Compute Credits¶
Compute credits are a unit of measure to quantify resource consumption when running analysis tools in IDEAS. These credits take into account the total RAM, memory, and GPUs required in addition to vCPUs. Refer to the following table to note how many compute credits are used for varying computing resources. Analyses requiring more computational resources to perform efficiently will use more compute credits.
| Computing Resource | Compute Credit (hours) Used |
|---|---|
| 4 vCPUs, 8GB RAM | 0.5 |
| 8 vCPUs, 32GB RAM | 1 |
| 16 vCPUs, 64GB RAM | 2 |
| 16 vCPUs, 128GB RAM | 2 |
| 32 vCPUs, 256GB RAM | 4 |
| 1 GPU (NVIDIA T4), 4 vCPUs, 16GB RAM | 2 |
Refer to the Resource Consumption documentation for where to find how many compute credits you've used. Note that each tool described in the Analysis Tools section lists how many compute credits are used by the tool.
Migration to compute credits
Starting in IDEAS 23.6, all previous 'CPU' usages will be converted to compute credits (compute credit usage will be equal to or less than CPU usage in all cases). Subscription CPU amounts will be directly converted into compute credits 1:1 (1000 CPU hours will become 1000 compute credits).