Introduce really rough linalg matrix proposal #404
Conversation
| necessarily what they do. One big part of this proposal is to take 5 steps back | ||
| and talk about what they do: linear algebra. | ||
|
|
||
| ## Motivation |
There was a problem hiding this comment.
Worth calling out relationship to https://github.com/microsoft/hlsl-specs/blob/main/proposals/0031-hlsl-vector-matrix-operations.md?
| Matrix operator+(Matrix); | ||
| Matrix operator-(Matrix); | ||
| Matrix operator*(Matrix); | ||
| Matrix operator/(Matrix); |
There was a problem hiding this comment.
What does operator/ do on a matrix?
There was a problem hiding this comment.
elementwise division, will add comments
There was a problem hiding this comment.
All of the operators are element wise aligning with our conventions for vector and matrix types.
| template <typename T> | ||
| requires ArithmeticScalar<T> | ||
| Matrix operator+(T); | ||
| template <typename T> | ||
| requires ArithmeticScalar<T> | ||
| Matrix operator-(T); | ||
| template <typename T> | ||
| requires ArithmeticScalar<T> | ||
| Matrix operator*(T); | ||
| template <typename T> | ||
| requires ArithmeticScalar<T> | ||
| Matrix operator/(T); |
There was a problem hiding this comment.
Yea, I'll add comments. This aligns with our convention for matrix and vector operators for other builtin types
| requires ArithmeticScalar<T> | ||
| Matrix operator/(T); | ||
|
|
||
| static Matrix Splat(ElTy Val); |
There was a problem hiding this comment.
| static Matrix Splat(ElTy Val); | |
| static Matrix Splat(ComponentTy Val); |
There was a problem hiding this comment.
Actually, I suspect that this code might have used all of T, ElTy and ComponentTy at some point to refer to the same thing?
There was a problem hiding this comment.
Oh... yea I crossed some template wires. Let me do fix that.
|
|
||
| ## Detailed design | ||
|
|
||
| ## Outstanding Questions |
There was a problem hiding this comment.
- Support for packed types?
- Support for other number formats that aren't natively supported by HLSL?
Co-authored-by: Damyan Pepper <[email protected]>
This is more as a straw man to start discussion than an actual proposal we'll move forward with as it is. This is a clear gap in the Direct3D and core HLSL feature sets and we should seek to address it.