Skip to content

Remove prefer-inline for non-trivial Matrix functions #347

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 16, 2025
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Next Next commit
Remove prefer-inline for non-trivial Matrix functions
  • Loading branch information
kevmoo committed May 13, 2025
commit 30f558bb8176f0ec7616aec06649dc9b7dcbd5b2
11 changes: 1 addition & 10 deletions lib/src/vector_math/matrix4.dart
Original file line number Diff line number Diff line change
Expand Up @@ -691,10 +691,7 @@ class Matrix4 {
}

/// Translate this matrix by x, y, z, w.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void translateByDouble(double tx, double ty, double tz, double tw) {
void translateByDouble(double tx, double ty, double tz, [double tw = 1.0]) {
Copy link
Member

Choose a reason for hiding this comment

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

Optional parameters compliate calling conventions, the function now needs to check every time whether tw was called or not. It's a good idea to avoid these in performance sensitive code, though I have no evidence that it matters in this particular case. We should benchmark and measure.

You can observe the difference by compiling to Wasm before and after and checking the code for this function. Make sure to call it with and without the optional parameter otherwise TFA will optimize it into a version that always takes tw or never takes it.

Alternatively you can inline it always, in which case you get optimized code with no "is tw passed" checks, as in each call site you'll either pass it or not, and generate the code based on that.

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Should I drop em? Please give me code 😄

Choose a reason for hiding this comment

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

Didn't I give you the code already in first round of review?

This usage of optional parameter isn't going to improve things - most likely the compiler will just make the call site implicitly pass 1.0 where before your changes this was done explicitly.

So it suffers from the same problem I outlined before, that even though a parameter is 1.0 constant, we don't inline, so we do unnecessary extra work.

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Happy to revert this change then.

I'm confusing about the private function then.

final t1 = _m4storage[0] * tx +
_m4storage[4] * ty +
_m4storage[8] * tz +
Expand Down Expand Up @@ -760,9 +757,6 @@ class Matrix4 {
}

/// Multiply this by a translation from the left.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void leftTranslateByDouble(double tx, double ty, double tz, double tw) {
// Column 1
final r1 = _m4storage[3];
Expand Down Expand Up @@ -941,9 +935,6 @@ class Matrix4 {
}

/// Scale this matrix.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void scaleByDouble(double sx, double sy, double sz, double sw) {
_m4storage[0] *= sx;
_m4storage[1] *= sx;
Expand Down
11 changes: 1 addition & 10 deletions lib/src/vector_math_64/matrix4.dart
Original file line number Diff line number Diff line change
Expand Up @@ -691,10 +691,7 @@ class Matrix4 {
}

/// Translate this matrix by x, y, z, w.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void translateByDouble(double tx, double ty, double tz, double tw) {
void translateByDouble(double tx, double ty, double tz, [double tw = 1.0]) {
final t1 = _m4storage[0] * tx +
_m4storage[4] * ty +
_m4storage[8] * tz +
Expand Down Expand Up @@ -760,9 +757,6 @@ class Matrix4 {
}

/// Multiply this by a translation from the left.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void leftTranslateByDouble(double tx, double ty, double tz, double tw) {
// Column 1
final r1 = _m4storage[3];
Expand Down Expand Up @@ -941,9 +935,6 @@ class Matrix4 {
}

/// Scale this matrix.
@pragma('wasm:prefer-inline')
@pragma('vm:prefer-inline')
@pragma('dart2js:prefer-inline')
void scaleByDouble(double sx, double sy, double sz, double sw) {
_m4storage[0] *= sx;
_m4storage[1] *= sx;
Expand Down
2 changes: 1 addition & 1 deletion lib/src/vector_math_64/quad.dart
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@ class Quad {
/// The third point of the quad.
Vector3 get point2 => _point2;

/// The third point of the quad.
/// The fourth point of the quad.
Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Just a missed typing thing here...

Vector3 get point3 => _point3;

/// Create a new, uninitialized quad.
Expand Down
Loading