Hi, I have encountered some problems using your company's vtune recently, see if anyone can help me locate it. Using vtune to collect 5 minutes of data, it takes several hours to analyze (vtune 2020 and 2022 have been tried, both are slow).
Some background information:
1. The operating system is Tencent tlinux, and the virtual machine is developed based on centos7 2. The cpu is Intel(R) Xeon(R) Platinum 8255C CPU @ 2.50GHz, 8 cores 3. gcc version 7.3.1, using c++17 standard, compile optimization option O2 4. The size of the executive program is 150M, the symbol table nm |wc has 150,000 lines, and 2G shared memory is used during runtime
Use perf to view the call hotspots when vtune finalize, basically malloc and free:
Start another vtune instance to collect vtune hotspots:
Thank you for posting in the Intel Communities.
As you mentioned, you are using Vtune 2022.1 on Tencent tlinux, which is not present in supported operating systems. So, could you please try it on a supported operating system and let us know?
Note: Please avoid trying on deprecated operating systems.
System requirements link:
We informed our team to make changes to the documentation to avoid confusion. We tried to run a 5-minute application with the command you provided and it was running fine without any issues on Centos 7, and it took a few seconds to analyze the data. Please go through the attached screen shots.
OS : Linux (Centos7 distribution)
Vtune Version: 2022.1
Could you please answer the following questions:
1. Are you trying to run your application inside of a VM?
2. Please, can you share the exact steps you followed and a sample reproducer (a sample application that is similar to the application you are trying to analyze)?