L1 cache reference L1 cache读取 |
0.5ns | ||
Branch mispredict 分支误预测 |
5ns | ||
L2 cache reference L2 cache读取 |
7ns | 14x L1 cache | |
Mutex lock/unlock 互斥锁 |
25ns | ||
Main memory reference 主内存读取 |
100ns | 20x L2 cache, 200x L1 cache | |
Compress 1K bytes with Zippy 压缩1KB数据 |
3,000ns | ||
Send 1K bytes over 1 Gbps network 在1Gbps网络上发送2K数据 |
10,000ns | 0.01ms | |
Read 4K randomly from SSD* 从SSD硬盘随机读取4K数据* |
150,000ns | 0.15ms | |
Read 1 MB sequentially from memory 从内存连续读取1MB数据 |
250,000ns | 0.25ms | |
Round trip within same datacenter 同一数据中心的往返时间(RTT) |
500,000ns | 0.5ms | |
Read 1 MB sequentially from SSD* 从SSD硬盘顺序读取1MB数据* |
1,000,000ns | 1ms | 4x memory |
Disk seek 磁盘寻道时间 |
10,000,000ns | 10ms | 20x datacenter roundtrip |
Read 1 MB sequentially from Network 从网络上顺序读取1M数据 |
10,000,000ns | 10ms | |
Read 1 MB sequentially from disk 从磁盘上顺序读取1M数据 |
20,000,000ns | 20ms | 80x memory, 20x SSD |
Send packet CA->Netherlands->CA 发送一个包从从加州->荷兰->加州 |
150,000,000ns | 150ms |
* Assuming ~1GB/sec SSD
注:原文来自https://gist.github.com/jboner/2841832,略有修改
Data Warehouse For Ever原创文章,转载请注明出处