6,075 questions
0
votes
0
answers
317
views
why does nginx worker process CPU usage is not 100% under wrk stress test?
when doing Nginx stress test by running test cmd: wrk -t1 -c40 -d50s http://127.0.0.1:80, I find the worker process of Nginx is not using 100% CPU, it is only 70% (from top command). I have tried ...
- 139
1
vote
0
answers
142
views
How does perf correlate hardware events with PMU registers?
I have programmatically set up perf in such a way to allow for direct reading of PMU registers by following perf docs and the ARM register summary.
Below is how I have set up perf to configure the CPU ...
- 193
0
votes
0
answers
154
views
How to monitor elastic beanstalk tomcat instance using a profiler
I am trying to figure out some performance issues in my spring boot application. A code block is running within a for loop and taking a lot of time(there are no network calls). I want to identify the ...
5
votes
0
answers
408
views
Profiling .NET core tests
The test suite I am executing is taking long time to finish. The suite is mainly integration tests involving database.
I'd like to find out possible bottlenecks and shorten the test execution time.
...
- 4,497
1
vote
1
answer
671
views
Strange pink flame graphs with Google pprof
The "new" flame graph in Google pprof is uniformly pink on my machine (Mac OS 10). Is this supposed to happen? It's not as nice as the old flame graph, so I wonder if something is not ...
- 5,891
4
votes
1
answer
634
views
Access frequency limitation for reading PMCCNTR_EL0?
I am using perf_event_open in my c profiling app to leverage perf in getting event data. In order to improve performance, I am reading the hardware registers directly by following the Perf Userspace ...
- 193
0
votes
0
answers
19
views
How to abstract typical memory behavior from a complex multi-threaded workload?
There is a complex multi-threaded program running in my system, which performs various types of operations. At the same time, I cannot access its source code.
I want to analyze the potential impact of ...
- 809
1
vote
0
answers
70
views
Why is the total execution time so much greater than the profiled cumulated time in Python?
I'm running a Python script to find flight routes using a dataset with around 20k flights containing ~4000 flights each day. When profiling the code with cProfile, I print that the cumulative time ...
- 964
1
vote
0
answers
284
views
External tools for a RAD Studio / C++ / Windows project? (How to get PDB from RAD Studio and/or what are other options?)
I have a project of a 64-bit Windows application in Rad Studio 11 (pre-existing project, not my choice, moving to another toolchain is not an option). The language is C++ (C++17). So far, I've been ...
- 1,662
1
vote
1
answer
335
views
Intellij Failed to activate JFR synchronization
Trying to profile my app locally and it isn't working.
It's saying
Profiling is not supported on java versions prior to 1.8. If you're using oracle jdk prior to version 11, or hotspot jdk prior to ...
- 465
1
vote
1
answer
205
views
How can a C# WinUI3 project be profiled when the lag is outside of the methods called?
I have an issue where some code is taking a very long time to complete, so I tried to profile the CPU usage, and the area of the graph that contains the lag has absolutely no methods to display. In ...
- 966
2
votes
1
answer
434
views
How to profile stack size for c programs
I am trying to optimize a program.
The program do not make use of heap.
I would like to profile the size of the stack to see:
How big is the stack?
Where the stack reach its peak?
What is the trend ...
1
vote
0
answers
142
views
GNU gprof showing calls in seconds instead of milliseconds
Is there an option i am missing that i couldnt find? i have searched the internets and all the articles show the flat profile in milliseconds without any added option tags, but my gprof outputs the ...
- 110
-2
votes
2
answers
634
views
how to find out which function(s) took most time during API invocation?
i am trying to debug why my spring boot application takes around 200-ms to respond to an API, and for subsequent hits of same API it takes around 10-ms.
no caching is involved in API.
I am using ...
- 454
2
votes
0
answers
92
views
perf_event_open - unable to monitor HW_CACHE write and prefetch events
I am attempting to monitor PERF_TYPE_HW_CACHE events. I have found that I am able to measure read accesses and misses for each cache, but if I try to measure write or prefetch accesses/misses the ...
- 193
0
votes
0
answers
43
views
unable to understand gprof output
I have captured the gprof profiler output while running a c++ program. I got below logs
I found that there are two functions which are consuming most time ie receiveXappMessages and load() function. I ...
- 11
0
votes
1
answer
59
views
fastavro.schemaless_reader performance loss when profiling is enabled
I am attempting to profile a Python app that uses the fastavro library.
I am profiling using the Datadog Profiler
I run the application using the command ddtrace-run python -m app.main
I enable the ...
2
votes
1
answer
644
views
Power Usage Profiling in Nsight?
New to Nsight and GPU programming. I need a way to evaluate the affect my code has on power usage in the GPU.
This article from 2013 shows that the feature was part of Nsight's toolset at some point, ...
- 23
1
vote
1
answer
1k
views
Is there a good way to profile threading in Python?
I am new to threading in Python. I am trying to make a game using PyGame where a CPU-heavy function gets its own thread, and where another function (that gets user input and renders graphics to the ...
- 142
0
votes
1
answer
148
views
Lettuce (Redis) + Mybatis (MySQL) take up most of the CPU of Spring Boot app in production - is it normal? how to optimize?
today I used https://github.com/async-profiler/async-profiler to check the CPU usage of my Spring Boot app (just a normal backend) in production. Surprisingly, Lettuce (Redis) + Mybatis (MySQL) take ...
- 17.9k
1
vote
1
answer
551
views
How to improve app startup profiling for flutter?
I am facing a delay in my flutter app at startup for android, the delay is before execution of dart code begins.
Mostly it is due to Geolocator plugin.
I have raised a ticket with them, however I want ...
2
votes
0
answers
111
views
Using inlined/expanded assembly to obtain a rough estimate of cpu cycle cost of a function?
I'm afraid I couldn't come up with a better term for what I want to ask about, other than "inlined/expanded assembly" - but let me try to explain through an example. The example will be for ...
- 5,948
0
votes
0
answers
306
views
First steps in performance profiler: why is "Memory Usage" not an available tool and what's the problem with remote analysis?
As mentioned in the title, I'm having my first steps, using the performance profiler in Visual Studio on a C# application, and I have following questions:
When starting up the performance profiler, ...
- 17.6k
1
vote
0
answers
181
views
MSDN Array error when training tensorflow model on Macbook M2 chip
I'm training reinforcement learning models using tensorflow (Python) but since few weeks I can't run my code anymore on my macbook air (Monteray 12.5) with M2 chip.
I get this error
/AppleInternal/...
- 394
0
votes
2
answers
1k
views
How to fix 'ModuleNotFoundError: No module named transformer.modules' while profiling a Python script with cProfile?
Trying to profile a python script using cProfile. It has subfiles being called encoder.py and decoder.py in the directory "modules" in "transformer". When I profile it, error shows ...
- 1