Creating a dump file is excruciating slow
Brief
Creating a dump file using any of the following methods is way to slow (> 2 minutes) to be usable for our current scenario.
procdump -ma notepadprocdump notepad- Open task-manager, select notepad, create dump file
Does anyone have any idea
- what might be causing this
- how to possibly speed this up
Edit: some extra information
- Using procexp and looking at the call stack dumping notepad, it seems to be taking 100% cpu of one core in a method from dbghelp:
dbghelp.dll!MiniDumpReadDumpStream+0x241a
fortunatly, we tried this first with notepad
Background
We are trying to create a dumpfile on a first-chance exception of a critical application with as minimal impact as possible. The application should not get suspended for more than a few seconds.
Solution
In the end, this turned out to be related to the combination of our AV and a specific dbghelp.dll version.
Tested combinations
- AV Enabled with dbghelp version 6.0 is way to slow
- AV Disabled with dbghelp version 6.0 is no problem
- AV Enabled with dbghelp version 6.3 is no problem
- Would adding a CustomerId to the Order entity violate DDD or database design principles to improve lookup performance?
- Efficient way to insert similar rows (with only one column changing) right after each row in Numpy or Pandas
- Performance using STDOUT (screen) vs regular file
- Why is this code 5 times slower in C# compared to Java?
- Dependency chain analysis
- Is performance reduced when executing loops whose uop count is not a multiple of processor width?
- Poor performance sorting multiple tables in mysql 8.4
- Looking for a more efficient way to replace matrix elements
- Strange thing about ArrayList Performance (Java)
- Keep PostgreSQL from sometimes choosing a bad query plan
- Python - Entire PC network gets slower during repeated request
- Fast check for NaN in NumPy
- Is it better to pre-allocate array in python or use arr.append()?
- Fastest way to calculate the sum of row-wise "v.T @ v" of a matrix
- Fastest way to merge columns of the same type in Numpy
- Why is the size of my pages so big as compared to Wikipedia?
- How do I benchmark queries in PostgreSQL?
- Fastest way to find all pairs of close numbers in a Numpy array
- Performance issues with recursive stored procedure for hierarchical data
- Fastest way to find the indices of two Numpy arrays that meet a condition
- SQL Query slow in .NET application but instantaneous in SQL Server Management Studio
- Do CSS3 transitions slow down websites?
- The reason behind slow performance in WPF
- Why Numpy fabs is much slower than abs?
- Get the second largest number in a list in linear time
- iOS - How can I preload the keyboard?
- Is simple but frequent usage of std::stringstream a premature pessimization?
- Which if statement would execute faster?
- WPF Performance when using frozen Brushes as resource
- Converting double to char* in C++ with high performance