Google Compute Engine Machine Type a2-highgpu-4g
Accelerator-optimized Google Compute Engine machine type a2-highgpu-4g with 48 vCPU and 340 GB memory. Available in 8 Google Cloud regions.
✅ Instance Picker with filter on machine type a2-highgpu-4g Compare machine type a2-highgpu-4g
Google Cloud VM a2-highgpu-4g
Technical facts about the Google Compute Engine machine type a2-highgpu-4g.
| Series | A2 |
|---|---|
| Family | Accelerator-optimized |
| vCPU | 48 |
| Memory | 340 GB |
| CPU Manufactur | Intel |
| CPU Platform |
|
| CPU Base Frequency | 2.2 GHz |
| CPU Turbo Frequency | 2.9 GHz |
| CPU Max. Turbo Frequency | 3.7 GHz |
| Accelerator (GPUs) | 4 |
| Accelerator Type | nvidia-tesla-a100 |
| EEMBC CoreMark Benchmark (?) | - |
| EEMBC CoreMark Standard Deviation | - |
| EEMBC CoreMark Sample Count | - |
| SAP Standard Benchmark (?) | - |
| Network Bandwidth | 50 Gbps |
| Network Tier 1 | - |
| Max. Disk Size | 257 TB |
| Max. Number of Disks | 128 |
| Local SSD | ✔️ |
| SAP Application | ❌ |
| SAP HANA | ❌ |
| Spot Provisioning Mode (Spot VM) | ✔️ |
| Sustained Use Discount (SUD) | ❌ |
| GCE API Description | Accelerator Optimized: 4 NVIDIA Tesla A100 GPUs, 48 vCPUs, 340GB RAM |
Share a2-highgpu-4g
Share a link to the Google Cloud machine type a2-highgpu-4g.
Link to share
Costs (Pricing) for a2-highgpu-4g
Average cost and pricing across all regions for Google Cloud VM a2-highgpu-4g.
| Avg. costs per hour | 15.4077 |
|---|---|
| Avg. costs per month | 11247.63 |
| Avg. costs per month with 1 year commitment (CUD) | 7054.38 |
| Avg. costs per month with 3 year commitment (CUD) | 3944.79 |
Costs and pricing for Google Compute Engine machine type a2-highgpu-4g in Google Cloud regions in which the VM is available.
| Region | Location | #Zones | #Platf. | Hour ⬇️ | Month | 1Y CUD | 3Y CUD |
|---|---|---|---|---|---|---|---|
| us-central1 | Iowa 🍃 | 4 | 1 | 14.6935 | 10726.28 | 6757.48 | 3754.27 |
| us-east1 | South Carolina | 2 | 1 | 14.6935 | 10726.28 | 6757.48 | 3754.27 |
| us-west1 | Oregon 🍃 | 1 | 1 | 14.6935 | 10726.28 | 6757.48 | 3754.27 |
| europe-west4 | Netherlands | 2 | 1 | 14.9919 | 10944.08 | 6894.69 | 3830.51 |
| us-west4 | Las Vegas | 1 | 1 | 15.7374 | 11488.32 | 7174.47 | 4037.07 |
| asia-southeast1 | Singapore | 2 | 1 | 16.055 | 11720.17 | 7320.53 | 4118.23 |
| asia-northeast1 | Tokyo | 2 | 1 | 16.1984 | 11824.81 | 7386.45 | 4154.86 |
| asia-northeast3 | Seoul | 2 | 1 | 16.1984 | 11824.81 | 7386.45 | 4154.86 |
Operating System Licenses Costs for a2-highgpu-4g
Monthly costs for paid premium operating system licenses for Google Compute Engine machine type a2-highgpu-4g.
| SUSE Linux Enterprise Server | 80.3 |
|---|---|
| Red Hat Enterprise Linux | 94.9 |
| Microsoft Windows Server | 1611.84 |
| SUSE Linux Enterprise Server for SAP | 299.3 |
| SUSE Linux Enterprise Server for SAP w. 1 year CUD | 119.94 |
| SUSE Linux Enterprise Server for SAP w. 3 year CUD | 107.75 |
| Red Hat Enterprise Linux for SAP | 164.25 |
Machine Types with 48 vCPU
| Name ⬇️ | vCPU | Memory | Avg. costs per hour | Avg. costs per month | |||||
|---|---|---|---|---|---|---|---|---|---|
| a2-highgpu-4g | 48 | 340 | Intel | 15.4077 | 11247.63 | ||||
| a2-ultragpu-4g | 48 | +340.00 | 680 | Intel | +6.23 | 21.6346 | +4545.62 | 15793.25 | Compare |
| n2-highcpu-48 | 48 | -292.00 | 48 | Intel | -13.32 | 2.0869 | -10028.57 | 1219.06 SUD | Compare |
| n2-highmem-48 | 48 | +44.00 | 384 | Intel SAP HANA | -11.59 | 3.8131 | -9020.21 | 2227.42 SUD | Compare |
| n2-standard-48 | 48 | -148.00 | 192 | Intel SAP | -12.58 | 2.8267 | -9596.42 | 1651.21 SUD | Compare |
| n2d-highcpu-48 | 48 | -292.00 | 48 | AMD | -13.66 | 1.7444 | -10228.65 | 1018.98 SUD | Compare |
| n2d-highmem-48 | 48 | +44.00 | 384 | AMD SAP | -12.22 | 3.1872 | -9385.82 | 1861.81 SUD | Compare |
| n2d-standard-48 | 48 | -148.00 | 192 | AMD SAP | -13.04 | 2.3627 | -9867.44 | 1380.19 SUD | Compare |
| t2a-standard-48 | 48 | -148.00 | 192 | Arm | -13.35 | 2.0536 | -9748.50 | 1499.13 | Compare |
| t2d-standard-48 | 48 | -148.00 | 192 | AMD SAP | -13.01 | 2.3933 | -9500.50 | 1747.13 | Compare |