Skip to content

RISCV64-CI: Update toolchain to unbreak qemu-riscv; add B/HFLOAT16 to VL256 target#5494

Closed
martin-frbg wants to merge 3 commits intoOpenMathLib:developfrom
martin-frbg:riscv-qemu
Closed

RISCV64-CI: Update toolchain to unbreak qemu-riscv; add B/HFLOAT16 to VL256 target#5494
martin-frbg wants to merge 3 commits intoOpenMathLib:developfrom
martin-frbg:riscv-qemu

Conversation

@martin-frbg
Copy link
Copy Markdown
Collaborator

No description provided.

@ChipKerchner
Copy link
Copy Markdown
Contributor

Thanks Martin!

@martin-frbg
Copy link
Copy Markdown
Collaborator Author

weird error where it apparently fails to include a valid time.h from common.h , but only in the compare_ tests

*****************************************************************************/
#include <stdio.h>
#include <stdint.h>
#include <time.h>
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why would only SHGEMV need this? Shouldn't the others also need it?

@martin-frbg
Copy link
Copy Markdown
Collaborator Author

I'm just trying if it fixes compilation for that one file... time.h should get included by common.h, so maybe that nightly build of the toolchain is incomplete/broken ? (The added -std=c11 has been there a few months)

@ChipKerchner
Copy link
Copy Markdown
Contributor

BTW, it is recommended that GCC be at least 15.0 for vector RISC-V. Probably 19+ on LLVM

@martin-frbg
Copy link
Copy Markdown
Collaborator Author

Idk how authoritative this riscv-collab project is actually... besides being endorsed by RISC-V International

@martin-frbg martin-frbg closed this Oct 9, 2025
@martin-frbg
Copy link
Copy Markdown
Collaborator Author

Seems the problem is that (at least) that toolchain does not define __USE_POSIX199309 although it appears to define POSIX_MONOTONIC_CLOCK, so the clock_gettime() function is undefined although time.h was included.

@martin-frbg
Copy link
Copy Markdown
Collaborator Author

Further to this, seems to be a glibc oddity where one has to define POSIX_C_SOURCE to 199309 to enable certain POSIX functions (or alternatively, use std=gnu11 instead of c11)

@ChipKerchner
Copy link
Copy Markdown
Contributor

Maybe you are using an older version of glibc?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants