Verified
16 July 2026
Vultr·2 vCPU · 8 GB RAM·United States
Overall
71.2
/ 100
Scores
0–100 vs comparable servers
Ping
89.5/100
Speed
81.2/100
CPU
80.2/100
Memory
96.9/100
Disk
21.1/100
Category scores
Internet speed
Ping to control points
CPU performance
Memory bandwidth
Disk — random IOPS
Disk — sequential throughput
Disk latency (p95 / p99)
Network
| download_speed | 1,800.7 | Mbps |
| upload_speed | 306 | Mbps |
| ping_cloudflare | 3.266 | ms |
| ping_google | 1.786 | ms |
| ping_quad9 | 1.811 | ms |
| mtu | 1,500 |
Disk
| random_4k_read_iops | 15,070 | IOPS |
| random_4k_read_bw | 58.9 | MB/s |
| random_4k_read_lat_p95 | 2.507 | ms |
| random_4k_read_lat_p99 | 2.638 | ms |
| random_4k_write_iops | 15,075 | IOPS |
| random_4k_write_bw | 58.9 | MB/s |
| random_4k_write_lat_p95 | 3.162 | ms |
| random_4k_write_lat_p99 | 4.047 | ms |
| sequential_1m_read_iops | 532 | IOPS |
| sequential_1m_read_bw | 532 | MB/s |
| sequential_1m_read_lat_p95 | 25.821 | ms |
| sequential_1m_read_lat_p99 | 50.07 | ms |
| sequential_1m_write_iops | 384 | IOPS |
| sequential_1m_write_bw | 384.4 | MB/s |
| sequential_1m_write_lat_p95 | 31.588 | ms |
| sequential_1m_write_lat_p99 | 33.817 | ms |
CPU
| sysbench_single | 2,075.27 | events/s |
| sysbench_multi | 2,079.42 | events/s |
| gzip_compress | 217.5 | MB/s |
Memory
| memory_write_mb_s | 26,279.22 | MiB/s |
| memory_read_mb_s | 73,600.37 | MiB/s |
| OS | Ubuntu 22.04.5 LTS |
| Kernel | 5.15.0-185-generic |
| Virtualization | kvm |
| CPU model | AMD EPYC-Turin Processor |
| vCPU | 2 |
| RAM (MB) | 7931 |
| Disk (GB) | 8.82 |
| Agent | 1.1.0 |
| Scoring | 2026.07b |
Exactly what the agent returned from this server — before rounding, with nothing edited out. Every number above is computed from this. Open separately →
Loading…
Could not load the data.
Plain-language summary
Tested on 16 July 2026 — Vultr 2 vCPU · 8 GB RAM, hosted in Piscataway, United States. Performance leans hardest on memory — 73,600 MiB/s, 97/100. storage is the soft side here — 15,070 IOPS (4K read) — worth keeping in mind. A balanced profile like this works well for typical web workloads. Ratings are computed from measurements only — no editorial picks, no paid placement.