blob: 9f9356a3451d0c913f0a760ec041d17a4f537380 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
|
#pragma once
#ifndef __CVTT_AGGREGATEDERROR_H__
#define __CVTT_AGGREGATEDERROR_H__
#include "ConvectionKernels_ParallelMath.h"
namespace cvtt
{
namespace Internal
{
template<int TVectorSize>
class AggregatedError
{
public:
typedef ParallelMath::UInt16 MUInt16;
typedef ParallelMath::UInt31 MUInt31;
typedef ParallelMath::Float MFloat;
AggregatedError()
{
for (int ch = 0; ch < TVectorSize; ch++)
m_errorUnweighted[ch] = ParallelMath::MakeUInt31(0);
}
void Add(const MUInt16 &channelErrorUnweighted, int ch)
{
m_errorUnweighted[ch] = m_errorUnweighted[ch] + ParallelMath::ToUInt31(channelErrorUnweighted);
}
MFloat Finalize(uint32_t flags, const float channelWeightsSq[TVectorSize]) const
{
if (flags & cvtt::Flags::Uniform)
{
MUInt31 total = m_errorUnweighted[0];
for (int ch = 1; ch < TVectorSize; ch++)
total = total + m_errorUnweighted[ch];
return ParallelMath::ToFloat(total);
}
else
{
MFloat total = ParallelMath::ToFloat(m_errorUnweighted[0]) * channelWeightsSq[0];
for (int ch = 1; ch < TVectorSize; ch++)
total = total + ParallelMath::ToFloat(m_errorUnweighted[ch]) * channelWeightsSq[ch];
return total;
}
}
private:
MUInt31 m_errorUnweighted[TVectorSize];
};
}
}
#endif
|