Hans Kristian Rosbach

RPI5 Aarch64 ### Develop Dec9 ``` Level Comp Comptime min/avg/max/stddev Decomptime min/avg/max/stddev Compressed size 1 48.164% 0.2779/0.2901/0.2940/0.0043 0.0982/0.1126/0.1178/0.0053 21,937,619 2 39.504% 0.4891/0.5038/0.5083/0.0044 0.0979/0.1092/0.1151/0.0048 17,993,257 3 37.948% 0.5918/0.6071/0.6117/0.0046 0.0961/0.1052/0.1111/0.0046 17,284,459 4...

@neurolabusc I am curious, what do the zlib-ng numbers look like when `WITH_OPTIM=OFF`? I'm also curious about whether there is any significant change with `WITH_NATIVE_INSTRUCTIONS=OFF` (and thus without LTO too),...

Oh, forgot to answer your question regarding minigzip using what zlib-ng api. Minigzip actually does both compat and native apis, and switches depending on what api you are building zlib-ng...

Hm, so for decompression, there is effectively nothing to gain from neither WITH_OPTIM or WITH_NATIVE_INSTRUCTIONS, in fact they rather seem to hurt performance a little according to those numbers. On...

@ccawley2011 I notice you work a lot on ARM stuff, you don't happen to have any ARM Apple systems, do you?

Interesting stuff @neurolabusc @KungFuJesus You don't have access to any such Apple M1-M4 systems, do you? @ccawley2011 does, if you or anyone else wants to try to implement an optimized...

While I see the benefit of auto-detecting cc as well, I do not see any reasoning for why you also change the default. Doing that will cause a potentially big...

I had another look, and I have no requests for changes at this point. I do want to do some benchmarking though, but I'll have to wait for my migraine...

> You should split these into separate PRs I agree. This was also meant to say it was a draft, but I forgot to set the title.

Having gone through the exercise of making this, I don't actually feel like this is what we want. The gain is currently solely a small library size reduction, but the...