Hi, I think it would be awesome if you could make available ARMv7+NEON builds, tuned for Cortex-A7; i.e. where libstd will be compiled with NEON auto-vectorization enabled. These builds would be very useful for maximizing performance for Raspberry Pi 2 projects, which is NEON-enabled Cortex-A7. I would use these write away for building and testing ring and my other projects for Raspberry Pi 2.
If there's something I can do to help facilitate this, let me know. Thanks!
Probably moot now, but with the exception of hardware division in libstd et al. (that's what tuning for Cortex-A7 implies) those builds already were ARMv7+NEON. (armv7- triple)
Hi, I think it would be awesome if you could make available ARMv7+NEON builds, tuned for Cortex-A7; i.e. where libstd will be compiled with NEON auto-vectorization enabled. These builds would be very useful for maximizing performance for Raspberry Pi 2 projects, which is NEON-enabled Cortex-A7. I would use these write away for building and testing ring and my other projects for Raspberry Pi 2.
If there's something I can do to help facilitate this, let me know. Thanks!