The HECToR Service is now closed and has been superceded by ARCHER.

HECToR Phase 3 Update - 24th Nov

Tests are progressing on the machine as planned. Further information will be issued in due course.

HECToR Phase 3 Update - 16th Nov

The Phase3 Upgrade Plan has now been revised.

Cray has been asked by the factory to suspend the handover of the upgraded HECToR system to users. There have been some failure modes identified that require additional checkout tests before the system can go into production. These additional tests are currently being run. We will provide updates as soon as we have further details of the revised upgrade schedule.

Maintenance Today

The planned maintenance today has gone ahead, however this is now to support electrical works and to fully partition the current Magny Cours and Interlagos cabinets. This work will also ensure that we have eliminated the job scheduling issues which occurred this week. Note - The system will return with a 10 cabinet Magny Cours configuration.

Interlagos Test Proposal

To reduce any risk and to ensure a reliable system once the service is moved to the new processors, we are now scheduling full onsite quality assurance testing. These test runs will last for roughly one week, which is the equivalent of a full test run on the Cray factory floor. Following on from today's maintenance, this testing will commence on the first 20 Interlagos cabinets.

The nature of the testing means we will be running the facility at very high capacity over an extended period of time, and we may have to review the test schedule should safety dictate this. For example, we may need to limit the number of cabinets that can be tested in this way to 10. Should this happen then of course the timeline will change. Having eliminated any problem nodes from the system, the service would then conduct 'user level' testing, focusing on the codes important to the HECToR users. We would then switch the service to the 20 Interlagos cabinets. The remaining 10 cabinets of Magny Cours would then be upgraded and tested in the same way, before being integrated back into the user service.

User Impact

So what does this mean? In summary the reduced sized system will remain in place for now. The exact dates of when things will happen is not included above, but we will issue these as soon as we can. We will keep you informed every few days on progress and any changes to the plan.

We understand that this is a great inconvenience, but we also hope that due to the nature of the problem you can understand that this change in plan cannot be avoided. Our sincere apologies for this. We will do our utmost to avoid unnecessary disruption during this process.

Any users requiring assistance should contact the HECToR Helpdesk.