diff --git a/docs/administer-science/prices/examples.md b/docs/administer-science/prices/examples.md index 7ca553729f..e37a3fea79 100644 --- a/docs/administer-science/prices/examples.md +++ b/docs/administer-science/prices/examples.md @@ -24,67 +24,60 @@ Head over to our [pricing estimator](/administer-science/prices/estimator) to ca This is a typical lab tailored for data analysis of structured data, such as health registries and survey data. -| Service | Description | Type | Cost/unit | Units | Total cost | -| ---- | ---- | ---- | ----: | ----: | ----: | -| Lab | subscription | 1 year | 9354 | 1 | 9354 | -| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5360 | 1 | 5360 | -| Store | 1 TB regular storage | 1 year | 2844 | 1 | 2844 | -| **PER YEAR** | | | | | **17558** | +| Service | Description | Type | Cost/unit | Units | Total cost | +| ------------ | ----------------------------------------------------------------- | ------ | --------: | ----: | ---------: | +| Lab | subscription | 1 year | 9600 | 1 | 9600 | +| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5385 | 1 | 5385 | +| Store | 1 TB regular storage | 1 year | 2864 | 1 | 2864 | +| **PER YEAR** | | | | | **17849** | The above example establishes your own lab in HUNT Cloud with access to commonly used tools in HUNT Workbench. Your lab holds sufficient resources for analysis for a small team: 2 vCPUs and 8 GB of memory, and the minimum allocation of 1 TB storage. Resources can easily be upgraded as your experiments grow. - - ## Genomics This is an example of a lab for larger genomics analysis, for example when you need a lab to store and analyze genomics data from several sources using a varied set of tools. -| Service | Description | Type | Cost/unit | Units | Total cost | -| ---- | ---- | ---- | ----: | ----: | ----: | -| Lab | subscription | 1 year | 9354 | 1 | 9354 | -| Compute | [default.b5](/administer-science/services/machine-types#b-series) | 1 year | 41462 | 1 | 41462 | -| Store | 10 TB regular storage | 1 year | 2844 | 10 | 28440 | -| **PER YEAR** | | | | | **79256** | +| Service | Description | Type | Cost/unit | Units | Total cost | +| ------------ | ----------------------------------------------------------------- | ------ | --------: | ----: | ---------: | +| Lab | subscription | 1 year | 9600 | 1 | 9600 | +| Compute | [default.b5](/administer-science/services/machine-types#b-series) | 1 year | 41654 | 1 | 41654 | +| Store | 10 TB regular storage | 1 year | 2864 | 10 | 28640 | +| **PER YEAR** | | | | | **79894** | -The above example establishes your own lab in HUNT Cloud, a powerful machine fit for larger analysis (32 vCPU and 64 GB memory), and storage capacity to run both experiments and archive larger data sets (10 TB). +The above example establishes your own lab in HUNT Cloud, a powerful machine fit for larger analysis (32 vCPU and 64 GB memory), and storage capacity to run both experiments and archive larger data sets (10 TB). The cost can be reduced by running a smaller machine size for day-to-day activities and then upgrading to larger on-demand machines in analysis intensive periods. - - ## Machine learning -This is an example of a machine learning lab with mixed activities that require data storage, computational power and graphical accelerators. +This is an example of a machine learning lab with mixed activities that require data storage, computational power and graphical accelerators. -| Service | Description | Type | Cost/unit | Units | Total cost | -| ---- | ---- | ---- | ----: | ----: | ----: | -| Lab | subscription | 1 year | 9354 | 1 | 9354 | -| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5360 | 1 | 5360 | -| Compute | [default.c4](/administer-science/services/machine-types#c-series) | 1 year | 32986 | 1 | 32986 | -| Compute | [nvidia.l40s](/administer-science/services/machine-types#gpu-accelerator-types) | 1 year | 23425 | 1 | 23425 | -| Store | 10 TB regular storage | 1 year | 2844 | 10 | 28440 | -| Store | 2 TB speed-optimized storage | 1 year | 5524 | 2 | 11048 | -| **PER YEAR** | | | | | **110613** | +| Service | Description | Type | Cost/unit | Units | Total cost | +| ------------ | ------------------------------------------------------------------------------- | ------ | --------: | ----: | ---------: | +| Lab | subscription | 1 year | 9600 | 1 | 9600 | +| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5385 | 1 | 5385 | +| Compute | [default.c4](/administer-science/services/machine-types#c-series) | 1 year | 33138 | 1 | 33138 | +| Compute | [nvidia.l40s](/administer-science/services/machine-types#gpu-accelerator-types) | 1 year | 23437 | 1 | 23437 | +| Store | 10 TB regular storage | 1 year | 2864 | 10 | 28640 | +| Store | 2 TB speed-optimized storage | 1 year | 5559 | 2 | 11118 | +| **PER YEAR** | | | | | **111318** | The above example establishes your own lab in HUNT Cloud. The compute resources are split in two machines: a smaller home machine (default.c1 with 2 vCPU and 8 GB memory) for data handling and day-to-day activities, one of the most typical machines used for machine learning (default.c4 with 16 vCPU and 64 GB memory) that has attached one commonly used enterprise GPU card (NVIDIA L40S with 48 GB memory). The lab holds 10 TB regular storage for archive and 2 TB of speed-optimized storage (NVMe) for the GPU machine. The cost may be reduced by using on-demand compute resources when you need the GPU resources. - ## Data archives This is an example of a large data archive, such as a biobank or image archive that manages data for reuse. The configuration exemplifies a data archive holding 200 TB of data with 90 days of intensive data quality control per year. +| Service | Description | Type | Cost/unit | Units | Total cost | +| ------------ | ----------------------------------------------------------------- | ------ | --------: | ----: | ---------: | +| Lab | subscription | 1 year | 9600 | 1 | 9600 | +| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5385 | 1 | 5385 | +| Compute | [default.b4](/administer-science/services/machine-types#b-series) | 1 day | 52.49 | 90 | 4724 | +| Store | First 10 TB regular storage | 1 year | 2864 | 10 | 28640 | +| Store | Next 90 TB regular storage | 1 year | 2203 | 90 | 198270 | +| Store | Next 100 TB regular storage | 1 year | 1652 | 100 | 165200 | +| **PER YEAR** | | | | | **411819** | -| Service | Description | Type | Cost/unit | Units | Total cost | -| ---- | ---- | ---- | ----: | ----: | ----: | -| Lab | subscription | 1 year | 9354 | 1 | 9354 | -| Compute | [default.c1](/administer-science/services/machine-types#c-series) | 1 year | 5360 | 1 | 5360 | -| Compute | [default.b4](/administer-science/services/machine-types#b-series) | 1 day | 52.25 | 90 | 4702 | -| Store | First 10 TB regular storage | 1 year | 2844 | 10 | 28440 | -| Store | Next 90 TB regular storage | 1 year | 2188 | 90 | 196920 | -| Store | Next 100 TB regular storage | 1 year | 1641 | 100 | 164100 | -| **PER YEAR** | | | | | **408876** | - -The above examples gives you a lab with a home machine intended for basic data activities. On-demand compute resources is allocated for 90 days for intensive quality control activities. +The above examples gives you a lab with a home machine intended for basic data activities. On-demand compute resources is allocated for 90 days for intensive quality control activities. Storage above 10 TB obtains a volume discounts ending at a mean TB price of NOK 1947. The total cost may be significantly reduced by initiating a layered preservation strategies that utilizes tapes for duplicate copies. -