇺🇸 DMIT T1 @4Gbps | 131Mbps | 0.47GB | 0 | 125Mbps | 0.45GB | 3 |
| Los Angeles 🇺🇸 RFCHost T1 | 94Mbps | 0.34GB | 0 | 73.5Mbps | 0.26GB | 0 |
| Los Angeles 🇺🇸 Legend LAX @1Gbps | 156Mbps | 0.56GB | 0 | 130Mbps | 0.47GB | 0 |
| Los Angeles 🇺🇸 RackNerd DC2 @1Gbps | 60Mbps | 0.22GB | 19 | 39Mbps | 0.14GB | 7544 |
| Seattle 🇺🇸 RackNerd SEA @1Gbps | 103Mbps | 0.37GB | 2889 | 144Mbps | 0.52GB | 132 |
| New York 🇺🇸 Legend NYC @1Gbps | 80Mbps | 0.29GB | 10 | 91Mbps | 0.33GB | 0 |
| New York 🇺🇸 Netcup NYC @2.5Gbps | 117Mbps | 0.42GB | 0 | 89Mbps | 0.32GB | 0 |
| Nuremberg 🇩🇪 Netcup NUE @2.5Gbps | 148Mbps | 0.53GB | 0 | 116Mbps | 0.42GB | 0 |
| Strasbourg 🇫🇷 RackNerd @1Gbps | 153Mbps | 0.55GB | 0 | 108Mbps | 0.39GB | 591 |
| London 🇬🇧 Upstream @1Gbps | 127Mbps | 0.46GB | 0 | 116Mbps | 0.41GB | 1 |
| Mumbai 🇮🇳 HostDZire @10Gbps | 166Mbps | 0.60GB | 0 | 108Mbps | 0.39GB | 0 |
🔗 About the Tests
All machines tested in this article were purchased at my own expense; tests only enabled BBR+FQ, with all other kernel parameters left at defaults; results may be time-sensitive.
