wasi-sdk
wasi-sdk copied to clipboard
WebAssembly
Add binary release for Linux on Apple Silicon M1
I think an increasing number of us are running on Linux on Apple Silicon with Parallels in a VM, and support would be great. I can always build from source, but I would guess there will be plenty of others who will run into this.
There is already an issue for this on MacOS #221.
Instructions for building from source are in the README: https://github.com/WebAssembly/wasi-sdk
I don't know of anyone working on a linux arm64 build. We build our images using github CI I guess it depends on when we can actions runing on arm64 linux.. or somebody could figure out to do a cross-compile on x86?
followup, needed
On Thu, 18 Aug 2022 at 11.50 Sam Clegg @.***> wrote:
Instructions for building from source are in the README: https://github.com/WebAssembly/wasi-sdk
I don't know of anyone working on a linux arm64 build. We build our images using github CI I guess it depends on when we can actions runing on arm64 linux.. or somebody could figure out to do a cross-compile on x86?
— Reply to this email directly, view it on GitHub https://github.com/WebAssembly/wasi-sdk/issues/236#issuecomment-1219011938, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC6MWVJR55QMOM6GCOWEYTDVZWXHPANCNFSM5WGVR3RQ . You are receiving this because you commented.Message ID: @.***>
https://github.blog/changelog/2022-08-09-github-actions-self-hosted-runners-now-support-apple-m1-hardware/
+1
In my case, I'm running a container on top of an Apple M1, so I'm missing the -linux-arm64 flavor.
More info:
Building from sources from a linux-arm64 container (tried with ubuntu:22.04 and bitnami/minideb) fails.
In both containers, the error message is:
/usr/bin/cc -DGTEST_HAS_RTTI=0 -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I/home/ubuntu/Home/Workspace/VMware/wasi-sdk/build/llvm/lib/Support/BLAKE3 -I/home/ubuntu/Home/Workspace/VMware/wasi-sdk/src/llvm-project/llvm/lib/Support/BLAKE3 -I/home/ubuntu/Home/Workspace/VMware/wasi-sdk/build/llvm/include -I/home/ubuntu/Home/Workspace/VMware/wasi-sdk/src/llvm-project/llvm/include -fPIC -Os -DNDEBUG -mavx512vl -MD -MT lib/Support/BLAKE3/CMakeFiles/LLVMSupportBlake3.dir/blake3_avx512_x86-64_unix.S.o -MF lib/Support/BLAKE3/CMakeFiles/LLVMSupportBlake3.dir/blake3_avx512_x86-64_unix.S.o.d -o lib/Support/BLAKE3/CMakeFiles/LLVMSupportBlake3.dir/blake3_avx512_x86-64_unix.S.o -c /home/ubuntu/Home/Workspace/VMware/wasi-sdk/src/llvm-project/llvm/lib/Support/BLAKE3/blake3_avx512_x86-64_unix.S
cc: error: unrecognized command-line option '-mavx512vl'
Any insight about how to workaround this? Thanks!
UPDATE: Issue described in #294
This issue is open for anyone to work on: from what I can tell, it would involve modifying the CI workflow, the CI scripts, and (perhaps) the Makefile.
I started looking into this, locally the build went fine, and editing the ci workflows/scripts is straight forward. I think the bigger potential hold back here is where it would run. From what I can tell, Github currently only provides m1 runners in their "larger" runners offerings, which can get pricey. Other option could be self-hosted runners. Thoughts?
