Because of the large size of single cell RNA-Seq data sets and the computationally-intensive tools used in single cell analysis, we have amended our system requirements and recommendations for installations with the Single Cell toolkit. 

Up to 100,000 cells per analysis

Required

• Linux: Ubuntu^® 12.04, Redhat^® 6.10, CentOS^® 6.10, or newer

• CPU: 64-bit 2 GHz quad-core processor

• Memory: 64 GB of RAM

• Local scratch space¹: 1 TB with cached or native speeds of 2GB/s or higher

• Storage: 1 TB available for data

Recommended

• Linux: Ubuntu^® 16.04, Redhat^® 7.6, CentOS^® 7.6, or newer

• CPU: 64-bit 2 GHz quad-core processor

• Memory: 128 GB of RAM

• Local scratch space¹: 2 TB with cached or native speeds of 2GB/s or higher

• Storage: 1 TB available for data

More than 100,000 cells per analysis

Required

• Linux: Ubuntu^® 16.04, Redhat^® 7.6, CentOS^® 7.6, or newer

• CPU: 64-bit 2 GHz quad-core processor

• Memory: 256 GB of RAM

• Local scratch space¹: 2 TB with cached or native speeds of 2GB/s or higher

• Storage: 4 TB available for data

Recommended

• Linux: Ubuntu^® 16.04, Redhat^® 7.6, CentOS^® 7.6, or newer

• CPU: 64-bit 2 GHz quad-core processor

• Memory: 512 GB of RAM

• Local scratch space¹: 10 TB with cached or native speeds of 2GB/s or higher

• Storage: 10 TB available for data

For fastest performance:

• Newer generation CPU cores with avx2 or avx-512 are recommended.

• Performance scales proportionality to the number of CPU cores available.

• Hyper thread cores (threads) scales performance for most operations other than principal component analysis.

¹ Contact Partek support for recommended setup of local scratch storage