14:47tertl8: hello folks
14:47tertl8: how about that darn asahi linux
14:50tertl8: nvidia is so snakey, they lock down their cards and then contribute to 'open source' in the form of python wrappers for cuda
14:57juri_: only the best.
18:58tertl8: im starting to consider getting a m1 mac
19:05diogenes_oftc: tertl8, you think apple is more open than nvidia?
19:06fdobridge: <gouz> When running the tessellation/clipping tests i still get an assertion:
19:06fdobridge: <gouz> "../src/compiler/nir/nir.h:2678: nir_src_comp_as_uint: Assertion `nir_src_is_const(src)' failed." when executing nir_lower_io pass.
19:06fdobridge: <gouz> This happens on dynamic index tests like: dEQP-VK.clipping.user_defined.clip_distance_dynamic_index.vert_tess.1
19:08fdobridge: <gouz> this does not happen on dynamic_index.vert tests
21:23HdkR: I mean, go for it. It's good hardware.
21:46tertl8: HdkR: it really is. im curious, have you guys looked at the RAPIDS api for cuda?
21:53tertl8: the only thing holding me back is spending $1000 on laptop and only getting 8GB ram
23:10karolherbst: I got the 24GB version myself
23:12karolherbst: and linux being linux it's kinda necessary, but that also depends. Webbrowsers tend to cache data depending on the total amount of RAM, so it's hard to tell how much you'd actually need