Commit graph

10 commits

Author SHA1 Message Date
serenityosrocks
4a6a7cf3c8 AK: Add generic sincos solution for non-x86 platforms 2022-04-03 00:31:41 +01:00
Hendiadyoin1
b1db1e4e4f AK: Add rsqrt and a SSE specific implementation for sqrt 2022-04-02 18:37:38 +02:00
Hendiadyoin1
cd21e03225 AK+Everywhere: Add sincos and use it in some places
Calculating sin and cos at once is quite a bit cheaper than calculating
them individually.
x87 has even a dedicated instruction for it: `fsincos`.
2022-03-15 11:39:42 +01:00
Gunnar Beutner
7a65bd8efb AK: Don't use x86 assembly when building for non-x86 targets
This allows Lagom to be built successfully for Apple M1.

Fixes #12644.
2022-02-19 13:13:22 +02:00
Hendiadyoin1
fbb798f98c AK: Move integral log2 and exp to IntegerMath.h 2022-02-06 17:52:33 +00:00
Pankaj Raghav
24c66b8442 AK: Fix log2 calculation for Integer
For a `N` bit integer X, the log2 calculation should be as follows:
(N - 1) - ctz(X)
2022-01-29 17:41:06 +02:00
Nick Johnson
08e4a1a4dc AK+Everywhere: Replace __builtin bit functions
In order to reduce our reliance on __builtin_{ffs, clz, ctz, popcount},
this commit removes all calls to these functions and replaces them with
the equivalent functions in AK/BuiltinWrappers.h.
2021-12-21 22:13:51 +01:00
Jelle Raaijmakers
dfbdd035da AK: Implement acos<T> correctly
This is a naive implementation based on the symmetry with `asin`.

Before, I'm not really sure what we were doing, but it was returning
wildly incorrect results.
2021-11-18 21:10:30 +01:00
Daniel Bertalan
5d32f543ec AK: Handle partial remainders
On x86, the `fprem` and `fmprem1` instructions may produce a 'partial
remainder', for which we should check by reading a FPU flag. If we don't
check for it, we may end up using values that are outside the expected
range of values.
2021-08-08 10:55:36 +02:00
Hendiadyoin1
c5f6ba6e71 AK: Introduce Math.h
This is to implement constexpr template based implementations for
mathematical functions

This also changes math.cpp to use these implementations.

Also adds a fastpath for floating point trucation for values smaller
than the signed 64 bit limit.
2021-07-19 16:34:21 +04:30