Google Cloud Compute Engine
m2-ultramemx-96
A Google Cloud Compute Engine machine type in the M2 machine series.
Google's catalog has no Spot price for m2-ultramemx-96.
On AWS
The nearest AWS machines to m2-ultramemx-96 by vCPU and memory, at On-demand Linux rates in Sydney (australia-southeast1 ↔ ap-southeast-2). m2-ultramemx-96 isn't sold in us-east4, so this compares the paired region where its rate is lowest.
| Machine | Size | On-demand Linux | vs m2-ultramemx-96 |
|---|---|---|---|
| m2-ultramemx-96This page · Google Cloud, australia-southeast1 | 96 vCPU · 600 GiB | $9.798009/hr | — |
| r5.24xlarge96 vCPU and 768 GiB for 96 vCPU and 600 GiB, the nearest size up; Intel on both | 96 vCPU · 768 GiB | $7.2480/hr | 26% less |
| r6i.24xlarge96 vCPU and 768 GiB for 96 vCPU and 600 GiB, the nearest size up; Intel on both | 96 vCPU · 768 GiB | $7.2480/hr | 26% less |
| r7i.24xlarge96 vCPU and 768 GiB for 96 vCPU and 600 GiB, the nearest size up; Intel on both | 96 vCPU · 768 GiB | $7.6104/hr | 22% less |
Matched on vCPU and memory, then CPU vendor, then rate; a larger size is offered only when nothing is the same size. On-demand Linux only: Spot and commitments work differently on the two clouds, so they aren't compared here.
How this price is built
Google bills a VM by its parts, each with its own SKU. These are the parts of the On-demand Linux rate in Iowa (us-central1). The total is the published price; the parts add up to it.
| Part | SKU | Google's rate | Quantity | Per hour |
|---|---|---|---|---|
| vCPU | Memory-optimized Instance Core running in Americas AE53-03EE-A981 | $0.0348 per hour | 96 vCPU | $3.3408 |
| Memory | Memory-optimized Instance Ram running in Americas 9DFC-84DE-A1D6 | $0.0051 per GiB-hour | 600 GiB | $3.0600 |
| vCPU premium | Memory Optimized Upgrade Premium for Memory-optimized Instance Core running in Americas 19A3-E481-8933 | $0.004524 per hour | 96 vCPU | $0.434304 |
| Memory premium | Memory Optimized Upgrade Premium for Memory-optimized Instance Ram running in Americas 646C-54DD-382D | $0.000663 per GiB-hour | 600 GiB | $0.3978 |
| On-demand, Linux | $7.232904 | |||
Windows adds the Windows Server licence: 96 vCPU × $0.046 per hour (SKU 92D2-8A8F-DF54) = $4.4160/hr, the same in every region and on every basis.
Rates by region
On-demand: Pay-as-you-go rate, composed from Google's per-vCPU and per-GiB SKUs
| Region | Linux (On-demand) | vs On-demand | Windows (On-demand) | vs On-demand |
|---|---|---|---|---|
australia-southeast1 Sydney | $9.798009 | $14.214009 | ||
southamerica-west1 HighestSantiago | $10.343053 | $14.759053 | ||
us-central1 LowestIowa | $7.232904 | $11.648904 |
No regions match the filter.
Monthly is 730 hours and annual is 8,760. Windows prices include the Windows Server licence.
What the prices include
- Regions. m2-ultramemx-96 is priced in the 3 regions where Compute Engine runs it in at least one zone. It may run in only some of a region's zones.
- Windows. The Windows Server licence costs $0.046 per visible vCPU per hour. Spot and commitments don't discount it, so it's added at the full rate on every basis.
- Commitments. The 1-yr and 3-yr rates are resource-based committed use discounts: you commit to an amount of M2 vCPU and memory in one region and pay for it whether you use it or not. Flexible (spend-based) commitments discount less and aren't shown.
- Sustained use. M2 earns a sustained-use discount of up to 30% off the On-demand rate when a VM runs the whole month. It starts once the VM runs more than a quarter of the month, arrives as a monthly credit, doesn't apply to usage a commitment covers, and only self-serve billing accounts get it. The prices here don't include it.
Other M2 machine types
| Machine type | vCPU | Memory (GiB) | Lowest On-demand, Linux |
|---|---|---|---|
| m2-ultramemx-96 | 96 | 600 | $7.232904/hr |
| m2-ultramem2x-96 | 96 | 1,250 | $10.978854/hr |
| m2-ultramem-208 | 208 | 5,888 | $42.111936/hr |
| m2-megamem-416 | 416 | 5,888 | $50.291328/hr |
| m2-hypermem-416 | 416 | 8,832 | $67.2576/hr |
| m2-ultramem-416 | 416 | 11,776 | $84.223872/hr |