I'm reversing a program with x64dbg and ida (I switched from ghidra), hence disassembled and decompiled, I need to understand what is happening with dynamic analysis, but I have a tricky part where about 100 threads are running operations on chunks of a big file.
The program can't run in a VM, since it makes extensive use of AVX AVX2 and GPU.
How can I force the program to run this paralleled part as monothread ? In x64dbg I can suspend threads, but it's very tedious to do this by hand. Is there something I can do with the debuggers or should I change the assembly to launch at most 1 working thread (while allowing several non working threads) ?