damondd / taf Goto Github PK

Taf(Trading Application Foundations) provides a framework for low-latency trading system development based on Intel x86-64.

• Based on Intel TBB, SIMD(AVX2+), lock-free queue, etc to optimize the performance.
• Nanosecond timestamp available.
• Developed with standard C++, corss-platform (windows and linux) is definitely supported.
• All dependencies included in pack, no need to download any other libs (like TBB).
• Provide C interface for multiple language calls (like Python).
• The delay for a simple application could be less 10us.

• Check your hardware availability to make sure that AVX2 or AVX512F is supported.
• C++17 compiler is needed.
• TBB libs and binary located in taf/libs/tbb/bin/

Windows

• Must use Visual Studio 2019 or above
• open taf/master/build/msvs/taf.sln
• Must select x64 mode
• Open taf/master/src/libs/taf/common/simd/simd.hpp to switch USE_AVX512 or USE_AVX512_BELOW micro (already existed in the file), depending on your hardware.
• Add /std:c++latest to C/C++ -> Command Line
• Config TBB env path in Configration->Debugging->Environment: path=..., could refer the properties config in example

Linux

• gcc 8.2 or above
• Example makefile is in taf/master/build/
• Must add -mavx2（-mavx512f） -std=c++17 to g++ command line
• Config TBB env path like LD_LIBRARY_PATH=... : /taf/libs/tbb/bin/linux
• example：
  $ cd taf/master/build
  $ make
  $ ./example