# Magic-trace: Diagnosing tricky performance issues easily with Intel Processor Trace

DevFeed: [Magic-trace: Diagnosing tricky performance issues easily with Intel Processor Trace](<https://devfeed.tech/articles/magic-trace-diagnosing-tricky-performance-issues-easily-with-intel-processor-trace-20193.md>)

Original publisher: [Read original article](<https://blog.janestreet.com/magic-trace/>)

Author: Tristan Hume

Published: 2022-01-11T00:00:00Z

Content type: article

Language: en

Sources: [Jane Street](<https://devfeed.tech/sources/jane-street.md>)

Topics: [tracing](<https://devfeed.tech/topics/tracing.md>), [Latency](<https://devfeed.tech/topics/latency.md>), [cpu](<https://devfeed.tech/topics/cpu.md>), [Hardware](<https://devfeed.tech/topics/hardware.md>), [systems](<https://devfeed.tech/topics/systems.md>), [Instrumentation](<https://devfeed.tech/topics/instrumentation.md>), [OCaml](<https://devfeed.tech/topics/ocaml.md>), [Google](<https://devfeed.tech/topics/google.md>)

Tags: [flow](<https://devfeed.tech/tags/flow.md>), [instrumentation](<https://devfeed.tech/tags/instrumentation.md>), [intel](<https://devfeed.tech/tags/intel.md>), [latency](<https://devfeed.tech/tags/latency.md>), [ocaml](<https://devfeed.tech/tags/ocaml.md>), [performance](<https://devfeed.tech/tags/performance.md>), [processor](<https://devfeed.tech/tags/processor.md>), [trace](<https://devfeed.tech/tags/trace.md>), [tracing](<https://devfeed.tech/tags/tracing.md>), [visualization](<https://devfeed.tech/tags/visualization.md>)

## AI overview

Jane Street describes magic-trace, an open-source tool built on Intel Processor Trace. It captures roughly 10 milliseconds of execution leading up to a selected function call and visualizes the call stack on a timeline, helping diagnose fine-grained performance and latency issues that sampling profilers may miss.

## Source excerpt

Intel Processor Trace is a hardware technology that can record all program execution flow along with timing information accurate to around 30ns. As far as I can tell almost nobody uses it, seemingly because capturing the data is tricky and, without any visualization tools, you're forced to read enormous text dumps.