Ascetic/common.cu at main · NKU-EmbeddedSystem/Ascetic · GitHub

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
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
#include "common.cuh"

void getMaxPartitionSize(unsigned long &max_partition_size, unsigned long &totalSize, uint testNumNodes, float param,
                         int edgeSize, int nodeParamsSize) {
    //caculate max memory
    int deviceID;
    cudaDeviceProp dev;
    cudaGetDevice(&deviceID);
    cudaGetDeviceProperties(&dev, deviceID);
    size_t totalMemory;
    size_t availMemory;
    cudaMemGetInfo(&availMemory, &totalMemory);
    long reduceMem = nodeParamsSize * sizeof(uint) * (long) testNumNodes;
    totalSize = (availMemory - reduceMem) / edgeSize;
    max_partition_size = param * totalSize;
    printf("total memory is %ld max memory is %ld, most edge size is %ld\n total edge size %ld \n multiprocessors %d \n",
           availMemory - reduceMem,
           dev.totalGlobalMem, max_partition_size, totalSize, dev.multiProcessorCount);
    if (max_partition_size > DIST_INFINITY) {
        printf("bigger than DIST_INFINITY\n");
        max_partition_size = DIST_INFINITY;
    }
    uint temp = max_partition_size % fragment_size;
    max_partition_size = max_partition_size - temp;
}


void getMaxPartitionSize(unsigned long &max_partition_size, unsigned long &totalSize, uint testNumNodes, float param,
                         int edgeSize, uint edgeListSize, int nodeParamsSize) {
    int deviceID;
    cudaDeviceProp dev{};
    cudaGetDevice(&deviceID);
    cudaGetDeviceProperties(&dev, deviceID);
    size_t totalMemory;
    size_t availMemory;
    cudaMemGetInfo(&availMemory, &totalMemory);
    long reduceMem = nodeParamsSize * sizeof(uint) * (long) testNumNodes;
    cout << "reduceMem " << reduceMem << " testNumNodes " << testNumNodes << " nodeParamsSize " << nodeParamsSize
         << endl;
    totalSize = (availMemory - reduceMem) / edgeSize;

    float adviseK = (10 - (float) edgeListSize / (float) totalSize) / 9;
    //uint dynamicDataMax = edgeListSize * edgeSize -
    /*double tempUpper = ((double) (availMemory - reduceMem) * 15 - (double)edgeListSize * edgeSize);
    double tempLower = (double) (availMemory - reduceMem) * 14;
    double adviseK = tempUpper / tempLower;*/
    cout<<"adviseK " << adviseK << endl;
    if (adviseK < 0) {
        adviseK = 0.5;
        cout<<"adviseK " << adviseK << endl;
    }
    if (adviseK > 1) {
        adviseK = 0.95;
        cout<<"adviseK " << adviseK << endl;
    }
    if (param > 0) {
        adviseK = param;
    }

    max_partition_size = adviseK * totalSize;
    cout << "availMemory " << availMemory << " totalMemory " << totalMemory << endl;
    printf("total memory is %ld totalGlobalMem is %ld, most edge size is %ld\n total edge size %ld \n multiprocessors %d adviseK %f\n",
           availMemory - reduceMem,
           dev.totalGlobalMem, max_partition_size, totalSize, dev.multiProcessorCount, adviseK);
    if (max_partition_size > DIST_INFINITY) {
        printf("bigger than DIST_INFINITY\n");
        max_partition_size = DIST_INFINITY;
    }
    uint temp = max_partition_size % fragment_size;
    max_partition_size = max_partition_size - temp;
}

void
checkNeedTransferPartition(bool *needTransferPartition, CommonPartitionInfo *partitionInfoList, bool *isActiveNodeList,
                           int partitionNum, uint testNumNodes, uint &activeNum) {
    uint tempMinNode = UINT_MAX;
    uint tempMaxNode = 0;
    for (uint j = 0; j < testNumNodes; j++) {
        if (isActiveNodeList[j]) {
            if (j < tempMinNode) {
                tempMinNode = j;
            }
            if (j > tempMaxNode) {
                tempMaxNode = j;
            }
            activeNum++;
        }
    }
    if (activeNum <= 0) {
        return;
    }
    for (int i = 0; i < partitionNum; i++) {
        needTransferPartition[i] = false;
        if (partitionInfoList[i].startVertex <= tempMaxNode && partitionInfoList[i].endVertex >= tempMinNode) {
            needTransferPartition[i] = true;
        }
    }
}

void checkNeedTransferPartitionOpt(bool *needTransferPartition, CommonPartitionInfo *partitionInfoList,
                                   bool *isActiveNodeList, int partitionNum, uint testNumNodes, uint &activeNum) {
    for (int i = 0; i < partitionNum; i++) {
        needTransferPartition[i] = false;
    }
    for (uint j = 0; j < testNumNodes; j++) {
        if (isActiveNodeList[j]) {
            for (int i = 0; i < partitionNum; i++) {
                if (partitionInfoList[i].startVertex <= j && partitionInfoList[i].endVertex >= j) {
                    needTransferPartition[i] = true;
                }
            }
            activeNum++;
        }
    }
}

void caculatePartInfoForEdgeList(uint *overloadNodePointers, uint *overloadNodeList, uint *degree,
                                 vector<PartEdgeListInfo> &partEdgeListInfoArr, uint overloadNodeNum,
                                 uint overloadMemorySize, uint overloadEdgeNum) {
    partEdgeListInfoArr.clear();
    if (overloadMemorySize < overloadEdgeNum) {
        uint left = 0;
        uint right = overloadNodeNum - 1;
        while ((overloadNodePointers[right] + degree[overloadNodeList[right]] - overloadNodePointers[left]) >
               overloadMemorySize) {
            uint start = left;
            uint end = right;
            uint mid;
            while (start <= end) {
                mid = (start + end) / 2;
                uint headDistance = overloadNodePointers[mid] - overloadNodePointers[left];
                uint tailDistance =
                        overloadNodePointers[mid] + degree[overloadNodeList[mid]] - overloadNodePointers[left];
                if (headDistance <= overloadMemorySize && tailDistance > overloadMemorySize) {
                    break;
                } else if (tailDistance <= overloadMemorySize) {
                    start = mid + 1;
                } else if (headDistance > overloadMemorySize) {
                    end = mid - 1;
                }
            }
            PartEdgeListInfo info;
            info.partActiveNodeNums = mid - left;
            info.partEdgeNums = overloadNodePointers[mid] - overloadNodePointers[left];
            info.partStartIndex = left;
            partEdgeListInfoArr.push_back(info);
            left = mid;
        }
        PartEdgeListInfo info;
        info.partActiveNodeNums = right - left + 1;
        info.partEdgeNums = overloadNodePointers[right] + degree[overloadNodeList[right]] - overloadNodePointers[left];
        info.partStartIndex = left;
        partEdgeListInfoArr.push_back(info);
    } else {
        PartEdgeListInfo info;
        info.partActiveNodeNums = overloadNodeNum;
        info.partEdgeNums = overloadEdgeNum;
        info.partStartIndex = 0;
        partEdgeListInfoArr.push_back(info);
    }
}