[RFC PATCH v2 08/12] ipa: libipa: vector: Add element-wise min() and max() functions

[Laurent Pinchart]

On Mon, Nov 18, 2024 at 02:41:54PM +0100, Milan Zamazal wrote:
> Laurent Pinchart <laurent.pinchart at ideasonboard.com> writes:
> 
> > Add functions to calculate the element-wise minimum and maximum of two
> > vectors or of a vector and a scalar. This will be used in algorithm
> > implementations.
> >
> > Signed-off-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> > ---
> >  src/ipa/libipa/vector.cpp | 28 ++++++++++++++++++++++++++++
> >  src/ipa/libipa/vector.h   | 20 ++++++++++++++++++++
> >  2 files changed, 48 insertions(+)
> >
> > diff --git a/src/ipa/libipa/vector.cpp b/src/ipa/libipa/vector.cpp
> > index a45f08fde493..14816cdb5d09 100644
> > --- a/src/ipa/libipa/vector.cpp
> > +++ b/src/ipa/libipa/vector.cpp
> > @@ -183,6 +183,34 @@ namespace ipa {
> >   * \return This vector
> >   */
> >  
> > +/**
> > + * \fn Vector::min(const Vector &other) const
> > + * \brief Calculate the minimum of this vector and \a other element-wise
> > + * \param[in] other The other vector
> > + * \return The element-wise minimum of this vector and \a other
> > + */
> > +
> > +/**
> > + * \fn Vector::min(T scalar) const
> > + * \brief Calculate the minimum of this vector and \a scalar element-wise
> > + * \param[in] scalar The scalar
> > + * \return The element-wise minimum of this vector and \a scalar
> > + */
> > +
> > +/**
> > + * \fn Vector::max(const Vector &other) const
> > + * \brief Calculate the maximum of this vector and \a other element-wise
> > + * \param[in] other The other vector
> > + * \return The element-wise maximum of this vector and \a other
> > + */
> > +
> > +/**
> > + * \fn Vector::max(T scalar) const
> > + * \brief Calculate the maximum of this vector and \a scalar element-wise
> > + * \param[in] scalar The scalar
> > + * \return The element-wise maximum of this vector and \a scalar
> > + */
> > +
> >  /**
> >   * \fn Vector::dot(const Vector<T, Rows> &other) const
> >   * \brief Compute the dot product
> > diff --git a/src/ipa/libipa/vector.h b/src/ipa/libipa/vector.h
> > index 722ffc57ea45..949dc4650aa4 100644
> > --- a/src/ipa/libipa/vector.h
> > +++ b/src/ipa/libipa/vector.h
> > @@ -154,6 +154,26 @@ public:
> >  		return apply(scalar, [](T a, T b) { return a / b; });
> >  	}
> >  
> > +	constexpr Vector min(const Vector &other) const
> > +	{
> > +		return apply(*this, other, [](T a, T b) { return std::min(a, b); });
> > +	}
> > +
> > +	constexpr Vector min(T scalar) const
> > +	{
> > +		return apply(*this, scalar, [](T a, T b) { return std::min(a, b); });
> > +	}
> > +
> > +	constexpr Vector max(const Vector &other) const
> > +	{
> > +		return apply(*this, other, [](T a, T b) { return std::max(a, b); });
> > +	}
> > +
> > +	constexpr Vector max(T scalar) const
> > +	{
> > +		return apply(*this, scalar, [](T a, T b) -> T { return std::max(a, b); });
> > +	}
> > +
> 
> I wonder whether there could be a way to use std::max and std::min
> directly, without wrapping them by lambdas.  I don't know so:

Unfortunately that doesn't seem to be the case.

> Reviewed-by: Milan Zamazal <mzamazal at redhat.com>
> 
> >  	constexpr T dot(const Vector<T, Rows> &other) const
> >  	{
> >  		T ret = 0;
> 

-- 
Regards,

Laurent Pinchart