Implement FEAT_SME_F8F32

[PATCH 00/10] Implement FEAT_SME_F8F32

Richard Henderson posted 10 patches 1 month, 1 week ago

Download series mbox

Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20260618041517.573469-1-richard.henderson@linaro.org

Maintainers: Peter Maydell <peter.maydell@linaro.org>, Pierrick Bouvier <pierrick.bouvier@oss.qualcomm.com>, Laurent Vivier <laurent@vivier.eu>, Helge Deller <deller@gmx.de>

target/arm/cpu-features.h        |  5 ++
target/arm/tcg/helper-fp8-defs.h |  4 ++
target/arm/tcg/vec_internal.h    | 31 +++++++++++
linux-user/aarch64/elfload.c     |  1 +
target/arm/tcg/cpu64.c           |  1 +
target/arm/tcg/fp8_helper.c      | 58 +++++++++++++++++++
target/arm/tcg/sme_helper.c      | 33 -----------
target/arm/tcg/translate-sme.c   | 96 ++++++++++++++++++++++++++++++++
docs/system/arm/emulation.rst    |  1 +
target/arm/tcg/sme.decode        | 29 ++++++++++
10 files changed, 226 insertions(+), 33 deletions(-)

Expand all Fold all

[PATCH 00/10] Implement FEAT_SME_F8F32

Posted by Richard Henderson 1 month, 1 week ago

Based-on: 20260618040718.572950-1-richard.henderson@linaro.org
("[PATCH v2 0/8] target/arm: Implement FEAT_SVE2p2")

This is (kinda-sorta) a prerequisite for (parts of) FEAT_SME_MOP4,
which is a prerequisite for FEAT_SME2p2.  It's fairly easy because
it's mostly re-using helpers from previous FP8 features.


r~


Richard Henderson (10):
  target/arm: Move tile_vslice_{index,offset} to vec_internal.h
  target/arm: Implement FMOPA (widening, 4-way fp8 to fp32)
  target/arm: Implement FMLALL (multiple and single vector)
  target/arm: Implement FMLALL (multiple vector)
  target/arm: Implement FMLALL (multiple and indexed vector)
  target/arm: Implement FDOT (multiple and single vector)
  target/arm: Implement FDOT (multiple vector)
  target/arm: Implement FDOT (multiple and indexed vector)
  target/arm: Implement FVDOTB, FVDOTT
  target/arm: Enable FEAT_SME_F8F32 for -cpu max

 target/arm/cpu-features.h        |  5 ++
 target/arm/tcg/helper-fp8-defs.h |  4 ++
 target/arm/tcg/vec_internal.h    | 31 +++++++++++
 linux-user/aarch64/elfload.c     |  1 +
 target/arm/tcg/cpu64.c           |  1 +
 target/arm/tcg/fp8_helper.c      | 58 +++++++++++++++++++
 target/arm/tcg/sme_helper.c      | 33 -----------
 target/arm/tcg/translate-sme.c   | 96 ++++++++++++++++++++++++++++++++
 docs/system/arm/emulation.rst    |  1 +
 target/arm/tcg/sme.decode        | 29 ++++++++++
 10 files changed, 226 insertions(+), 33 deletions(-)

-- 
2.43.0

Re: [PATCH 00/10] Implement FEAT_SME_F8F32

Posted by Peter Maydell 1 month ago

On Thu, 18 Jun 2026 at 05:15, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Based-on: 20260618040718.572950-1-richard.henderson@linaro.org
> ("[PATCH v2 0/8] target/arm: Implement FEAT_SVE2p2")
>
> This is (kinda-sorta) a prerequisite for (parts of) FEAT_SME_MOP4,
> which is a prerequisite for FEAT_SME2p2.  It's fairly easy because
> it's mostly re-using helpers from previous FP8 features.
>



Applied to target-arm.next, thanks.

-- PMM