Testing a performance-oriented route VPS I got, to see how it performs 
1. Configuration
| Plan | 🇭🇰 GoMAMi Turin.Mini |
|---|---|
| Location | China, Hong Kong |
| CPU | 2 x AMD EPYC 9575F vCPU |
| RAM | 4GB |
| Disk | 100GB |
| Traffic | 1000GB(one-way) |
| Speed | 2Gbps |
| Network | 1 x IPv4 |
| Route | premium return routes for all three Chinese networks |
| Price | Monthly USD$55.20 / Quarterly USD$155.25 / Yearly USD$579.60 |
2. NodeQuality
💻 Basic Information
Performance is extremely strong. Compared with the 7763 I have, the 9575F is 123.8% higher in GB5 single-core (1101) and 130% higher in multi-core (2116)
Basic System Information:
---------------------------------
Uptime : 0 days, 18 hours, 3 minutes
Processor : AMD EPYC 9575F 64-Core Processor
CPU cores : 2 @ 3299.996 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 3.8 GiB
Swap : 0.0 KiB
Disk : 101.0 GiB
Distro : Debian GNU/Linux 12 (bookworm)
Kernel : 6.12.43+deb13-cloud-amd64
VM Type : STANDARD PC (I440FX + PIIX, 1996)
IPv4/IPv6 : ✔ Online / ❌ Offline
IPv4 Network Information:
---------------------------------
ISP : Spare and Care Engineers Private Limited
ASN : AS132729 SPARE AND CARE ENGINEERS PRIVATE LIMITED
Host : Hong Kong Broadband Network Limited
Location : Ha Kwai Chung, Kwai Tsing (NKT)
Country : Hong Kong
fio Disk Speed Tests (Mixed R/W 50/50) (Partition -):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 609.54 MB/s (152.3k) | 5.32 GB/s (83.1k)
Write | 611.15 MB/s (152.7k) | 5.35 GB/s (83.5k)
Total | 1.22 GB/s (305.1k) | 10.67 GB/s (166.7k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 6.11 GB/s (11.9k) | 6.61 GB/s (6.4k)
Write | 6.44 GB/s (12.5k) | 7.05 GB/s (6.8k)
Total | 12.55 GB/s (24.5k) | 13.66 GB/s (13.3k)
Geekbench 5 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 2465
Multi Core | 4870
Full Test | https://browser.geekbench.com/v5/cpu/23863433
Geekbench 6 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 3202
Multi Core | 5911
Full Test | https://browser.geekbench.com/v6/cpu/14639521
SysBench CPU Test (Fast Mode, 1-Pass @ 5sec)
---------------------------------
1 thread test (single-core) score: 6229 Scores
2 thread test (multi-core) score: 12438 Scores
SysBench Memory Test (Fast Mode, 1-Pass @ 5sec)
---------------------------------
Single-thread read test: 96293.13 MB/s
Single-thread write test: 37029.43 MB/s
Geekbench 6 results are provided by YABS
curl -sL https://yabs.sh | bash -s -- -i -f -6
🎬 IP Quality
Announced IP; geolocation, unlocking, and risk score are all good
🌐 Network Quality
📍 Return Routes
Premium return routes for all three Chinese networks
| China Telecom | China Unicom | China Mobile |
|---|---|---|
| CN2GIA | 10099->9929 | CMIN2 |
3. Outbound Routes (itdog.cn)
Outbound uses direct backbone routes
| China Telecom | China Unicom | China Mobile |
|---|---|---|
| 163 | 4837->10099 | CMI |
| Location | China Telecom | China Unicom | China Mobile |
|---|---|---|---|
| Beijing | |||
| Shanghai | |||
| Guangzhou |
4. TCP-PING (itdog.cn, 2025-10-26 21:00)
5. Streaming Unlock
bash <(curl -L -s check.unlock.media)
6. Major website & DNS latency tests
bash <(wget -qO- https://raw.githubusercontent.com/Cd1s/network-latency-tester/main/latency.sh)
- Website test
| Ping | Real connection |
|---|---|
- DNS test
| Ping | DNS resolution |
|---|---|
7. Proxy Speed Test (2025-10-24 22:00-23:00)
Tested during Friday evening peak
- Single-thread
| Node | Average speed(Mbps) | Maximum speed(Mbps) | TLS RTT(ms) |
|---|---|---|---|
| Shenzhen Telecom@2Gbps (Haitun) | 1256 | 1454 | 11 |
| Shanghai Telecom@1Gbps (Impart) | 639 | 700 | 53 |
| Foshan Unicom@2.3Gbps (Haitun) | 1358 | 1511 | 13 |
| Guangdong Unicom@2.5Gbps (Impart) | 1190 | 1420 | 27 |
| Shanghai Unicom9929@2.5Gbps(Impart) | 34 | 98 | 51 |
| Shanghai Unicom@1Gbps (Impart) | 589 | 626 | 52 |
| Beijing Unicom9929@2.5Gbps (Impart) | 122 | 164 | 86 |
| Hangzhou Mobile@2Gbps (Haitun) | 743 | 806 | 61 |
| ShanghaiVolcengine Cloud@2Gbps (Impart) | 758 | 808 | 44 |
| Beijing Education network@1Gbps (Impart) | 375 | 391 | 67 |
| HangzhouIX@20Gbps (Impart) | 273 | 324 | 49 |
- Multi-thread
| Node | Average speed(Mbps) | Maximum speed(Mbps) | TLS RTT(ms) |
|---|---|---|---|
| Shenzhen Telecom@2Gbps (Haitun) | 1666 | 1958 | 88 |
| Shanghai Telecom@1Gbps (Impart) | 914 | 943 | 48 |
| Foshan Unicom@2.3Gbps (Haitun) | 1840 | 1910 | 13 |
| Guangdong Unicom@2.5Gbps (Impart) | 1890 | 1940 | 26 |
| Shanghai Unicom9929@2.5Gbps (Impart) | 1180 | 1340 | 53 |
| Shanghai Unicom@1Gbps (Impart) | 894 | 926 | 88 |
| Beijing Unicom9929@2.5Gbps (Impart) | 975 | 1210 | 55 |
| Hangzhou Mobile@2Gbps (Haitun) | 1793 | 1942 | 66 |
| ShanghaiVolcengine Cloud@2Gbps (Impart) | 1360 | 1900 | 50 |
| Beijing Education network@1Gbps (Impart) | 843 | 930 | 75 |
| HangzhouIX@20Gbps (Impart) | 377 | 1090 | 49 |
8. Speedtest multi-thread speed test
curl -sL nws.sh | bash
---------------------------------- nws.sh ---------------------------------
A simple script to bench network performance using speedtest-cli
---------------------------------------------------------------------------
Version : v2025.10.10
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 0.99 ms 0.0% 1925.44 Mbps 1999.38 Mbps FPT Telecom - Hong Kong
Bangalore, IN 70.09 ms 0.0% 1911.88 Mbps 1127.23 Mbps Bharti Airtel Ltd - Bangalore
Chennai, IN 65.22 ms 0.0% 1860.61 Mbps 1259.45 Mbps RailTel Corporation of India Ltd - Chennai
Mumbai, IN 186.50 ms 0.0% 1815.84 Mbps 407.39 Mbps Melbicom - Mumbai
Seattle, US 146.01 ms N/A 1068.41 Mbps 561.92 Mbps Comcast - Seattle, WA
Los Angeles, US 143.97 ms 0.0% 1282.17 Mbps 386.23 Mbps ReliableSite Hosting - Los Angeles, CA
Dallas, US 171.02 ms 0.0% 1256.13 Mbps 461.76 Mbps Hivelocity - Dallas, TX
Miami, US 210.29 ms 0.0% 1109.67 Mbps 394.56 Mbps Frontier - Miami, FL
New York, US 181.92 ms 0.0% 1781.66 Mbps 462.35 Mbps GSL Networks - New York, NY
Toronto, CA 234.13 ms 0.0% 1123.34 Mbps 347.86 Mbps Rogers - Toronto, ON
Mexico City, MX 401.67 ms N/A 478.93 Mbps 220.67 Mbps INFINITUM - Ciudad de México
London, UK 179.40 ms 0.0% 1778.03 Mbps 455.33 Mbps VeloxServ Communications - London
Amsterdam, NL 183.22 ms 0.0% 1816.99 Mbps 457.41 Mbps 31173 Services AB - Amsterdam
Paris, FR 193.53 ms 0.0% 1840.04 Mbps 438.55 Mbps Scaleway - Paris
Frankfurt, DE 204.07 ms 0.0% 879.17 Mbps 411.46 Mbps Clouvider Ltd - Frankfurt am Main
Warsaw, PL 247.82 ms 0.0% 1522.26 Mbps 335.50 Mbps Play - Warszawa
Bucharest, RO 203.77 ms 0.0% 10.36 Mbps 4.56 Mbps Vodafone Romania Mobile - Bucharest - Bucharest
Moscow, RU 167.38 ms 0.0% 1659.13 Mbps 503.12 Mbps Misaka Network, Inc. - Moscow
Jeddah, SA 222.41 ms 0.0% 1661.75 Mbps 374.16 Mbps Saudi Telecom Company
Dubai, AE 112.19 ms 0.0% 1958.12 Mbps 406.93 Mbps e& UAE - Dubai
Istanbul, TR 298.64 ms 0.0% 1476.88 Mbps 263.57 Mbps Turkcell - Istanbul
Tehran, IR 529.39 ms 0.0% 678.06 Mbps 260.56 Mbps Irancell - Tehran
Cairo, EG 212.74 ms 0.0% 1292.60 Mbps 356.62 Mbps Telecom Egypt - Cairo
Tokyo, JP 46.17 ms 0.7% 1852.23 Mbps 1811.73 Mbps fdcservers.net - Tokyo
Shanghai, CU-CN FAILED
Hong Kong, CN 0.56 ms 0.0% 1910.52 Mbps 1778.07 Mbps Misaka Network, Inc. - Hong Kong
Singapore, SG 39.97 ms 0.0% 1910.33 Mbps 2106.74 Mbps ViewQwest - Singapore
Jakarta, ID 51.67 ms 0.0% 1909.46 Mbps 1609.19 Mbps PT Solnet Indonesia - Jakarta
Sydney, AU 122.04 ms 0.0% 1931.40 Mbps 690.32 Mbps Aussie Broadband - Sydney
---------------------------------------------------------------------------
Avg DL Speed : 1489.34 Mbps
Avg UL Speed : 710.45 Mbps
Total DL Data : 58.21 GB
Total UL Data : 27.75 GB
Total Data : 85.96 GB
---------------------------------------------------------------------------
Duration : 14 min 27 sec
System Time : 24/10/2025 - 20:29:14 CST
Total Script Runs : 411747
---------------------------------------------------------------------------
Result : https://result.nws.sh/r/1761308116_94RE5V_GLOBAL.txt
---------------------------------------------------------------------------
9. iPerf3 TCP single-thread speed test (single 30-second test, 2025-10-26 14:00-15:00)
Full speed within Hong Kong, and international connectivity is also good
- 🇭🇰 Local
| Peer | Download from peer(Mbps) | Retransmits | Upload to peer(Mbps) | Retransmits | Average latency(ms) |
|---|---|---|---|---|---|
| 🇭🇰 Hong Kong DMIT T1 @4Gbps | 1880 | 3605 | 2060 | 588690 | 0.479 |
| 🇭🇰 Hong Kong RFC T1 JINX | 1870 | 0 | 2070 | 890643 | 0.442 |
| 🇭🇰 Hong Kong RFC T1 CTC | 1910 | 663 | 2070 | 904682 | 0.342 |
| 🇭🇰 Hong Kong GreenCloud DC1 (xTom) @10Gbps | 1840 | 7080 | 2070 | 818638 | 0.527 |
| 🇭🇰 Hong Kong GreenCloud DC2 (Leaseweb) @10Gbps | 1850 | 5986 | 2070 | 836274 | 0.397 |
| 🇭🇰 Hong Kong NNC (Akari) | 1840 | 3424 | 2060 | 608147 | 0.519 |
| 🇭🇰 Hong Kong BAGE HKS (CTC) @1Gbps | 1000 | 219015 | 988 | 0 | 0.440 |
| 🇭🇰 Hong Kong BAGE HKL (Hytron) @1Gbps | 1030 | 511135 | 916 | 396237 | 0.578 |
| 🇭🇰 Hong Kong Claw Intl @1Gbps | 1030 | 31075 | 1000 | 230934 | 2.284 |
| 🇭🇰 Hong Kong Servps (Zenlayer) @1Gbps | 1030 | 234797 | 914 | 0 | 1.051 |
| 🇭🇰 Hong Kong Kaze HKBN @2.5Gbps | 1860 | 17293 | 2060 | 161716 | 3.420 |
| 🇭🇰 Hong Kong Boil HKT @1Gbps | 819 | 3815 | 863 | 67 | 2.886 |
- 🌏 Asia-Pacific
| Peer | Download from peer(Mbps) | Retransmits | Upload to peer(Mbps) | Retransmits | Average latency(ms) |
|---|---|---|---|---|---|
| 🇹🇼 Taipei Misaka | 1120 | 1 | 779 | 415 | 19.834 |
| 🇹🇼 Taipei NNC (Akari) | 1370 | 90 | 1380 | 0 | 16.134 |
| 🇹🇼 Taichung Boil HiNet @777Mbps | 622 | 254 | 418 | 0 | 24.384 |
| 🇯🇵 Tokyo DMIT T1 @4Gbps | 481 | 0 | 490 | 0 | 45.687 |
| 🇯🇵 Tokyo RFC JP1T1 | 496 | 0 | 492 | 0 | 45.130 |
| 🇯🇵 Tokyo BAGE JP (GSL) @5Gbps | 480 | 8 | 485 | 0 | 45.635 |
| 🇯🇵 Tokyo BAGE JPS (CTC) @5Gbps | 692 | 0 | 477 | 0 | 46.012 |
| 🇯🇵 Tokyo GreenCloud Softbank (xTom) @10Gbps | 474 | 0 | 476 | 0 | 46.168 |
| 🇯🇵 Tokyo GreenCloud IIJ (xTom) @10Gbps | 467 | 450 | 432 | 13223 | 46.563 |
| 🇯🇵 Tokyo DDPS (CDN77) @10Gbps | 474 | 0 | 481 | 0 | 46.477 |
| 🇯🇵 Tokyo AWS Lightsail @1Gbps | 357 | 3 | 362 | 7 | 49.427 |
| 🇸🇬 Singapore RFC T1 | 597 | 0 | 731 | 0 | 30.235 |
| 🇸🇬 Singapore BAGE SG (GSL) @5Gbps | 433 | 4000 | 595 | 28 | 37.246 |
| 🇸🇬 Singapore BAGE SGS (CTC) @5Gbps | 729 | 4 | 726 | 0 | 30.418 |
| 🇸🇬 Singapore LegendVPS (CTC) @1Gbps | 612 | 0 | 634 | 446 | 30.176 |
| 🇸🇬 Singapore IDC.Care (Singtel) @1Gbps | 594 | 0 | 591 | 0 | 37.502 |
| 🇸🇬 Singapore AWS Lightsail @1Gbps | 452 | 0 | 391 | 0 | 45.467 |
- 🌎 International
| Peer | Download from peer(Mbps) | Retransmits | Upload to peer(Mbps) | Retransmits | Average latency(ms) |
|---|---|---|---|---|---|
| 🇺🇸 Los Angeles DMIT T1 @4Gbps | 146 | 0 | 144 | 0 | 143.172 |
| 🇺🇸 Los Angeles RFC T1 | 118 | 0 | 90.2 | 0 | 143.816 |
| 🇺🇸 Los Angeles BAGE LA1 (GSL) @1Gbps | 144 | 0 | 139 | 895 | 143.356 |
| 🇺🇸 Los Angeles BAGE LA2 (FiberState) @1Gbps | 139 | 0 | 134 | 608 | 148.038 |
| 🇺🇸 Los Angeles BAGE LA3 (WebNX) @1Gbps | 143 | 0 | 144 | 0 | 143.546 |
| 🇺🇸 Los Angeles ZGO (Kurun) @1Gbps | 109 | 50 | 99.4 | 2184 | 144.867 |
| 🇺🇸 San Francisco HostDZire (Leaseweb) @10Gbps | 136 | 0 | 137 | 0 | 152.796 |
| 🇺🇸 Salt Lake City BAGE (FiberState) @1Gbps | 79.5 | 17 | 70.8 | 3 | 245.287 |
| 🇩🇪 Frankfurt BAGE (dataforest) @5Gbps | 109 | 0 | 108 | 0 | 185.902 |
| 🇩🇪 Nuremberg Netcup @1Gbps | 149 | 0 | 112 | 13 | 179.145 |
| 🇬🇧 London Yunyoo (Scaleblade) @3Gbps | 76 | 0 | 56.5 | 0 | 234.847 |
| 🇬🇧 Coventry BAGE (Pebble) @1Gbps | 96.1 | 0 | 86.9 | 1278 | 193.644 |
| 🇮🇳 Mumbai HostDZire @10Gbps | 91.3 | 360 | 93.2 | 0 | 213.429 |
*During the first HiNet test, I found severe throttling and drops (limited to 8Mbps). I immediately retested other Hong Kong machines and did not see the issue; five minutes later I could not reproduce it. Cause unknown.
10. Summary
A 2Gbps port on an optimized route is very nice. Performance is extremely strong, suitable for websites or upgrading to host game servers. There are no obvious weaknesses. Pricing is high, and if used purely as a proxy, the performance is serious overkill 
11. About The Tests
The "GoMAMi Turin Mini" in this article was provided by the merchant for testing. These results do not guarantee retail-machine performance. All other machines participating in the tests (including all iPerf3 peers) were purchased at my own expense; TCP tuning was limited to enabling BBR+FQ, with all other parameters left at defaults; results may be time-sensitive
12. Acknowledgements
Thanks to GoMAMi for providing the "GoMAMi Turin Mini" test machine; thanks to "Impart Speedtest"(@Kyoma) and "Haitun Speedtest" for providing public-benefit speed testing services