← All VM Families
C3
Compute OptimizedHighest per-core performance on Google Cloud. Built for compute-intensive workloads that need maximum single-thread speed and sustained clock frequencies.
Specifications
vCPUs
4-176
Memory
8-1408 GB
Network
Up to 100 Gbps
Synthetic Benchmarks
| Category | Metric | Result | Source |
|---|---|---|---|
| cpu | Geekbench 6 Single-Core | 1,850 points | Geekbench Browser |
| cpu | Geekbench 6 Multi-Core | 14,200 points | Geekbench Browser |
| memory | STREAM Bandwidth | 45.8 GB/s | GCP Performance Benchmarks |
| memory | Memory Latency | 62 ns | GCP Performance Benchmarks |
| network | Max Bandwidth | 100 Gbps | GCP Documentation |
| storage | Local SSD IOPS (Read) | 900,000 IOPS | GCP Documentation |
| storage | Local SSD IOPS (Write) | 450,000 IOPS | GCP Documentation |
| storage | Local SSD Latency (Read) | 55 μs | GCP Documentation |
| storage | Local SSD Throughput (Read) | 7.2 GB/s | GCP Documentation |
Workload Performance
Workload Performance Summary
Best result per workload category (mixed units -- see table below for details)
| Workload | Metric | Machine Type | Result | Notes |
|---|---|---|---|---|
| Web Serving | Nginx Requests/sec | c3-standard-8 | 82,000 req/s | Static file serving, 100 concurrent connections, wrk benchmark |
| Web Serving | Node.js API Requests/sec | c3-standard-8 | 24,200 req/s | Express.js JSON API, 50 concurrent connections |
| Database | PostgreSQL TPS | c3-standard-8 | 7,400 TPS | pgbench TPC-B, 32 clients, local SSD |
| Database | Redis Operations/sec | c3-standard-8 | 420,000 ops/s | redis-benchmark, 50 parallel connections, GET/SET mix |
| ML Inference | Llama 2 7B Tokens/sec | c3-standard-44 | 3.2 tokens/s | Llama 2 7B, CPU-only inference with llama.cpp, Q4 quantized |
| Video Encoding | FFmpeg H.265 Encoding | c3-standard-8 | 34 fps | 1080p source, libx265 medium preset, 8 threads |
| HPC | LINPACK GFLOPS | c3-standard-44 | 680 GFLOPS | HPL benchmark, 44 vCPUs, Intel MKL optimized |
| HPC | OpenFOAM Iterations/sec | c3-standard-44 | 6.8 iter/s | motorbike tutorial, 44 cores, OpenMPI |
Pricing
On-demand pricing in us-central1. Spot and committed-use discounts shown for comparison.
| Machine Type | vCPUs | Memory | On-Demand/hr | Spot/hr | 1yr CUD/hr | 3yr CUD/hr |
|---|---|---|---|---|---|---|
| c3-standard-4 | 4 | 16 GB | $0.209 | $0.063 | $0.132 | $0.094 |
| c3-standard-8 | 8 | 32 GB | $0.418 | $0.125 | $0.263 | $0.188 |
| c3-standard-22 | 22 | 88 GB | $1.150 | $0.345 | $0.724 | $0.518 |
| c3-standard-44 | 44 | 176 GB | $2.300 | $0.690 | $1.449 | $1.035 |