Uploaded November 2026 | Updated November 2026, 5 days ago
SREcon26 Europe/Middle East/Africa - Performance Prompts
Brendan Gregg
The rising cost of AI datacenters is a call for performance engineering: not just to use the best tools and techniques of yesterday, but to build new AI performance capabilities that are easier and faster. In 2000, performance experts were wizards, calculating system behavior from limited metrics and tools (e.g., from vmstat, iostat, and mpstat): It was the "performance math" era, and had many blind spots. By 2015, custom dashboards, dynamic instrumentation, eBPF, flame graphs, and an expectation of open source, enabled direct measurement almost everywhere, creating the "performance tools" era: Many tools, but also little time to learn them. By 2030 we will be well into the "performance prompts" era: AI prompts, skills, and agents to analyze all sources, including flame graphs, eBPF, Ftrace, PMCs, and MSRs. This talk summarizes these eras and shares performance prompts for code analysis, benchmarking, system metrics, and CPU analysis.
View the full SREcon26 Europe/Middle East/Africa program at usenix.org/conference/srecon26emea/program
SREcon26 Europe/Middle East/Africa - Performance Prompts
Brendan Gregg
The rising cost of AI datacenters is a call for performance engineering: not just to use the best tools and techniques of yesterday, but to build new AI performance capabilities that are easier and faster. In 2000, performance experts were wizards, calculating system behavior from limited metrics and tools (e.g., from vmstat, iostat, and mpstat): It was the "performance math" era, and had many blind spots. By 2015, custom dashboards, dynamic instrumentation, eBPF, flame graphs, and an expectation of open source, enabled direct measurement almost everywhere, creating the "performance tools" era: Many tools, but also little time to learn them. By 2030 we will be well into the "performance prompts" era: AI prompts, skills, and agents to analyze all sources, including flame graphs, eBPF, Ftrace, PMCs, and MSRs. This talk summarizes these eras and shares performance prompts for code analysis, benchmarking, system metrics, and CPU analysis.
View the full SREcon26 Europe/Middle East/Africa program at usenix.org/conference/srecon26emea/program










![USENIX Security 26 - PICS: Private Intersection over Committed (and reusable) Sets
USENIX Security 26 - PICS: Private Intersection over Committed (and reusable) Sets
Aarushi Goel, Rutgers University; Peihan Miao and Phuoc Van Long Pham, Brown University; Satvinder Singh, Purdue University
Private Set Intersection (PSI) enables two parties to compute the intersection of their private sets without revealing any additional information. While maliciously secure PSI protocols prevent many attacks, adversaries can still exploit them by using inconsistent inputs across multiple sessions. This limitation stems from the definition of malicious security in secure multiparty computation, but is particularly problematic in PSI because: (1) real-world applications—such as Apples PSI protocol for CSAM detection and private contact discovery in messaging apps—often require multiple PSI executions over consistent inputs, and (2) the PSI functionality makes it relatively easy for adversaries to infer additional information.
We propose Private Intersection over Committed Sets (PICS), a new framework that enforces input consistency across multiple sessions via committed sets. Building on the state-of-the-art maliciously secure PSI framework (i.e., VOLE-PSI [EUROCRYPT 2021]), we present an efficient instantiation of PICS using lightweight cryptographic tools. Our protocol achieves strong receiver-side input consistency (i.e., the receiver uses the exact committed set) and weak sender-side input consistency (i.e., the sender cannot inject new elements into the committed set but can potentially use a subset of the committed set). We implement our protocol to demonstrate concrete efficiency. Compared to VOLE-PSI, our communication overhead is a small constant between 1.57 - 2.04× for set sizes between 216-224, and the total end-to-end running time overhead is 1.22 - 1.98× across various network settings.
View the full USENIX Security 26 program at https://www.usenix.org/conference/usenixsecurity26/technical-sessions USENIX Security 26 - PICS: Private Intersection over Committed (and reusable) Sets](https://i.ytimg.com/vi/Q6rhdJr_J4M/mqdefault.jpg)