Intel® Memory Resilience Technology
A memory RAS technology to analyze and predict memory failures and isolate fault bits
Intel® Memory Resilience Technology Use Case
Through our multi-dimensional model and algorithms, memory errors are mined and analyzed at the micro-architectural level, Intel® Memory Resilience Technology can be used for:
- Defective memory part detection & report
- Predictive page offlining to isolate fault bits
- Reducing unnecessary DIMM replacement and maintenance efforts
- Reducing e-waste and extend electric component life
Case Studies and White Papers
Tencent: Intel® Memory Failure Prediction at Tencent
Intel® Memory Failure Prediction substantially improves memory reliability through online machine learning and reduces downtime.
Meituan: Intel® Memory Failure Prediction Improves Reliability
Intel® Memory Failure Prediction uses machine learning to send potential memory failure alerts prior to hardware failure and thus reducing impact of downtime.
JD Cloud: Intel® Memory Failure Prediction at JD Cloud
Machine Check Architecture Recovery (MCA Recovery) + Intel MFP helps JD Cloud provide efficient and stable services.
ByteDance: Fault-Aware Prediction-Guided Page Offlining for Uncorrectable Memory Error Prevention
Empirical evaluation in ByteDance, fault-aware prediction-guide page offlining prevents 20X more UEs than the traditional OS policy.
Samsung: Improving Memory Reliability at Data Centers
Intel® Memory Resilience Technology reduces memory errors at data centers with predictive failure analysis and intelligent prevention using fine grain telemetry and artificial intelligence
Solution Providers
AMI
AMI enhanced memory module support through integrated Intel® Memory Resilience Technology
Supervyse® Memory/FP™ with Intel® Memory Resilience Technology
Reduce server downtime by enabling memory failure analysis and prediction with Supervyse Memory/FP add-in for Supervyse BMC firmware.
Articles and Papers
In the News
Using AI to Boost Data Center Uptime and Lower TCO
February 9, 2021 | George Clement, and Zachary Bobroff
How Memory Failure Prediction Keeps Data Centers and the Digital Economy Up and Running
December 3, 2020 | Jeff Klaus, Data Center News
Notices and Disclaimers
Intel does not control or audit third-party data. You should consult other sources to evaluate accuracy.
Memory failure prediction results provided through the use of Intel MFP are estimated and may vary based on differences in system hardware, software, or configuration. Results are derived using multi-dimensional models and algorithms to predict potential memory failures and do not constitute a representation or guarantee regarding memory failure.