a2-ultragpu-1g as a GKE node
12 vCPU170 GiBx86_641 GPUGoogle Cloud us-central1A2 Ultragpu 1g
Where the a2-ultragpu-1g goes
Default GKE setup: GKE ≤ 1.36.
CPU11.70 of 12 vCPU left for your pods
Memory159.45 GiB of 170 GiB left for your pods
your podsDaemonSets (200m, 300 MiB assumed)kubelet reservedeviction
110max pods per node
$832.72on demand / month · $1.14/h
$498.74spot / month (varies)
a2-ultragpu-1g GKE max pods and allocatable memory — quick answers
- How many pods can a a2-ultragpu-1g run as a node with GKE ≤ 1.36? Up to 110; kubelet keeps 10504 MiB of memory and 100m CPU, leaving 159.74 GiB and 11.90 vCPU for pods.
- How many pods can a a2-ultragpu-1g run as a node with GKE 1.37+ (Container-Optimized OS)? Up to 110; kubelet keeps 2250 MiB of memory and 100m CPU, leaving 167.80 GiB and 11.90 vCPU for pods.
- How many 500m / 1 GiB pods fit? 23 per node after DaemonSets; 1 nodes for 10 replicas, about $832.72/month on demand.
What kubelet reserves on a2-ultragpu-1g
| Setup | Max pods | kube-reserved CPU | Memory reserved + eviction | Allocatable CPU | Allocatable memory | Memory usable |
|---|---|---|---|---|---|---|
| GKE ≤ 1.36 tiered memory reservation | 110 | 100m | 10404 Mi + 100 Mi | 11.90 | 159.74 GiB | 94.0% |
| GKE 1.37+ (Container-Optimized OS) 15 MiB × max pods + 500 MiB, CPU capped at 1 vCPU | 110 | 100m | 2150 Mi + 100 Mi | 11.90 | 167.80 GiB | 98.7% |
How these numbers are computed: reservation formulas. Check a live node with kubectl describe node → Allocatable.
How many pods fit on one a2-ultragpu-1g
Assumes DaemonSets (CNI, kube-proxy, log and metrics agents) request 200m CPU, 300 Mi and 3 pods per node. Try your own numbers in the node calculator.
| Pod requests (CPU / memory) | GKE ≤ 1.36 | GKE 1.37+ (Container-Optimized OS) | $ per pod / month |
|---|---|---|---|
| 100m / 128 Mi | 107 | 107 | $7.78 |
| 250m / 256 Mi | 46 | 46 | $18.10 |
| 250m / 512 Mi | 46 | 46 | $18.10 |
| 500m / 512 Mi | 23 | 23 | $36.21 |
| 500m / 1024 Mi | 23 | 23 | $36.21 |
| 1000m / 1024 Mi | 11 | 11 | $75.70 |
| 1000m / 2048 Mi | 11 | 11 | $75.70 |
| 2000m / 4096 Mi | 5 | 5 | $166.54 |
| 4000m / 8192 Mi | 2 | 2 | $416.36 |
Similar GKE node types
| Type | vCPU | GiB | Alloc CPU | Alloc memory | Max pods | $/month | $ per alloc GiB |
|---|---|---|---|---|---|---|---|
| a2-highgpu-1g | 12 | 85 | 11.90 | 78.16 GiB | 110 | $539.82 | $6.91 |
| a2-highgpu-2g | 24 | 170 | 23.87 | 159.74 GiB | 110 | $1,079.64 | $6.76 |
| a3-highgpu-1g | 26 | 234 | 25.86 | 222.46 GiB | 110 | $923.23 | $4.15 |
| g2-standard-16 | 16 | 64 | 15.89 | 58.42 GiB | 110 | $428.63 | $7.34 |
| g2-standard-12 | 12 | 48 | 11.90 | 43.38 GiB | 110 | $321.47 | $7.41 |
| g2-standard-24 | 24 | 96 | 23.87 | 88.50 GiB | 110 | $642.95 | $7.26 |
| g2-standard-32 | 32 | 128 | 31.85 | 118.58 GiB | 110 | $857.27 | $7.23 |
| a2-ultragpu-2g | 24 | 340 | 23.87 | 326.34 GiB | 110 | $1,665.45 | $5.10 |
Closest match on other clouds
| Type | vCPU | GiB | Alloc CPU | Alloc memory | Max pods | $/month | $ per alloc GiB |
|---|---|---|---|---|---|---|---|
| gr6f.4xlarge | 16 | 128 | 15.89 | 125.14 GiB | 234 | $778.18 | $6.22 |
| gr6.4xlarge | 16 | 128 | 15.89 | 125.14 GiB | 234 | $1,123.62 | $8.98 |
| g6e.4xlarge | 16 | 128 | 15.89 | 125.14 GiB | 234 | $2,193.10 | $17.53 |
| g7e.4xlarge | 16 | 128 | 15.89 | 122.22 GiB | 506 | $2,918.66 | $23.88 |
| g3.4xlarge | 16 | 122 | 15.89 | 120.47 GiB | 110 | $832.20 | $6.91 |
| p5.4xlarge | 16 | 256 | 15.89 | 254.39 GiB | 118 | $5,022.40 | $19.74 |