Skip to content

Commit 63364ab

Browse files
committed
riscv announcement
1 parent 216c14d commit 63364ab

2 files changed

Lines changed: 54 additions & 0 deletions

File tree

_posts/2025-03-12-dropping-riscv.md

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4,6 +4,9 @@ layout: post
44
excerpt_separator: <!--more-->
55
---
66

7+
**UPDATE March 20 2025:** The architecture is not being dropped
8+
for now after all. See the newer article for details.
9+
710
The next set of images will drop RISC-V support. The builder is
811
currently still going but within the next few days it will stop,
912
and the repositories will stay in place but frozen.
Lines changed: 51 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,51 @@
1+
---
2+
title: Not dropping RISC-V support after all (maybe)
3+
layout: post
4+
excerpt_separator: <!--more-->
5+
---
6+
7+
As circumstances have changed, we are not dropping RISC-V repos
8+
for the time being. Instead, newly rebuilt repositories are introduced,
9+
built on hardware, with tests.
10+
11+
This support is provisional for now, with the new builder still being
12+
evaluated to see how it holds up in the long term.
13+
14+
<!--more-->
15+
16+
## The situation now
17+
18+
Shortly after announcing the drop, we were offered remote access to
19+
a Milk-V Pioneer machine by [Zach van Rijn](https://zv.io) of Adélie
20+
Linux. This machine was originally intended for another purpose which
21+
never ended up materializing.
22+
23+
I proceeded to do a full world rebuild on this machine, after some
24+
environment setup to allow our infra bits to run. This world rebuild
25+
is now finished, and makes up the new repository.
26+
27+
For most part, it was relatively stable during the build (we had to
28+
build our own kernel to prevent the draft RVV 0.7 in the CPU from
29+
interfering, and there were two crashes but it was also under total
30+
continuous load the whole time).
31+
32+
The performance is fairly acceptable, though nowhere near my original
33+
idea of being similar to Cortex-A72; the cores are more comparable
34+
to Cortex-A55 in practical performance, especially since we have to
35+
disable vectors. As there is still 64 of them, most of the large
36+
projects build fairly fast (anything written in Rust builds very
37+
slowly, however).
38+
39+
By now, the original repositories have been replaced, and the new
40+
machine is plugged into the infrastructure. Do keep this in mind when
41+
upgrading existing installations, and use the `--available` flag with
42+
`apk` (every package in your system will be reinstalled).
43+
44+
Either way, we will continue to monitor the builds and see how the
45+
new machine holds up. If it works well, it will stay; if significant
46+
issues arise, we might end up dropping the architecture after all,
47+
at least until something significantly better is available.
48+
49+
The current repository is in the same tier as the LoongArch64 repo.
50+
The specifics are very similar - i.e. no LTO, tests on and enforced.
51+
The overall coverage is also fairly equivalent.

0 commit comments

Comments
 (0)