Skip to content

Fix Clang model bug, improve SSE intrinsics code and add tests for q_math transform functions #1452

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.

Sign up for GitHub

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
slipher merged 5 commits into from
Dec 5, 2024
Merged

Fix Clang model bug, improve SSE intrinsics code and add tests for q_math transform functions #1452

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
4786300
Use operator = instead of TransCopy
slipher Dec 1, 2024
f96a37d
Remove some unused q_math transform_t functions
slipher Dec 2, 2024
5fbe900
Add unit tests for most q_math transform_t functions
slipher Dec 1, 2024
4a54fa1
SSE transform math: remove 2 unnecessary mask ops
slipher Dec 2, 2024
ac6c3fe
Rework SSE intrinsics to avoid Clang 19 bugs
slipher Dec 2, 2024
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
23 changes: 14 additions & 9 deletions src.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,6 +86,18 @@ if (DAEMON_PARENT_SCOPE_DIR)
set(COMMONLIST ${COMMONLIST} PARENT_SCOPE)
endif()

# Tests for code shared by engine and gamelogic
set(COMMONTESTLIST
${LIB_DIR}/tinyformat/TinyformatTest.cpp
${COMMON_DIR}/ColorTest.cpp
${COMMON_DIR}/FileSystemTest.cpp
${COMMON_DIR}/StringTest.cpp
${COMMON_DIR}/cm/unittest.cpp
${COMMON_DIR}/MathTest.cpp
${COMMON_DIR}/UtilTest.cpp
${ENGINE_DIR}/qcommon/q_math_test.cpp
)

set(RENDERERLIST
${ENGINE_DIR}/renderer/DetectGLVendors.cpp
${ENGINE_DIR}/renderer/DetectGLVendors.h
Expand Down Expand Up @@ -279,15 +291,8 @@ else()
)
endif()

# Tests for engine-lib
set(ENGINETESTLIST
${LIB_DIR}/tinyformat/TinyformatTest.cpp
${COMMON_DIR}/ColorTest.cpp
${COMMON_DIR}/FileSystemTest.cpp
${COMMON_DIR}/StringTest.cpp
${COMMON_DIR}/cm/unittest.cpp
${COMMON_DIR}/MathTest.cpp
${COMMON_DIR}/UtilTest.cpp
# Tests runnable for any engine variant
set(ENGINETESTLIST ${COMMONTESTLIST}
${ENGINE_DIR}/framework/CommandSystemTest.cpp
)

Expand Down
52 changes: 6 additions & 46 deletions src/engine/qcommon/q_math.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3199,39 +3199,19 @@ void TransInit( transform_t *t )
t->scale = 1.0f;
}

// copy a transform
void TransCopy( const transform_t *in, transform_t *out )
{
memcpy( out, in, sizeof( transform_t ) );
}

// apply a transform to a point
void TransformPoint( const transform_t *t, const vec3_t in, vec3_t out )
{
QuatTransformVector( t->rot, in, out );
VectorScale( out, t->scale, out );
VectorAdd( out, t->trans, out );
}
// apply the inverse of a transform to a point
void TransformPointInverse( const transform_t *t, const vec3_t in, vec3_t out )
{
VectorSubtract( in, t->trans, out );
VectorScale( out, 1.0f / t->scale, out );
QuatTransformVectorInverse( t->rot, out, out );
}

// apply a transform to a normal vector (ignore scale and translation)
void TransformNormalVector( const transform_t *t, const vec3_t in, vec3_t out )
{
QuatTransformVector( t->rot, in, out );
}
// apply the inverse of a transform to a normal vector (ignore scale
// and translation)
void TransformNormalVectorInverse( const transform_t *t, const vec3_t in,
vec3_t out )
{
QuatTransformVectorInverse( t->rot, in, out );
}

// initialize a transform with a pure rotation
void TransInitRotationQuat( const quat_t quat, transform_t *t )
Expand All @@ -3240,16 +3220,6 @@ void TransInitRotationQuat( const quat_t quat, transform_t *t )
VectorClear( t->trans );
t->scale = 1.0f;
}
void TransInitRotation( const vec3_t axis, float angle, transform_t *t )
{
float sa = sinf( 0.5f * angle );
float ca = cosf( 0.5f * angle );
quat_t q;

VectorScale( axis, sa, q );
q[3] = ca;
TransInitRotationQuat( q, t );
}
// initialize a transform with a pure translation
void TransInitTranslation( const vec3_t vec, transform_t *t )
{
Expand Down Expand Up @@ -3292,16 +3262,6 @@ void TransAddRotationQuat( const quat_t quat, transform_t *t )
QuatMultiply2( tmp, t->rot );
QuatCopy( tmp, t->rot );
}
void TransAddRotation( const vec3_t axis, float angle, transform_t *t )
{
float sa = sinf( 0.5f * angle );
float ca = cosf( 0.5f * angle );
quat_t q;

VectorScale( axis, sa, q );
q[3] = ca;
TransAddRotationQuat( q, t );
}

// add a scale to the start of an existing transform
void TransInsScale( float factor, transform_t *t )
Expand All @@ -3319,9 +3279,9 @@ void TransAddScale( float factor, transform_t *t )
void TransInsTranslation( const vec3_t vec, transform_t *t )
{
vec3_t tmp;

TransformPoint( t, vec, tmp );
VectorAdd( t->trans, tmp, t->trans );
QuatTransformVector( t->rot, vec, tmp );
VectorScale( tmp, t->scale, tmp );
VectorAdd( tmp, t->trans, t->trans );
}

// add a translation at the end of an existing transformation
Expand All @@ -3334,7 +3294,7 @@ void TransAddTranslation( const vec3_t vec, transform_t *t )
void TransCombine( const transform_t *a, const transform_t *b,
transform_t *out )
{
TransCopy( a, out );
*out = *a;

TransAddRotationQuat( b->rot, out );
TransAddScale( b->scale, out );
Expand All @@ -3345,15 +3305,15 @@ void TransCombine( const transform_t *a, const transform_t *b,
void TransInverse( const transform_t *in, transform_t *out )
{
quat_t inverse;
static transform_t tmp; // static for proper alignment in QVMs
transform_t tmp;

TransInit( &tmp );
VectorNegate( in->trans, tmp.trans );
TransAddScale( 1.0f / in->scale, &tmp );
QuatCopy( in->rot, inverse );
QuatInverse( inverse );
TransAddRotationQuat( inverse, &tmp );
TransCopy( &tmp, out );
*out = tmp;
}

// lerp between transforms
Expand Down
208 changes: 208 additions & 0 deletions src/engine/qcommon/q_math_test.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,208 @@
/*
===========================================================================
Daemon BSD Source Code
Copyright (c) 2024, Daemon Developers
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the Daemon developers nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL DAEMON DEVELOPERS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
===========================================================================
*/

#include <gtest/gtest.h>
#include <gmock/gmock.h>

#include "common/Common.h"

// TODO: add transform tests with quats with negative real component
// the calculator I used always made it positive

namespace {

using ::testing::Pointwise;
using ::testing::FloatNear;

// Using a function for this in case we want to change how transform_t is set
// to avoid undefined behaviors with the union
transform_t MakeTransform(
const std::array<float, 4>& quat,
const std::array<float, 3>& translation,
float scale)
{
transform_t t;
Vector4Copy(quat, t.rot);
VectorCopy(translation, t.trans);
t.scale = scale;
return t;
}

void ExpectTransformEqual(
const transform_t& t,
const std::array<float, 4>& expectedQuat,
const std::array<float, 3>& expectedTranslation,
float expectedScale)
{
const char* base = reinterpret_cast<const char*>(&t);
std::array<float, 4> actualQuat;
std::array<float, 3> actualTranslation;
float actualScale;
memcpy(&actualQuat, base + offsetof(transform_t, rot), sizeof(actualQuat));
memcpy(&actualTranslation, base + offsetof(transform_t, trans), sizeof(actualTranslation));
memcpy(&actualScale, base + offsetof(transform_t, scale), sizeof(actualScale));
EXPECT_THAT(actualQuat, Pointwise(FloatNear(1e-4), expectedQuat));
EXPECT_THAT(actualTranslation, Pointwise(FloatNear(1e-3), expectedTranslation));
EXPECT_THAT(actualScale, FloatNear(1e-5, expectedScale));
}

TEST(QMathTransformTest, TransInit)
{
transform_t t;
TransInit(&t);
ExpectTransformEqual(t, {0, 0, 0, 1}, {0, 0, 0}, 1);
}

TEST(QMathTransformTest, TransformPoint)
{
transform_t t = MakeTransform(
{0.3155654, 0.121273, 0.7211766, 0.6046616}, {-11, -26, 55}, 1.2);
const vec3_t pointIn = {12, 19, -30};
vec3_t pointOut;
TransformPoint(&t, pointIn, pointOut);
const vec3_t expectedPoint = {-51.8073, -10.3551, 44.3603};
EXPECT_THAT(pointOut, Pointwise(FloatNear(0.001), expectedPoint));
}

TEST(QMathTransformTest, TransformNormalVector)
{
const transform_t t = MakeTransform(
{0.1641059, -0.6753088, 0.2253332, -0.6828266}, {235, 52, 42}, 8);
const vec3_t vector = { 0.48, 0.64, 0.6 };
vec3_t transformedVector;
TransformNormalVector(&t, vector, transformedVector);
const vec3_t expectedVector = {0.6462654,0.2383020,-0.7249505};
EXPECT_THAT(transformedVector, Pointwise(FloatNear(1e-6), expectedVector));
}

TEST(QMathTransformTest, TransInitRotationQuat)
{
const quat_t q = { -0.1423769, 0.5422508, 0.7235357, -0.402727 };
transform_t t;
TransInitRotationQuat(q, &t);
ExpectTransformEqual(t, {-0.1423769, 0.5422508, 0.7235357, -0.402727}, {0, 0, 0}, 1);
}

TEST(QMathTransformTest, TransInitTranslation)
{
const vec3_t translation = {-9, -32, -0.5};
transform_t t;
TransInitTranslation(translation, &t);
ExpectTransformEqual(t, {0, 0, 0, 1}, {-9, -32, -0.5}, 1);
}

TEST(QMathTransformTest, TransInitScale)
{
transform_t t;
TransInitScale(4.3, &t);
ExpectTransformEqual(t, {0, 0, 0, 1}, {0, 0, 0}, 4.3);
}

TEST(QMathTransformTest, TransInsRotationQuat)
{
const quat_t q = {0.3694065, -0.6226631, 0.1538449, -0.6724294};
transform_t t = MakeTransform(
{-0.5881667, -0.3748035, 0.6614826, 0.2757227}, {1.8, -1.8, -4}, 3);
TransInsRotationQuat(q, &t);
ExpectTransformEqual(t, {0.8515735, 0.4151890, 0.1023027, -0.3032735}, {1.8, -1.8, -4}, 3);
}

TEST(QMathTransformTest, TransAddRotationQuat)
{
const quat_t rot = {-0.4217441, -0.6010819, -0.3683214, -0.5702384};
transform_t t = MakeTransform(
{-0.1167623, -0.9282776, 0.3395844, 0.096694}, {12.5, 93.7, -14.1}, 2.9);
TransAddRotationQuat(rot, &t);
ExpectTransformEqual(t, {-0.5202203, 0.6574423, 0.09205336, -0.5372771},
{-5.823749, 47.07177, 82.97640}, 2.9);
}

TEST(QMathTransformTest, TransInsScale)
{
transform_t t = MakeTransform(
{-0.8678937, 0.0774454, -0.2533126, -0.4202326}, {25, 25, 17}, 4.6);
TransInsScale(3, &t);
ExpectTransformEqual(t,
{-0.8678937, 0.0774454, -0.2533126, -0.4202326}, {25, 25, 17}, 13.8);
}

TEST(QMathTransformTest, TransAddScale)
{
transform_t t = MakeTransform(
{-0.9884186, -0.0914321, 0.1088879, 0.053031}, {-9, -12, -70}, 6);
TransAddScale(1.2, &t);
ExpectTransformEqual(t,
{-0.9884186, -0.0914321, 0.1088879, 0.053031}, {-10.8, -14.4, -84}, 7.2);
}

TEST(QMathTransformTest, TransInsTranslation)
{
const vec3_t translation = {3.6, 20, 5.3};
transform_t t = MakeTransform(
{0.5002998, 0.3780266, 0.7196359, 0.2981948}, {-8, 28, 27}, 5.5);
TransInsTranslation(translation, &t);
ExpectTransformEqual(t,
{0.5002998, 0.3780266, 0.7196359, 0.2981948}, {7.592672, -7.848988, 135.6898}, 5.5);
}

TEST(QMathTransformTest, TransAddTranslation)
{
const vec3_t translation = { 18, -1.1, -6};
transform_t t = MakeTransform(
{0.3917585, -0.3411706, -0.1357459, 0.8436237}, {-23, 56, 52}, 4);
TransAddTranslation(translation, &t);
ExpectTransformEqual(t,
{0.3917585, -0.3411706, -0.1357459, 0.8436237}, {-5, 54.9, 46}, 4);
}

TEST(QMathTransformTest, TransCombine)
{
const transform_t left = MakeTransform(
{-0.5029552, 0.4741776, -0.5393551, 0.4809239}, {-3, -50, 7}, 1.9);
const transform_t right = MakeTransform(
{0.4461722, 0.8836869, 0.1121932, -0.0862585}, {-25, 11, 16}, 1.6);
transform_t combined;
TransCombine(&right, &left, &combined);
ExpectTransformEqual(combined,
{0.7877796, 0.1998672, -0.5555394, -0.1755917}, {29.72799, -5.378298, -16.55849}, 3.04);
}

TEST(QMathTransformTest, TransInverse)
{
const transform_t t = MakeTransform(
{0.4833702, -0.42157, -0.7551386, -0.1356377}, {2.5, 3.4, 1.4}, 2.5);
transform_t inverse;
TransInverse(&t, &inverse);
// This quat multiplied by -1 would also be acceptable
ExpectTransformEqual(inverse,
{-0.4833702, 0.42157, 0.7551386, -0.1356377}, {1.244436,1.155842,-0.5278334}, 0.4);
}

} // namespace
Loading