Research Computing Core Policies


The Harvard Medical School Research Computing Core (HMS RC Core) provides IT services to investigators from HMS, HMS-affiliated institutions, and other Harvard Schools who wish to use HMS's infrastructure for their scientific computing needs.

The HMS RC Core provides services that include High-Performance Computing, Data Storage, and application services.  As described in the HMS IT External Use of HMS High-Performance Computing (HPC) Resources policy, researchers from labs whose PIs do not have a primary or secondary faculty appointment in an HMS Quad department (external users) will be charged for use of HMS High-Performance Computing resources.

  • Core user: This is the person using, or consuming the service(s) provided by the HMS RC Core.
  • PI or Lab Head: Person that is defined as a responsible party who will approve HMS RC use requests for members of their research group and, in the case of external labs, this person will be responsible for paying the invoice for RC services consumed by their lab members.
  • Invoices:  Bills that will be sent to the lab/group financial manager identified by each PI or Lab Head.
  • HMS PPMS RC Core system:  The core facilities management system used to manage and send invoices from the HMS RC Core.  This system required you have an HMS ID/eCommons ID.  The web address is
  • A purchase order (PO) is required for all external PI's or Lab Heads as defined by the External use of HMS High-Performance Computing (HPC) Resources Policy.
  • For users and lab groups from Harvard University, a Harvard 33-digit number will be accepted by the Research Computing Core via the HMS PPMS RC Core system.
  • Initially, billing for all services offered by HMS RC Core will occur quarterly, following the Harvard fiscal year, July 1- June 30; Q1 starts July 1, Q2 starts Oct. 1, etc. For quarterly invoices, invoices for the previous quarter will be sent at the beginning of the next quarter. Once RC Core billing processes become more established, the core will likely move to monthly invoicing; users will be notified in advance of this change.
  • The lab/group financial manager identified by each PI or Lab Head will receive the invoices.
  • For external users, the RC Core will be utilizing the Office of Controller, which is part of Harvard Financial Administration Department, for all accounts receivable (AR) payments.  All processes and procedures related to accounts receivable will follow the Harvard Financial Administration guidelines.  For more information about Harvard AR please visit
  • Invoices will contain all charges for core users for which a PI or Lab head is responsible.
  • In the event of job failure, which is a rare occurrence beyond the control of the RC Core, the user is responsible to cover the cost of those failed jobs.
Billing Disputes:
  • The Core user, PI, or Lab Head has up to 60 days from the day an invoice is sent to dispute any charges on that invoice.
  • All billing disputes, questions, or concerns should be sent to HMS RC Core will confirm receipt of the billing query within 5 business days.
  • HMS RC Core will have up to 30 business days to research and respond to billing queries.
  • The Core user, PI, or Lab Head should be available, if needed, to have a call or Zoom meeting with the HMS RC Core to gather more details about disputed charges or suspected errors.
Cancellation of service:
  • The Core user, PI, or the Lab Head can request cancellation of their use of RC services at any time; they will be responsible for all outstanding invoices, including any charges after the cancelation request has been submitted.
  • In order to cancel any RC Services, please use the HMS RC Off-boarding form.  
  • Once the HMS RC Core receives a request from the Core user, PI, or the Lab Head, to cancel use of RC services, the HMS RC Core will confirm receipt of the cancellation request and the requested cancellation date, via email to the requestor and to the PI/Lab Head.
  • All data should be migrated or transfer out from the relevant RC Core account(s) before the proposed cancellation date.
  • After the cancellation date, cancelled Core user accounts will be frozen for 30 days before they are deleted from HMS systems. 
  • For external PI/Lab groups: the network storage share (the total amount of storage available to each group) is currently limited to 10 Tebibytes (TiB, or 1.0995 Terabytes) of data.


Billing policies related to each HMS RC Core service


Compute Resources on High-Performance Computing (HPC) O2 system:
  • The cost of compute jobs consists of CPU and RAM utilization.
  • Each compute job will be charged for both CPU and RAM usage.
  • Compute usage of the O2 system will be measured in seconds consumed but billed in hours and fractions thereof.
  • HMS RC Core billing rates are listed at  These rates will be reviewed and updated annually.
  • Compute jobs that fail or are killed will be charged for the amount of usage before the job failed or was killed.
  • Any jobs that are submitted that fail or are killed while running on the HMS O2 High-Performance Cluster are billable for the number of resources consumed before the failure or are killed and not available for any credits or refunds.
  • The GPU jobs will be billed based on GPU usage only, as per the HMS RC Core billing rates.
  • Compute or GPU jobs will only be billed once the job has finished.
HMS Storage Services:
  • The storage usage charges are based on the sum of usage measured daily and divided by the number of days in the billing period.
  • Storage usage will be measured in bytes and tracked by current total usage for all labs that use over 1 Tebibyte of storage.
  • Charges include 60 days of backup and data replication at no extra cost.
  • Usage is billed in units of Tebibytes (1.0995 Terabytes).
  • HMS Storage services provided by the HMS RC Core can only store data at Data Security Level 3 as defined by the Harvard University Information Security Policy at