Using AVX CPU instructions: Poor performance without “/arch:AVX”

by

2021 update: Modern versions of MSVC don’t need manual use of _mm256_zeroupper() even when compiling AVX intrinsics without /arch:AVX. VS2010 did.

The behavior that you are seeing is the result of expensive state-switching.

See page 102 of Agner Fog’s manual:

http://www.agner.org/optimize/microarchitecture.pdf

Every time you improperly switch back and forth between SSE and AVX instructions, you will pay an extremely high (~70) cycle penalty.

When you compile without /arch:AVX, VS2010 will generate SSE instructions, but will still use AVX wherever you have AVX intrinsics. Therefore, you’ll get code that has both SSE and AVX instructions – which will have those state-switching penalties. (VS2010 knows this, so it emits that warning you’re seeing.)

Therefore, you should use either all SSE, or all AVX. Specifying /arch:AVX tells the compiler to use all AVX.

It sounds like you’re trying to make multiple code paths: one for SSE, and one for AVX.
For this, I suggest you separate your SSE and AVX code into two different compilation units. (one compiled with /arch:AVX and one without) Then link them together and make a dispatcher to choose based on the what hardware it’s running on.

If you need to mix SSE and AVX, be sure to use _mm256_zeroupper() or _mm256_zeroall() appropriately to avoid the state-switching penalties.

More Related Contents:

  1. Why does changing 0.1f to 0 slow down performance by 10x?
  2. How to solve the 32-byte-alignment issue for AVX load/store operations?
  3. How to efficiently perform double/int64 conversions with SSE/AVX?
  4. How to check if a CPU supports the SSE3 instruction set?
  5. Loading 8 chars from memory into an __m256 variable as packed single precision floats
  6. Loop unrolling to achieve maximum throughput with Ivy Bridge and Haswell
  7. cpu dispatcher for visual studio for AVX and SSE
  8. AVX/SSE version of xorshift128+
  9. Get sum of values stored in __m256d with SSE/AVX
  10. Why does C++ code for testing the Collatz conjecture run faster than hand-written assembly?
  11. Is there any advantage of using map over unordered_map in case of trivial keys?
  12. Static linking vs dynamic linking
  13. How to build Qt for Visual Studio 2010
  14. Efficient string concatenation in C++
  15. Using scanf() in C++ programs is faster than using cin?
  16. Relative performance of std::vector vs. std::list vs. std::slist?
  17. std::to_string – more than instance of overloaded function matches the argument list
  18. Visual Studio 2010 not autolinking static libraries from projects that are dependencies as it should be supposed to
  19. Generating m distinct random numbers in the range [0..n-1]
  20. Is the ranged based for loop beneficial to performance?
  21. What is the performance cost of having a virtual method in a C++ class?
  22. Why is iterating though `std::vector` faster than iterating though `std::array`?
  23. Visual Studio 2010’s strange “warning LNK4042”
  24. Difference between rdtscp, rdtsc : memory and cpuid / rdtsc?
  25. c++11 regex slower than python
  26. What is a good random number generator for a game?
  27. Debug and Release Library Linking with CMAKE (VISUAL STUDIO)
  28. Simple for() loop benchmark takes the same time with any loop bound
  29. C++ most efficient way to convert string to int (faster than atoi)
  30. Where can I find an official reference listing the operation of SSE intrinsic functions?

Leave a Comment