For whatever it's worth, you can get some very performant ARM and RISC V processors. Software support gaps are less of an issue for ARM considering how long it's been around for. The apple silicon macs are all arm based and seem to perform very well under specific scenarios and workloads.
But I've had some struggles recently with very obscure software packages not playing nicley on my raspberry 5 with x86 emulation, so there are some definite hurdles still.
no worries. RISC V is an interesting, promising yet emerging platform.
You may be able to use a RISC V system as a general computer but there's likely to be gaps in terms of software support.
I suppose you could try and see how far can you get by using a Raspberry Pi (or similar device) as your primary computer as a sort of benchmark (bearing in mind that the RPi is ARM based, not RISC V)
With all of that said, I'm really looking forward to the day of high performance, general purpose RISC V PC systems.
That can be the case for most mainstream distros in expecteded platforms. You may find some quirks with RISC-V.
General package availability is fairly high but there's bound to be gaps in software you need. (You should be able to find this out in advance on a per-app or library basis). Projects like Box86 and FexEmu can maybe be applied here as well but that's another layer of complexity added to an already significant jump you're making.
Make the exploration of this arch a side project rather than a main goal for now. There are some very interesting SBCs available, the PineTab V looks pretty cool as well, but I'd by lying to you if I said you could depend on these devices as your primary system.
Very welcome! I originally neglected the notion that I tested this from a 'supported' ASIC, however.
I'm not sure how this will behave on NV33; you may need to employ the aforementioned env variable workaround for any luck, I'll try to find a link for it.
E1: I believe RX 6700XT (NV23) users set the following env var to spoof their device as NV21
HSA_OVERRIDE_GFX_VERSION=10.3.0
I'll see if there's one more suitable to your GPU.
These newer modules are lower profile than SODIMM, and do not carry the same frequency/ throughput and latency limitations. LPCAMM effectively eliminates the need to solder RAM to mobile platform main boards, though we'll see how vendors react.
AV1 decode is supported on the RTX 3000 series, encode + decode in the RTX 4000 series
For Intel Arc, AV1 encode + decode support is present on all Arc Alchemist GPUs,
For AMD, AV1 decode is on RX 6000 series, encode + decode on all RX 7000 series GPUs
As someone else has recommended, a low end Intel Arc alchemist GPU is pretty great for stuff like Jellyfin, very low price to entry for gfx accelerated AV1 transcoding.
No prob.
For whatever it's worth, you can get some very performant ARM and RISC V processors. Software support gaps are less of an issue for ARM considering how long it's been around for. The apple silicon macs are all arm based and seem to perform very well under specific scenarios and workloads.
But I've had some struggles recently with very obscure software packages not playing nicley on my raspberry 5 with x86 emulation, so there are some definite hurdles still.