Not sure if it's on 2.5 too, but, giving duron/athlon on the cpu type it
passes -march=athlon on gcc. Some athlons had no sse support when
e.g. -march=athlon-4 can be used for new amd cpus that have sse support.
For a duron 1.3 here, that shows sse on cpuinfo, passed march=athlon-4
on the arch Makefile, having no problems after two weeks uptime with
that kernel.
Are there any issues covering this detail or it's just not right?
- fs