02-22-2019 12:27 AM
I am trying to implement ImageMagick application on Zynq ZCU102. When I was building it, the compiler complained about __float128 and _Float128 and the build failed.
/usr/include/bits/floatn.h:87:9: error: ‘__float128’ does not name a type; did you mean ‘__cfloat128’? typedef __float128 _Float128; ^~~~~~~~~~ __cfloat128 from /usr/local/include/ImageMagick-6/Magick++.h:9, from /users/Experiments/SDx_workspace/ImageMagick_Zynq/src/Magick++/tests/appendImages.cpp:8: /usr/include/stdlib.h:152:8: error: ‘_Float128’ does not name a type; did you mean ‘_Float32x’? extern _Float128 strtof128 (const char *__restrict __nptr, ^~~~~~~~~ _Float32x /usr/include/stdlib.h:245:4: error: ‘_Float128’ has not been declared _Float128 __f) ^~~~~~~~~
I tried some of the solutions of the following link, but it did not work.
Is there any workaround to support _Float128 in SDSoC?
02-22-2019 12:44 AM
Use below option if the default behavior results in target dependent compilation errors.
add "-hls-target 1" to sds++ compiler->inferred options->software platform
When set to 1, in Vivado HLS add_files commands, insert -targetand Arm GNU toolchain include options in addition to -m32 or -m64 options.
When set to 0, insert -m32 or -m64options.
02-24-2019 05:40 PM
Hi @nutang ,
Thanks for your reply. I tried it but it did not work. I am gettign the same error.
I am basically trying to build and run this application on Arm processor only (no hardware function). Next I will select HW functions after profiling.
Waiting for your reply.