ObjFW  History of src/OFMatrix4x4.m of merge-in:668be94

Changing the alignment of OFVector4D globally (as was done previously)
would have technically been an ABI break in extremely rare cases.
However, since we only need the alignment for the methods added after
1.0, it's better to only have the alignment there and get back to full
ABI compatibility with 1.0. file: [0c5b86a89a] check-in: [1af54eb2c2] user: js, branch: trunk, size: 10345 [annotate] [blame] [check-ins using] [diff]

This should work now that objects are properly aligned. file: [67c177d7d9] check-in: [71237f1a49] user: js, branch: trunk, size: 10332 [annotate] [blame] [check-ins using] [diff]

This gives another nice speed improvement. file: [a4f9a5f812] check-in: [5edf0d083d] user: js, branch: trunk, size: 9943 [annotate] [blame] [check-ins using] [diff]

This new SSE1 implementation is better than the SSE4.1 implementation,
hence this also deletes the SSE4.1 implementation. file: [ac56c8b767] check-in: [cf955413ab] user: js, branch: trunk, size: 9770 [annotate] [blame] [check-ins using] [diff]

This results in a ~ 16% performance improvement on a Duron 750. file: [b31d5e9464] check-in: [3280466d35] user: js, branch: trunk, size: 9256 [annotate] [blame] [check-ins using] [diff]

The new 3DNow! implementation is now better than the Enhanced 3DNow!
implementation, hence this also deletes the Enhanced 3DNow!
implementation. file: [66a4a58f6e] check-in: [34b48a4208] user: js, branch: trunk, size: 8587 [annotate] [blame] [check-ins using] [diff]

The Clang alignment bug has been worked around in the previous commit,
so now it works. file: [52a22d3f60] check-in: [0eb97e4612] user: js, branch: trunk, size: 11699 [annotate] [blame] [check-ins using] [diff]

This requires the vectors to be 16 byte aligned. In order to achieve
this, the OFVector4D type is changed to have an alignment of 16 bytes.
However, this does *not* break ABI because the only method actually
requiring 16 byte alignment is -[transformVectors:count:], which was not
in ObjFW 1.0. Hence binaries compiled for ObjFW 1.0 have no 16 byte
alignment for OFVector4D, but also cannot ever call into any code that
needs it. (-[transformedVector:] calls into -[transformVectors:count:],
but creates a properly aligned copy that it passes.) file: [46f1ac4383] check-in: [7f304f573b] user: js, branch: trunk, size: 11520 [annotate] [blame] [check-ins using] [diff]

The outputs aren't consumed as they are used as inputs that can be
modified. file: [b42a0641ea] check-in: [f949f7775b] user: js, branch: trunk, size: 9582 [annotate] [blame] [check-ins using] [diff]

This way, `result` doesn't need to be initialized. file: [b494f24572] check-in: [d53c87e7bb] user: js, branch: trunk, size: 6634 [annotate] [blame] [check-ins using] [diff]

It's broken in older versions of Clang (e.g. Clang 12). file: [c5f86994f5] check-in: [d9af65de97] user: js, branch: trunk, size: 6546 [annotate] [blame] [check-ins using] [diff]

This should make vectorization easier. file: [ccf7b44cda] check-in: [cf4d6a3dfa] user: js, branch: trunk, size: 3699 [annotate] [blame] [check-ins using] [diff]

This allows writing matrices more naturally, without having to transpose
them in the source. file: [37363ee709] check-in: [975a812f36] user: js, branch: trunk, size: 5526 [annotate] [blame] [check-ins using] [diff]

A singleton for the identity matrix makes no sense since OFMatrix4x4 is
mutable. file: [81834eec17] check-in: [b924571763] user: js, branch: trunk, size: 4218 [annotate] [blame] [check-ins using] [diff]

For now, we're only interested in transformed points. file: [630e704e06] check-in: [db7991d07b] user: js, branch: trunk, size: 5642 [annotate] [blame] [check-ins using] [diff]

These will be useful for 3D graphics. file: [8c09d10573] check-in: [44eb35b212] user: js, branch: trunk, size: 5287 [annotate] [blame] [check-ins using]