> On Jan 2, 2017, at 3:06 PM, Felix Fietkau <n...@nbd.name> wrote: > > On 2017-01-02 19:36, Philip Prindeville wrote: >> Right, this is why I’m trying to create a new target (or subtarget) >> called “xeon” which is optimized for Xeon targets and leverages the >> on-chip crypto-accelerators. >> >> We’ve come a long way since the Athalon-64 (k8) in 2004. > I think it should be easily possible to enable the crypto accelerator > support by default, it probably gets enabled only on CPUs that actually > support it. > > If you still want to add a subtarget to use different CFLAGS, you need > to prove (with hard numbers) that it has a meaningful impact and is more > than just a potentially pointless micro-optimization. > > - Felix
And I’d like to do that. (a) we’d need to decide what benchmarks are meaningful… are the openssl encryption benchmarks sufficient? (b) but even before that, it’s not clear that the CPU_TYPE target makefile machinery has ever worked correctly… so we’d need to remedy that. -Philip _______________________________________________ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
