C语言字符串压缩之ZSTD算法详解

发布时间:

最近项目上有大量的字符串数据需要存储到内存,并且需要储存至一定时间,于是自然而然的想到了使用字符串压缩算法对“源串”进行压缩存储。由此触发了对一些优秀压缩算法的调研。

字符串压缩,我们通常的需求有几个,一是高压缩率,二是压缩速率高,三是解压速率高。不过高压缩率与高压缩速率是鱼和熊掌的关系,不可皆得,优秀的算法一般也是采用压缩率与性能折中的方案。从压缩率、压缩速率、解压速率考虑,zstd与lz4有较好的压缩与解压性能,最终选取zstd与lz4进行调研。

zstd是facebook开源的提供高压缩比的快速压缩算法(参考https://github.com/facebook/zstd),很想了解一下它在压缩与解压方面的实际表现。

C语言字符串压缩之ZSTD算法详解

一、zstd压缩与解压

ZSTD_compress属于ZSTD的Simple API范畴,只有压缩级别可以设置。

ZSTD_compress函数原型如下:

size_tZSTD_compress(void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel)

ZSTD_decompress函数原型如下:

size_t ZSTD_decompress( void* dst, size_t dstCapacity, const void* src, size_t compressedSize);  我们先来看看zstd的压缩与解压缩示例。

#include <stdio.h>
#include <string.h>
#include <sys/time.h>
#include <malloc.h>
#include <zstd.h>
#include <iostream>
using namespace std;
int main()
{
// compress
size_t com_space_size;
size_t peppa_pig_text_size;
char *com_ptr = NULL;
char peppa_pig_buf[2048] = "Narrator: It is raining today. So, Peppa and George cannot play outside.Peppa: Daddy, it's stopped raining. Can we go out to play?Daddy: Alright, run along you two.Narrator: Peppa loves jumping in muddy puddles.Peppa: I love muddy puddles.Mummy: Peppa. If you jumping in muddy puddles, you must wear your boots.Peppa: Sorry, Mummy.Narrator: George likes to jump in muddy puddles, too.Peppa: George. If you jump in muddy puddles, you must wear your boots.Narrator: Peppa likes to look after her little brother, George.Peppa: George, let's find some more pud dles.Narrator: Peppa and George are having a lot of fun. Peppa has found a lttle puddle. George hasfound a big puddle.Peppa: Look, George. There's a really big puddle.Narrator: George wants to jump into the big puddle first.Peppa: Stop, George. | must check if it's safe for you. Good. It is safe for you. Sorry, George. It'sonly mud.Narrator: Peppa and George love jumping in muddy puddles.Peppa: Come on, George. Let's go and show Daddy.Daddy: Goodness me.Peppa: Daddy. Daddy. Guess what we' ve been doing.Daddy: Let me think... Have you been wa tching television?Peppa: No. No. Daddy.Daddy: Have you just had a bath?Peppa: No. No.Daddy: | know. You've been jumping in muddy puddles.Peppa: Yes. Yes. Daddy. We've been jumping in muddy puddles.Daddy: Ho. Ho. And look at the mess you're in.Peppa: Oooh....Daddy: Oh, well, it's only mud. Let's clean up quickly before Mummy sees the mess.Peppa: Daddy, when we've cleaned up, will you and Mummy Come and play, too?Daddy: Yes, we can all play in the garden.Narrator: Peppa and George are wearing their boots. Mummy and Daddy are wearing their boots.Peppa loves jumping up and down in muddy puddles. Everyone loves jumping up and down inmuddy puddles.Mummy: Oh, Daddy pig, look at the mess you're in. .Peppa: It's only mud.";
peppa_pig_text_size = strlen(peppa_pig_buf);
com_space_size= ZSTD_compressBound(peppa_pig_text_size);
com_ptr = (char *)malloc(com_space_size);
if(NULL == com_ptr) {
cout << "compress malloc failed" << endl;
return -1;
}
size_t com_size;
com_size = ZSTD_compress(com_ptr, com_space_size, peppa_pig_buf, peppa_pig_text_size, ZSTD_fast);
cout << "peppa pig text size:" << peppa_pig_text_size << endl;
cout << "compress text size:" << com_size << endl;
cout << "compress ratio:" << (float)peppa_pig_text_size / (float)com_size << endl << endl;
// decompress
char* decom_ptr = NULL;
unsigned long long decom_buf_size;
decom_buf_size = ZSTD_getFrameContentSize(com_ptr, com_size);
decom_ptr = (char *)malloc((size_t)decom_buf_size);
if(NULL == decom_ptr) {
cout << "decompress malloc failed" << endl;
return -1;
}
size_t decom_size;
decom_size = ZSTD_decompress(decom_ptr, decom_buf_size, com_ptr, com_size);
cout << "decompress text size:" << decom_size << endl;
if(strncmp(peppa_pig_buf, decom_ptr, peppa_pig_text_size)) {
cout << "decompress text is not equal peppa pig text" << endl;
}
free(com_ptr);
free(decom_ptr);
return 0;
}

执行结果:

C语言字符串压缩之ZSTD算法详解

从结果可以发现,压缩之前的peppa pig文本长度为1827,压缩后的文本长度为759,压缩率为2.4,解压后的长度与压缩前相等。

另外,上文提到可以调整ZSTD_compress函数的压缩级别,zstd的默认级别为ZSTD_CLEVEL_DEFAULT = 3,最小值为0,最大值为ZSTD_MAX_CLEVEL = 22。另外也提供一些策略设置,例如ZSTD_fast, ZSTD_greedy, ZSTD_lazy, ZSTD_lazy2, ZSTD_btlazy2。压缩级别越高,压缩率越高,但是压缩速率越低。

二、ZSTD压缩与解压性能探索

上面探索了zstd的基础压缩与解压方法,接下来再摸索一下zstd的压缩与解压缩性能。

测试方法是,使用ZSTD_compress连续压缩同一段文本并持续10秒,最后得到每一秒的平均压缩速率。测试压缩性能的代码示例如下:

#include <stdio.h>
#include <string.h>
#include <sys/time.h>
#include <malloc.h>
#include <zstd.h>
#include <iostream>
using namespace std;
int main()
{
int cnt = 0;
size_t com_size;
size_t com_space_size;
size_t peppa_pig_text_size;
char *com_ptr = NULL;
char peppa_pig_buf[2048] = "Narrator: It is raining today. So, Peppa and George cannot play outside.Peppa: Daddy, it's stopped raining. Can we go out to play?Daddy: Alright, run along you two.Narrator: Peppa loves jumping in muddy puddles.Peppa: I love muddy puddles.Mummy: Peppa. If you jumping in muddy puddles, you must wear your boots.Peppa: Sorry, Mummy.Narrator: George likes to jump in muddy puddles, too.Peppa: George. If you jump in muddy puddles, you must wear your boots.Narrator: Peppa likes to look after her little brother, George.Peppa: George, let's find some more pud dles.Narrator: Peppa and George are having a lot of fun. Peppa has found a lttle puddle. George hasfound a big puddle.Peppa: Look, George. There's a really big puddle.Narrator: George wants to jump into the big puddle first.Peppa: Stop, George. | must check if it's safe for you. Good. It is safe for you. Sorry, George. It'sonly mud.Narrator: Peppa and George love jumping in muddy puddles.Peppa: Come on, George. Let's go and show Daddy.Daddy: Goodness me.Peppa: Daddy. Daddy. Guess what we' ve been doing.Daddy: Let me think... Have you been wa tching television?Peppa: No. No. Daddy.Daddy: Have you just had a bath?Peppa: No. No.Daddy: | know. You've been jumping in muddy puddles.Peppa: Yes. Yes. Daddy. We've been jumping in muddy puddles.Daddy: Ho. Ho. And look at the mess you're in.Peppa: Oooh....Daddy: Oh, well, it's only mud. Let's clean up quickly before Mummy sees the mess.Peppa: Daddy, when we've cleaned up, will you and Mummy Come and play, too?Daddy: Yes, we can all play in the garden.Narrator: Peppa and George are wearing their boots. Mummy and Daddy are wearing their boots.Peppa loves jumping up and down in muddy puddles. Everyone loves jumping up and down inmuddy puddles.Mummy: Oh, Daddy pig, look at the mess you're in. .Peppa: It's only mud.";
timeval st, et;
peppa_pig_text_size = strlen(peppa_pig_buf);
com_space_size= ZSTD_compressBound(peppa_pig_text_size);
gettimeofday(&st, NULL);
while(1) {
com_ptr = (char *)malloc(com_space_size);
com_size = ZSTD_compress(com_ptr, com_space_size, peppa_pig_buf, peppa_pig_text_size, ZSTD_fast);
free(com_ptr);
cnt++;
gettimeofday(&et, NULL);
if(et.tv_sec - st.tv_sec >= 10) {
break;
}
}
cout << "compress per second:" << cnt/10 << " times" << endl;
return 0;
}

执行结果:

C语言字符串压缩之ZSTD算法详解

结果显示ZSTD的压缩性能大概在每秒6-7万次左右,这个结果其实并不是太理想。需要说明的是压缩性能与待压缩文本的长度、字符内容也是有关系的。

我们再来探索一下ZSTD的解压缩性能。与上面的测试方法类似,先对本文进行压缩,然后连续解压同一段被压缩过的数据并持续10秒,最后得到每一秒的平均解压速率。测试解压性能的代码示例如下:

#include <stdio.h>
#include <string.h>
#include <sys/time.h>
#include <malloc.h>
#include <zstd.h>
#include <iostream>
using namespace std;
int main()
{
int cnt = 0;
size_t com_size;
size_t com_space_size;
size_t peppa_pig_text_size;
timeval st, et;
char *com_ptr = NULL;
char peppa_pig_buf[2048] = "Narrator: It is raining today. So, Peppa and George cannot play outside.Peppa: Daddy, it's stopped raining. Can we go out to play?Daddy: Alright, run along you two.Narrator: Peppa loves jumping in muddy puddles.Peppa: I love muddy puddles.Mummy: Peppa. If you jumping in muddy puddles, you must wear your boots.Peppa: Sorry, Mummy.Narrator: George likes to jump in muddy puddles, too.Peppa: George. If you jump in muddy puddles, you must wear your boots.Narrator: Peppa likes to look after her little brother, George.Peppa: George, let's find some more pud dles.Narrator: Peppa and George are having a lot of fun. Peppa has found a lttle puddle. George hasfound a big puddle.Peppa: Look, George. There's a really big puddle.Narrator: George wants to jump into the big puddle first.Peppa: Stop, George. | must check if it's safe for you. Good. It is safe for you. Sorry, George. It'sonly mud.Narrator: Peppa and George love jumping in muddy puddles.Peppa: Come on, George. Let's go and show Daddy.Daddy: Goodness me.Peppa: Daddy. Daddy. Guess what we' ve been doing.Daddy: Let me think... Have you been wa tching television?Peppa: No. No. Daddy.Daddy: Have you just had a bath?Peppa: No. No.Daddy: | know. You've been jumping in muddy puddles.Peppa: Yes. Yes. Daddy. We've been jumping in muddy puddles.Daddy: Ho. Ho. And look at the mess you're in.Peppa: Oooh....Daddy: Oh, well, it's only mud. Let's clean up quickly before Mummy sees the mess.Peppa: Daddy, when we've cleaned up, will you and Mummy Come and play, too?Daddy: Yes, we can all play in the garden.Narrator: Peppa and George are wearing their boots. Mummy and Daddy are wearing their boots.Peppa loves jumping up and down in muddy puddles. Everyone loves jumping up and down inmuddy puddles.Mummy: Oh, Daddy pig, look at the mess you're in. .Peppa: It's only mud.";
size_t decom_size;
char* decom_ptr = NULL;
unsigned long long decom_buf_size;
peppa_pig_text_size = strlen(peppa_pig_buf);
com_space_size= ZSTD_compressBound(peppa_pig_text_size);
com_ptr = (char *)malloc(com_space_size);
com_size = ZSTD_compress(com_ptr, com_space_size, peppa_pig_buf, peppa_pig_text_size, 1);
gettimeofday(&st, NULL);
decom_buf_size = ZSTD_getFrameContentSize(com_ptr, com_size);
while(1) {
decom_ptr = (char *)malloc((size_t)decom_buf_size);
decom_size = ZSTD_decompress(decom_ptr, decom_buf_size, com_ptr, com_size);
if(decom_size != peppa_pig_text_size) {
cout << "decompress error" << endl;
break;
}
free(decom_ptr);
cnt++;
gettimeofday(&et, NULL);
if(et.tv_sec - st.tv_sec >= 10) {
break;
}
}
cout << "decompress per second:" << cnt/10 << " times" << endl;
free(com_ptr);
return 0;
}

执行结果:

C语言字符串压缩之ZSTD算法详解

结果显示ZSTD的解压缩性能大概在每秒12万次左右,解压性能比压缩性能高。

三、zstd的高级用法

zstd提供了一个名为PZSTD的压缩和解压工具。PZSTD(parallel zstd),并行压缩的zstd,是一个使用多线程对待压缩文本进行切片分段,且进行并行压缩的命令行工具。

其实高版本(v1.4.0及以上)的zstd也提供了指定多线程对文本进行并行压缩的相关API接口,也就是本小节要介绍的zstd高级API用法。下面我们再来探索一下zstd的多线程压缩使用方法。

多线程并行压缩的两个关键API,一个是参数设置API,另一个是压缩API。

参数设置API的原型是:

size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)

压缩API的原型是:

size_t ZSTD_compress2(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)

下面给出zstd并行压缩的示例demo,通过ZSTD_CCtx_setParameter设置线程数为3,即指定宏ZSTD_c_nbWorkers为3,通过ZSTD_compress2压缩相关文本。另外,为了展示zstd确实使用了多线程,需要先读取一个非常大的文件,作为zstd的压缩文本源,尽量使zstd运行较长时间。

#include <stdio.h>
#include <string.h>
#include <sys/time.h>
#include <malloc.h>
#include <zstd.h>
#include <iostream>
using namespace std;
int main()
{
size_t com_size;
size_t com_space_size;
FILE *fp = NULL;
unsigned int file_len;
char *com_ptr = NULL;
char *file_text_ptr = NULL;
fp = fopen("xxxxxx", "r");
if(NULL == fp){
cout << "file open failed" << endl;
return -1;
}
fseek(fp, 0, SEEK_END);
file_len = ftell(fp);
fseek(fp, 0, SEEK_SET);
cout << "file length:" << file_len << endl;
// malloc space for file content
file_text_ptr = (char *)malloc(file_len);
if(NULL == file_text_ptr) {
cout << "malloc failed" << endl;
return -1;
}
// malloc space for compress space
com_space_size = ZSTD_compressBound(file_len);
com_ptr = (char *)malloc(com_space_size);
if(NULL == com_ptr) {
cout << "malloc failed" << endl;
return -1;
}
// read text from source file
fread(file_text_ptr, 1, file_len, fp);
fclose(fp);
ZSTD_CCtx* cctx;
cctx = ZSTD_createCCtx();
// set multi-thread parameter
ZSTD_CCtx_setParameter(cctx, ZSTD_c_nbWorkers, 3);
ZSTD_CCtx_setParameter(cctx, ZSTD_c_compressionLevel, ZSTD_btlazy2);
com_size = ZSTD_compress2(cctx, com_ptr, com_space_size, file_text_ptr, file_len);
free(com_ptr);
free(file_text_ptr);
return 0;
}

运行上述demo,可见zstd确实启动了3个线程对文本进行了并行压缩。且设置的线程数越多,压缩时间越短,这里就不详细展示了,读者可以自行实验。

需要说明的是,zstd当前默认编译单线程的库文件,要实现多线程的API调用,需要在make的时候指定编译参数ZSTD_MULTITHREAD。

C语言字符串压缩之ZSTD算法详解

另外,zstd还支持线程池的方式,线程池的函数原型:

POOL_ctx* ZSTD_createThreadPool(size_t numThreads)

线程池可以避免在多次、连续压缩场景时频繁的去创建线程、撤销线程产生的非必要开销,使得算力主要开销在文本压缩方面。

四、总结

本篇分享了zstd压缩与解压缩使用的基本方法,对压缩与解压的性能进行了摸底,最后探索了zstd多线程压缩的使用方法。

从压缩测试来看,zstd的压缩比其实已经比较好了,比原文所占用空间缩小了一半以上,当然压缩比也跟待压缩文本的内容有关。

从性能执行结果来看,zstd的压缩与解压性能表现比较勉强,我认为zstd在鱼(性能)和熊掌(压缩比)之间更偏向熊掌一些,不过对一些性能要求不太高的,但是要高压缩比的场景是比较符合的。

多线程并行压缩,在有大文本需要连续多次压缩的场景下,结合线程池可以很好的提升压缩速率。