Subho Sankar Banerjee
Subho is a Software Engineer focusing on fleet-scale reliability and ML-driven systems initiatives at Google. He holds a PhD in Computer Science from the University of Illinois at Urbana-Champaign.
Research Areas
Authored Publications
Sort By
ITHICA: Intra-Thread Instruction Checking Approach for Defect-Induced Silent Data Corruptions
Ioanna Vavelidou
Eric Xue Liu
Mike Fuller
Subhasish Mitra
Caroline Trippel
59th IEEE/ACM International Symposium on Microarchitecture (MICRO) (2026)
Preview abstract
Hyperscalers are reporting silent data corruptions (SDCs), presumed to be caused by silicon manufacturing defects, as a threat to datacenter reliability. To support datacenter testing efforts to detect defective CPU servers, this paper presentsITHICA, an approach and tool for automatically generating functional tests for defect-induced errors from arbitrary programs by inserting intra-thread, instruction-level error checks, primarily leveraging instruction duplication and output comparison. Our key insight is that the most pernicious defects (those most likely to escape manufacturing testing) cause inconsistent errors: two executions of the same instruction given the same inputs within the same thread can produce different architectural outputs depending on the execution context in which they run. By exploiting this insight, ITHICA uniquely enables arbitrary programs to serve as tests and localizes affected instructions concurrently with error detection. We use ITHICA to transform industrial hyperscaler tests (our baseline), datacenter programs, and common libraries into functional tests, and evaluate them on over 3,000 CPU servers. ITHICA checks detect 39% more defective servers than baseline industrial checks and yield novel findings on defect behavior that challenge conclusions drawn by prior hyperscaler fleet studies.
View details
Silent Data Corruption by 10× Test Escapes Threatens Reliable Computing
Subhasish Mitra
Rama Govindaraju
Eric Liu
Mike Fuller
IEEE Design & Test, 42, Issue 6 (2025), pp. 40-53
Preview abstract
Summary:
Silent Data Corruption by 10x Test Escapes Threatens Reliable Computing" highlights a critical issue: manufacturing defects, dubbed "test escapes," are evading current testing methods at an alarming rate, ten times higher than industry targets. These defects lead to Silent Data Corruption (SDC), where applications produce incorrect outputs without error indications, costing companies significantly in debugging, data recovery, and service disruptions. The paper proposes a three-pronged approach: quick diagnosis of defective chips directly from system-level behaviors, in-field detection using advanced testing and error detection techniques like CASP, and new, rigorous test experiments to validate these solutions and improve manufacturing testing practices.
View details
