From jacobnavia@21:1/5 to All on Fri Oct 13 09:37:01 2017
XPost: comp.lang.c
Hi.
Here are the first results for the port of lcc-win to the ARM64 processor.
All times are in seconds. The source code of lcc is used as test bed.
This code consists of 70206 lines of C, and 1169 lines of machine
description that are translated into a C source of 11696 lines. In
total, we have then 81902 lines (2 254 910 bytes).
The machine running the tests is an ARM64 with 4GB of RAM and a 1TB SSD
disk under the linux operating system. Its name is "Rock64", a cheap
machine costing some dozens of dollars. Both gcc and lcc use exactly the
same hardware and exactly the same source code.
Results:
Raw Speed:
Time needed for the gcc's generated program and for the lcc generated
program to compile all the source code. Both compilers are run under
full optimizations
GCC: 7.541
LCC: 9.716
Compilation time:
GCC: 130.379
LCC: 9.716
This is the time needed by gcc to produce the optimized executable that
runs in only 7.5 seconds. For lcc is the same time as above since the
benchmark is run under full optimization also.
Executable size: (only the .text section is shown here)