One of the low-level features of C++ is the ability to specify the precise alignment of objects in memory to take maximum advantage of a specific hardware architecture. By default, the compiler aligns class and struct members on their size value: bool and char on 1-byte boundaries, short on 2-byte boundaries, int, long, and float on 4-byte boundaries, and long long, double, and long double on 8-byte boundaries.

In most scenarios, you never have to be concerned with alignment because the default alignment is already optimal. In some cases, however, you can achieve significant performance improvements, or memory savings, by specifying a custom alignment for your data structures. Before Visual Studio 2015 you could use the Microsoft-specific keywords __alignof and declspec(alignas) to specify an alignment greater than the default. Starting in Visual Studio 2015 you should use the C++11 standard keywords alignof and alignas for maximum code portability. The new keywords behave in the same way under the hood as the Microsoft-specific extensions. The documentation for those extensions also applies to the new keywords. For more information, see __alignof Operator and align. The C++ standard doesn't specify packing behavior for alignment on boundaries smaller than the compiler default for the target platform, so you still need to use the Microsoft #pragma pack in that case.

Use the aligned_storage class for memory allocation of data structures with custom alignments. The aligned_union class is for specifying alignment for unions with non-trivial constructors or destructors.

Alignment and memory addresses

Alignment is a property of a memory address, expressed as the numeric address modulo a power of 2. For example, the address 0x0001103F modulo 4 is 3. That address is said to be aligned to 4n+3, where 4 indicates the chosen power of 2. The alignment of an address depends on the chosen power of 2. The same address modulo 8 is 7. An address is said to be aligned to X if its alignment is Xn+0.

CPUs execute instructions that operate on data stored in memory. The data are identified by their addresses in memory. A single datum also has a size. We call a datum naturally aligned if its address is aligned to its size. It's called misaligned otherwise. For example, an 8-byte floating-point datum is naturally aligned if the address used to identify it has an 8-byte alignment.

Compiler handling of data alignment

Compilers attempt to make data allocations in a way that prevents data misalignment.

For simple data types, the compiler assigns addresses that are multiples of the size in bytes of the data type. For example, the compiler assigns addresses to variables of type long that are multiples of 4, setting the bottom 2 bits of the address to zero.

The compiler also pads structures in a way that naturally aligns each element of the structure. Consider the structure struct x_ in the following code example:

C++Copy

struct x_
{
   char a;     // 1 byte
   int b;      // 4 bytes
   short c;    // 2 bytes
   char d;     // 1 byte
} bar[3];


The compiler pads this structure to enforce alignment naturally.

The following code example shows how the compiler places the padded structure in memory:

C++Copy

// Shows the actual memory layout
struct x_
{
   char a;            // 1 byte
   char _pad0[3];     // padding to put 'b' on 4-byte boundary
   int b;            // 4 bytes
   short c;          // 2 bytes
   char d;           // 1 byte
   char _pad1[1];    // padding to make sizeof(x_) multiple of 4
} bar[3];


Both declarations return sizeof(struct x_) as 12 bytes.

The second declaration includes two padding elements:


	
char _pad0[3] to align the int b member on a 4-byte boundary.
	
	

	
char _pad1[1] to align the array elements of the structure struct _x bar[3]; on a four-byte boundary.
	
The padding aligns the elements of bar[3] in a way that allows natural access.

The following code example shows the bar[3] array layout:

OutputCopy

adr offset   element
------   -------
0x0000   char a;         // bar[0]
0x0001   char pad0[3];
0x0004   int b;
0x0008   short c;
0x000a   char d;
0x000b   char _pad1[1];

0x000c   char a;         // bar[1]
0x000d   char _pad0[3];
0x0010   int b;
0x0014   short c;
0x0016   char d;
0x0017   char _pad1[1];

0x0018   char a;         // bar[2]
0x0019   char _pad0[3];
0x001c   int b;
0x0020   short c;
0x0022   char d;
0x0023   char _pad1[1];


alignof and alignas

The alignas type specifier is a portable, C++ standard way to specify custom alignment of variables and user defined types. The alignof operator is likewise a standard, portable way to obtain the alignment of a specified type or variable.

Example

You can use alignas on a class, struct or union, or on individual members. When multiple alignas specifiers are encountered, the compiler will choose the strictest one, (the one with the largest value).

C++Copy

// alignas_alignof.cpp
// compile with: cl /EHsc alignas_alignof.cpp
#include <iostream>

struct alignas(16) Bar
{
    int i;       // 4 bytes
    int n;      // 4 bytes
    alignas(4) char arr[3];
    short s;          // 2 bytes
};

int main()
{
    std::cout << alignof(Bar) << std::endl; // output: 16
}

刚一开始理解错了题意，当我wa了两次之后，重新思考，发现在队列里的士兵可以从左边或者从右边没有比他高的即可。
由此，此题和合唱队形差不多，求一个最长升序列，以及一个最长下降序列，枚举中间的兵和他左边最长下降序列之和。
时间复杂度为o(n^2)
#include <iostream>
#include <cstdio>
#include <cstring>
#include <cmath>
#include <algorithm>

using namespace std;


double a[1010];
int inc[1010],de[1010];


int main(){
    int len;
    while(~scanf("%d",&len))
    {
        for(int i=1;i<=len;i++)
        scanf("%lf",&a[i]);

        memset(inc,0,sizeof(inc));
        memset(de,0,sizeof(de));

        for(int i=1;i<=len;i++)
        {
            for(int j=0;j<i;j++)
            {
                if(inc[i]<=inc[j]&&a[i]>a[j])
                inc[i]=inc[j]+1;
            }
        }
        for(int i=len;i>0;i--)
        {
            for(int j=i+1;j<=len+1;j++)
            {
                if(de[i]<=de[j]&&a[i]>a[j])
                {
                    de[i]=de[j]+1;
                }
            }
        }
        int ans=0,tem;
        for(int i=1;i<=len;i++)
        {
            tem=i;
            for(int j=i+1;j<=len;j++)
            {
                if(de[tem]<=de[j])
                tem=j;
            }
            if(tem==i)
            ans=ans<(de[tem]+inc[i]-1)?(de[tem]+inc[i]-1):ans;
            else
            ans=ans<(de[tem]+inc[i])?(de[tem]+inc[i]):ans;
        }
        printf("%d\n",len-ans);
    }

    return 0;
}

reads比对到参考序列后，bam文件中会有2048、2064这样的flag，表示supplementary alignment 。 为了理解这个概念，可能需要以下知识。



Linear Alignment

An alignment of a read to a single reference sequence that may include insertions, deletions, skips and clipping, but may not include direction changes (i.e. one portion of the alignment on forward strand and another portion of alignment on reverse strand). 1



Chimeric Alignment

An alignment of a read that cannot be represented as a linear alignment. Typically, one of the linear alignments in a chimeric alignment is considered the “representative” alignment, and the others are called “supplementary” and are distinguished by the supplementary alignment flag. 1

Chimeric reads are indicative of structural variation in DNA-seq and it may indicate the presence of chimeric genes in RNA-seq. 2

In short, chimeric reads can be split in to two or more parts, each part would be mapped to reference(it’s not hard-clipped), the total length of the mapped part is longger than read length. 3



Representative alignment

A chimeric alignment that is represented as a set of linear alignments that do not have large overlaps typically has one linear alignment that is considered the representative alignment.

I don’t understand representative alignment with the word “representative” in my mother tongue and could not find more information(figure) about it. One read can align to multiple positions, we can find one alignmnet position which sequence do not have large overlaps, it called representative alighment, for other alignment positions, we called them supplementary alignment.

It seems that GATK can realignment those representative reads to the correctly position via RealignerTargetCreator and IndelRealigner. (WARNING: I am not quite sure if I understand this correctly. If someone could help me, please leave me a message below, thanks, thanks.)



Supplementary Alignment

A chimeric reads but not a representative reads.



Primary Alignment and Secondary Alignment

A read may map ambiguously to multiple locations, e.g. due to repeats. Only one of the multiple read alignments is considered primary, and this decision may be arbitrary. All other alignments have the secondary alignment flag.