Had a bunch of thoughts about the recent safety stuff, way more than fit in social media post... Blog post story time! (It's a bit of a ramble, sorry about that...)
https://chandlerc.blog/posts/2024/11/story-time-bounds-checking/
chandlerc.blog · Story-time: C++, bounds checking, performance, and compilersRecently, several of my colleagues at Google shared the story of how we are
retrofitting spatial safety onto our monolithic C++ codebase:
https://security.googleblog.com/2024/11/retrofitting-spatial-safety-to-hundreds.html
I wanted to have a bit of story-time about some of the strange ways that all
this came to be, at least as I remember things. There are some really
interesting developments that led us here, and some important lessons to learn
from that history.
Do note that this is just my retrospective memory. It’s entirely possible I’m
misremembering some of it (let me know if so!). It’s also limited to my
perspective, and others may have seen very different aspects of things (please
share!).
![L] H N AMD EPYC™ 9xx4-series Processors e L . . . . Sl i Compiler Options Quick Reference Guide . GNU compiler collection Microsoft® Visual Studio 2022 Latest release: GCC 13.2 , July 2023 Latest release: 17.6.3, June 2023 Recommended version: GCC 13.1 or later https://visualstudio.microsoft.com/ User Guide http://gcc.gnu.org Architecture Architecture Generate instructions that runs on AMD -march=znver4 Generate instructions that runson /arch:[AVX|AVX2|AVX512] 4" Gen EPYC ™ and AMD 4" Gen AMD 4" Gen EPYC ™ and AMD 4" RyzenT'\’I Gen RyzenT'\’I Generate instructions supported in the -march=native Optimize for 64-bit AMD processors /favor:AMD64 given machine Optimization Levels Optimization Levels Disable optimizations /0d Disables all optimizations {default) -00 . S — Maximum optimizations (favor](https://cdn.masto.host/floss/media_attachments/files/113/293/821/020/497/922/original/cf14c0bf5e8d3cd2.webp "L] H N AMD EPYC™ 9xx4-series Processors e L . . . . Sl i Compiler Options Quick Reference Guide . GNU compiler collection Microsoft® Visual Studio 2022 Latest release: GCC 13.2 , July 2023 Latest release: 17.6.3, June 2023 Recommended version: GCC 13.1 or later https://visualstudio.microsoft.com/ User Guide http://gcc.gnu.org Architecture Architecture Generate instructions that runs on AMD -march=znver4 Generate instructions that runson /arch:[AVX|AVX2|AVX512] 4\" Gen EPYC ™ and AMD 4\" Gen AMD 4\" Gen EPYC ™ and AMD 4\" RyzenT'\’I Gen RyzenT'\’I Generate instructions supported in the -march=native Optimize for 64-bit AMD processors /favor:AMD64 given machine Optimization Levels Optimization Levels Disable optimizations /0d Disables all optimizations {default) -00 . S — Maximum optimizations (favor")
![™ H o N AMD EPYC™ 9xx4-series Processors LLLr]
. : . . = Compiler Options Quick Reference Guide . AOCC compiler (C/C++/Fortran)
Latest release: 4.1, August 2023 https://www.amd.com/en/developer/aocc.html Architecture Other options Generate instructions that runs on -march=znver4 Enables faster, less precise -ffast-math AMD 4% Gen EPYC ™ and AMD 4% math operations {part of Ofast) -freciprocal-math ™ Gen Ryzen OpenMP® threads and affinity ~ export OMP_NUM_THREADS=N Generate instructions supported -march=native (N number of cores) export GOMP_CPU_AFFINITY="0- in the given machine {N-1}" Optimization Levels Link to AMD library -L/libm-install-dir/lib -lamdlibm -Im Disables all optimizations -00 Enables vector library -lamdlibm -fveclib=AMDLIBM -Im Enables minimal level optimiza- -01 Enables faster library -lamdlibm -fsclrlib=AMDLIBM - tions lamdlibmfast -Im Enables moderate level optimiza- -02/-O For Fortran Workloads tions (Default from AOCC 4.1) . Compiles Fortran free form -ffree-form Enables all optimizations that -03 attempt to make programs run faster .. . . AMD Optimized Libraries Enables O3 with other aggressive -Ofast Latest release: 4.1, August 2023 ER R R https://www.amd.com/en/developer/aocl.html strict compliance and precisions Enables link time optimization -flto . AMD pProf (Performance & Power Profiler) More advanced optimizations - -zopt Latest release: 4.1 August 2023 Enables](https://cdn.masto.host/floss/media_attachments/files/113/293/821/010/884/970/original/89f737236f29228c.webp "™ H o N AMD EPYC™ 9xx4-series Processors LLLr]
. : . . = Compiler Options Quick Reference Guide . AOCC compiler (C/C++/Fortran)
Latest release: 4.1, August 2023 https://www.amd.com/en/developer/aocc.html Architecture Other options Generate instructions that runs on -march=znver4 Enables faster, less precise -ffast-math AMD 4% Gen EPYC ™ and AMD 4% math operations {part of Ofast) -freciprocal-math ™ Gen Ryzen OpenMP® threads and affinity ~ export OMP_NUM_THREADS=N Generate instructions supported -march=native (N number of cores) export GOMP_CPU_AFFINITY=\"0- in the given machine {N-1}\" Optimization Levels Link to AMD library -L/libm-install-dir/lib -lamdlibm -Im Disables all optimizations -00 Enables vector library -lamdlibm -fveclib=AMDLIBM -Im Enables minimal level optimiza- -01 Enables faster library -lamdlibm -fsclrlib=AMDLIBM - tions lamdlibmfast -Im Enables moderate level optimiza- -02/-O For Fortran Workloads tions (Default from AOCC 4.1) . Compiles Fortran free form -ffree-form Enables all optimizations that -03 attempt to make programs run faster .. . . AMD Optimized Libraries Enables O3 with other aggressive -Ofast Latest release: 4.1, August 2023 ER R R https://www.amd.com/en/developer/aocl.html strict compliance and precisions Enables link time optimization -flto . AMD pProf (Performance & Power Profiler) More advanced optimizations - -zopt Latest release: 4.1 August 2023 Enables")
![L] H N AMD EPYC™ 9xx4-series Processors LT . : . ) ™ Compiler Options Quick Reference Guide Intel® oneAPI DPC++/C++ Compiler Latest release: 2023.1.0 http://software.intel.com Architecture Generate instructions that runs on -axCORE-AVX512 AMD 4" Gen EPYC ™ and AMD 4" Gen RyzenT'\’I Optimization Levels Disable all optimizations -00 Speed optimization without code -01 growth Enables optimization for speed in- -02 cluding vectorization Enables 02 and aggressive loop -03 Enables set of aggressive options to ~ -Ofast Additional Optimizations Sets function inline level -inline-level=<value> Sets unroll loop maximum threshold -unroll<value> Disable improved precision floating -no-prec-div divides Enables vectorization -vec Enables inter procedural optimiza- -ipo Enables use of OpenMP® directives -gopenmp Enables profile generated optimiza- -prof-gen and -prof-use Other Options Enables floating point accuracy tun- -fp-model Compiles Fortran free form layout -free Advanced Micro Devices © 2020-23 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, AMD EPYC, AMD Ryzen and One AMD Place, P.O. Box 3453 combinations thereof are trademarks of Advanced Micro Devices, Inc. Linux is a registered trademark of Linus Torvalds. Sunnyvale, CA 94088-3453 OpenMP. Microsoft, Intel are for informational purposes only and may be trademarks of their respective owners.](https://cdn.masto.host/floss/media_attachments/files/113/293/821/001/432/090/original/a90b50100f6ae921.webp "L] H N AMD EPYC™ 9xx4-series Processors LT . : . ) ™ Compiler Options Quick Reference Guide Intel® oneAPI DPC++/C++ Compiler Latest release: 2023.1.0 http://software.intel.com Architecture Generate instructions that runs on -axCORE-AVX512 AMD 4\" Gen EPYC ™ and AMD 4\" Gen RyzenT'\’I Optimization Levels Disable all optimizations -00 Speed optimization without code -01 growth Enables optimization for speed in- -02 cluding vectorization Enables 02 and aggressive loop -03 Enables set of aggressive options to ~ -Ofast Additional Optimizations Sets function inline level -inline-level=<value> Sets unroll loop maximum threshold -unroll<value> Disable improved precision floating -no-prec-div divides Enables vectorization -vec Enables inter procedural optimiza- -ipo Enables use of OpenMP® directives -gopenmp Enables profile generated optimiza- -prof-gen and -prof-use Other Options Enables floating point accuracy tun- -fp-model Compiles Fortran free form layout -free Advanced Micro Devices © 2020-23 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, AMD EPYC, AMD Ryzen and One AMD Place, P.O. Box 3453 combinations thereof are trademarks of Advanced Micro Devices, Inc. Linux is a registered trademark of Linus Torvalds. Sunnyvale, CA 94088-3453 OpenMP. Microsoft, Intel are for informational purposes only and may be trademarks of their respective owners.")
), aside from some shitposts, expect toots about: