Hacker News new | past | comments | ask | show | jobs | submit login

I’m glad that RISC-V SBCs are starting to become usable.



Depends on what you mean by usable. If you don't need long term support, strong software ecosystem support, and performance parity, it's great. It'll be a few more years before everything comes together, but hopefully not too many.


>If you don't need

Reads like FUD, as you go on to list a bunch of items that RISC-V actually already delivers.

Point per point:

>long term support

Is achieved via upstreaming drivers[0] and providing documentation[1], something that e.g. StarFive is doing much better than Raspberry Pi ever has.

>strong software ecosystem support

RISC-V is rapidly building the strongest ecosystem.

>performance parity

JH7110 SoC used in boards like VisionFive 2 provides CPU performance between Raspberry Pi 3b and 4, at much lower power consumption.

TH1520 SoC used in boards like Sipeed Lichee Pi4A[2] provides performance above Raspberry Pi 4.

Both SoCs provide faster GPU (JH7110 is 4x that of Pi 4, TH1520 is faster), better hardware video codec blocks, cryptography acceleration, faster memory interface, faster I/O outside of the SoC and otherwise better and more built in peripherals.

Note: Raspberry Pi have been used as reference points as they are, by far, the most popular ARM SBCs.

>It'll be a few more years

As proven above, it's already there against the Raspberry Pi line.

But next year it'll be better, as RISC-V will finally compete with the fastest cores available. This is based solely on what's already announced (Ascalon, Veyron, P570 and so on).

RISC-V enables the best processors.

0. https://rvspace.org/en/project/JH7110_Upstream_Plan

1. https://doc-en.rvspace.org/index.html

2. https://news.ycombinator.com/item?id=37201754


The RISC V fanboism is rampant here with bunch of them repeating their marketing slogan in the comments. There is nothing inevitable about RISC V. If anything it has lot of mountains to climb before it can be considered a challenge to the existing instruction set. It may or it may not do that. Being open does not automagically make it happen.


"Being open does not automagically make it happen." Absolutely true. Many many "open" things have never gained traction and have fallen by the wayside. It was a plausible argument about RISC-V in 2016. It is not in 2023, when it has already happened.

Lots of companies have made their decisions, projects are kicking off -- for example Samsung just started porting .NET and Tizen to RISC-V for use in their future TVs and other products. LG similarly. It will take five years for the products resulting from that to emerge, but they are coming.

The RISC-V ISA didn't even formally exist as a frozen spec a little over four years ago. Ratification in July 2019 was the starting gun for many to start projects. The results of that are just emerging in the last months -- the VisionFive 2, tne Lichee Pi 4A, the 64 core Milk-V Pioneer.

A lot more things have gone into the pipeline since then, and will be emerging in the next two to three years. Things up to around Apple M1 performance ... from multiple companies.


>If anything it has lot of mountains to climb before it can be considered a challenge to the existing instruction set.

That's not what ARM thinks[0].

>There is nothing inevitable about RISC V.

As it turns out, the decisions have long been made, and RISC-V has been chosen.

0. https://thechipletter.substack.com/p/risc-v-part-2-ambitious...




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: