FlameGraph
memray
FlameGraph | memray | |
---|---|---|
53 | 27 | |
16,438 | 12,545 | |
- | 0.9% | |
4.5 | 9.0 | |
16 days ago | 12 days ago | |
Perl | Python | |
- | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
FlameGraph
-
JVM Profiling in Action
We'll use async-profiler and flame graphs for profiling. To simplify the process, we'll run the code using JBang.
-
Memray – A Memory Profiler for Python
And flame graphs excel and this kind of thing
https://www.brendangregg.com/flamegraphs.html
-
All my favorite tracing tools: eBPF, QEMU, Perfetto, new ones I built and more
which can output in a format understood by Brendan Gregg's flame frames (https://www.brendangregg.com/flamegraphs.html)
But that's not quite the kind of tracing you're talking about. We also built a printf-style interface to our recording files, which seems closer:
-
Recap of Werner Vogels' Keynote at re:Invent 2023
Strategies included discontinuing or resizing underutilized services, transitioning to more cost-effective solutions, reducing the current resources to the amount of resources that we need for our application, and conducting detailed analyses of computing resource utilization through tools like flamegraphs. This detailed scrutiny helped identify and rectify significant cost-driving areas, such as garbage collection and application configurations.
-
Pinpoint performance regressions with CI-Integrated differential profiling
Flame Graphs by Brendan Gregg
-
Flameshow: A Terminal Flamegraph Viewer
Historically brendangregg's since AIUI he basically invented flamegraphs
https://www.brendangregg.com/flamegraphs.html
So if you can make your tool eat whatever https://github.com/brendangregg/FlameGraph is fed with you're going to support a lot of existing tooling across OSes and languages.
-
Introducing Flame graphs: It’s getting hot in here
“Flame graphs are a visualization of hierarchical data, created to visualize stack traces of profiled software so that the most frequent code-paths to be identified quickly and accurately.”
-
Using SVG to create simple sparkline charts
SVGs are amazing for interactive visualisation too. Like Flamegraphs: https://www.brendangregg.com/flamegraphs.html
-
Good example of using flame graphs to speed up java code (50x improvement)
This may be a good example of the application of a flame graph but it is not a good demonstration of flame graphs; the graph is nearly incidental. The source has an actual explanation.
-
Intro to PostGraphile V5 (Part 1): Replacing the Foundations
A profiling flame graph from Graphile Crystal (a precursor to Grafast) using GraphQL.js' executor (each tick is 1ms, total: 29ms). As we removed more and more responsibilities from GraphQL.js, we ended up only using it for output. Replacing this final responsibility with a custom implementation in Graphile Crystal itself, we reduced execution time for this query down to 15.5ms (effectively removing the majority of the yellow portion of the flame graph).
memray
-
Memray – A Memory Profiler for Python
I collected a list of profilers (also memory profilers, also specifically for Python) here: https://github.com/albertz/wiki/blob/master/profiling.md
Currently I actually need a Python memory profiler, because I want to figure out whether there is some memory leak in my application (PyTorch based training script), and where exactly (in this case, it's not a problem of GPU memory, but CPU memory).
I tried Scalene (https://github.com/plasma-umass/scalene), which seems to be powerful, but somehow the output it gives me is not useful at all? It doesn't really give me a flamegraph, or a list of the top lines with memory allocations, but instead it gives me a listing of all source code lines, and prints some (very sparse) information on each line. So I need to search through that listing now by hand to find the spots? Maybe I just don't know how to use it properly.
I tried Memray, but first ran into an issue (https://github.com/bloomberg/memray/issues/212), but after using some workaround, it worked now. I get a flamegraph out, but it doesn't really seem accurate? After a while, there don't seem to be any new memory allocations at all anymore, and I don't quite trust that this is correct.
There is also Austin (https://github.com/P403n1x87/austin), which I also wanted to try (have not yet).
Somehow this experience so far was very disappointing.
(Side node, I debugged some very strange memory allocation behavior of Python before, where all local variables were kept around after an exception, even though I made sure there is no reference anymore to the exception object, to the traceback, etc, and I even called frame.clear() for all frames to really clear it. It turns out, frame.f_locals will create another copy of all the local variables, and the exception object and all the locals in the other frame still stay alive until you access frame.f_locals again. At that point, it will sync the f_locals again with the real (fast) locals, and then it can finally free everything. It was quite annoying to find the source of this problem and to find workarounds for it. https://github.com/python/cpython/issues/113939)
-
Microservice memory profiling
second time was nastier. I used https://github.com/bloomberg/memray to try to spot it - that's the tool you should try out. You load your service through memray, and it will get you some stats that you can export as a flamegraph. I can't really afford to make it run on production so I ran it in a docker image and repeatedly ran the scenario I thought was responsible. Didn't find anything. I know what I did wrong: I assumed one particular codepath was the problem. If would have find the issue if I had a really complete scenario that covers broadly every possible endpoint and condition. Can't blame memray, that tool is really promising.
-
Big Data Is Dead
This is an excellent summary, but it omits part of the problem (perhaps because the author has an obvious, and often quite good solution, namely DuckDB).
The implicit problem is that even if the dataset fits in memory, the software processing that data often uses more RAM than the machine has. It's _really easy_ to use way too much memory with e.g. Pandas. And there's three ways to approach this:
* As mentioned in the article, throw more money at the problem with cloud VMs. This gets expensive at scale, and can be a pain, and (unless you pursue the next two solutions) is in some sense a workaround.
* Better data processing tools: Use a smart enough tool that it can use efficient query planning and streaming algorithms to limit data usage. There's DuckDB, obviously, and Polars; here's a writeup I did showing how Polars uses much less memory than Pandas for the same query: https://pythonspeed.com/articles/polars-memory-pandas/
* Better visibility/observability: Make it easier to actually see where memory usage is coming from, so that the problems can be fixed. It's often very difficult to get good visibility here, partially because the tooling for performance and memory is often biased towards web apps, that have different requirements than data processing. In particular, the bottleneck is _peak_ memory, which requires a particular kind of memory profiling.
In the Python world, relevant memory profilers are pretty new. The most popular open source one at this point is Memray (https://bloomberg.github.io/memray/), but I also maintain Fil (https://pythonspeed.com/fil/). Both can give you visibility into sources of memory usage that was previous painfully difficult to get. On the commercial side, I'm working on https://sciagraph.com, which does memory and also performance profiling for Python data processing applications, and is designed to support running in development but also in production.
-
Check Python Memory Usage
bloomberg/memray: Memray is a memory profiler for Python
- What Python library do you wish existed?
- Modules Import and Optimisation
-
The hand-picked selection of the best Python libraries and tools of 2022
Memray — a memory profiler
-
Python 3.11 delivers.
Python profiling is enabled primarily through cprofile, and can be visualized with help of tools like snakeviz (output flame graph can look like this). There are also memory profilers like memray which does in-depth traces, or sampling profilers like py-spy.
-
Memory Profiling for Python
I've been using this recently for memory profiling with Python, it works pretty well: https://github.com/bloomberg/memray
-
What stack or tools are you using for ensuring code quality and best practices in medium and large codebases ?
great suggestions in this thread. i also recommend performance testing your codebase. these include techniques such as: - creating micro performance benchmarks - using [cProfile] (and learning how to plot / read flame graphs) - memory profiling (e.g. via memray)
What are some alternatives?
hotspot - The Linux perf GUI for performance analysis.
scalene - Scalene: a high-performance, high-precision CPU, GPU, and memory profiler for Python with AI-powered optimization proposals
benchmark - A microbenchmark support library
pyinstrument - 🚴 Call stack profiler for Python. Shows you why your code is slow!
tracing-bunyan-formatter - A Layer implementation for tokio-rs/tracing providing Bunyan formatting for events and spans.
MemoryProfiler - memory_profiler for ruby
HeatMap - Heat map generation tools
viztracer - VizTracer is a low-overhead logging/debugging/profiling tool that can trace and visualize your python code execution.
node-clinic - Clinic.js diagnoses your Node.js performance issues
magic-trace - magic-trace collects and displays high-resolution traces of what a process is doing
pmu-tools - Intel PMU profiling tools
py-spy - Sampling profiler for Python programs