Benchmarks test for Go character stitching

This is a creation in Article, where the information may have evolved or changed.

in the use of Go to do a SQL generation tool, the performance of his string splicing is curious, in Google search for half a day, as if serious foreigners do not have this result, so get one of them!

First come to the conclusion

Times 1th time 2nd 3rd times 4th 5th 2 parameters/1 million times Fmt. Sprintf 0.5550317 0.5800332 0.5800332 0.5090291 0.5160295 bytes. Newbuffer 0.9450541 1.0010573 0.9430539 0.9500544 0.9600549 + 0.1460083 0.1700097 0.1470085 0.1450083 0.1560089 Strings. Join 0.279016 0.3030173 0.2820161 0.2760158 0.3080176 4 parameters/1 million times Fmt. Sprintf 1.0260587 0.944054 1.0270587 0.944054 0.9530545 bytes. Newbuffer 1.4470828 1.3770788 1.3970799 1.5100864 1.4890852 + 0.2920167 0.2580147 0.2630151 0.2880164 0.2860163 Strings. Join 0.4670267 0.4560261 0.4700269 0.4620265 0.4630265 6 parameters/1 million times Fmt. Sprintf 1.4090806 1.3880794 1.4110808 1.4290817 1.3900795 bytes. Newbuffer 1.5800904 1.5240871 1.5440883 1.5990915 1.57209 + 0.4000229 0.4030231 0.3940225 0.4110235 0.4080233 Strings. Join 0.5630322 0.6270358 0.6280359 0.5840334 0.6340363

Oh, yes, the string + is the fastest, of course, here is just a simple test results, the situation is to choose the most consistent with the use of the scene method.

If there is a new result, it will be updated to this side: Http://git.oschina.net/janpoem/go-benchmarks

Test code

Package Mainimport ("Git.oschina.net/janpoem/go-agi") Import ("FMT" "Math/rand" "StrConv" "bytes" "Stri NGS ") Func Randstr () (string) {return StrConv. Formatint (Rand. Int63 (),}var sp_str1 = Randstr () var sp_str2 = Randstr () var sp_str3 = Randstr () var sp_str4 = Randstr () var SP_STR5 = Ran Dstr () var sp_str6 = Randstr () func Test_2_args () {max: = 1000000 dur1: = Agi. Timestest (Max, func () {fmt. Sprintf ("Select%s from%s", SP_STR1, SP_STR2)}) Dur2: = Agi. Timestest (Max, func () {bf: = bytes. Newbufferstring ("select") bf. WriteString ("") bf. WriteString (SP_STR1) bf. WriteString ("") bf. WriteString ("from") bf. WriteString (SP_STR2)}) Dur3: = Agi.    Timestest (Max, func () {bf: = "Select" + Sp_str1 + "from" + sp_str2 If Len (BF) > 0 {}}) DUR4: = Agi. Timestest (Max, func () {bf: = strings.      Join ([]string {"Select", Sp_str1, "from", sp_str2}, "") If Len (BF) > 0 {  }}) Fmt. Println (Dur1. Seconds ()) fmt. Println (DUR2. Seconds ()) fmt. Println (DUR3. Seconds ()) fmt. Println (DUR4. Seconds ())}func Test_4_args () {max: = 1000000 dur1: = Agi. Timestest (Max, func () {fmt. Sprintf ("Select%s from%s WHERE%s is ORDER by%s", Sp_str1, Sp_str2, SP_STR3, SP_STR4)}) Dur2: = Agi. Timestest (Max, func () {bf: = bytes. Newbufferstring ("select") bf. WriteString (SP_STR1) bf. WriteString ("") bf. WriteString ("from") bf. WriteString (SP_STR2) bf. WriteString ("") bf. WriteString ("WHERE") bf. WriteString (SP_STR3) bf. WriteString ("") bf. WriteString ("ORDER by") Bf. WriteString (SP_STR4)}) Dur3: = Agi. Timestest (Max, func () {bf: = "Select" + Sp_str1 + "from" + sp_str2 + "WHERE" + Sp_str3 + "ORDER by" + Sp_st R4 If Len (BF) > 0 {}}) Dur4: = Agi. Timestest (Max, func () {bf: = strings. Join ([]string {"Select", Sp_str1, "from", Sp_str2, "WHERE", Sp_str3, "ORDER by", SP_STR4}, "") If Len (BF) > 0 {}}) fmt. Println (Dur1. Seconds ()) fmt. Println (DUR2. Seconds ()) fmt. Println (DUR3. Seconds ()) fmt. Println (DUR4. Seconds ())}func Test_6_args () {max: = 1000000 dur1: = Agi. Timestest (Max, func () {fmt. Sprintf ("Select%s from%s WHERE%s is the ORDER by%s GROUP by%s LIMIT%s", Sp_str1, Sp_str2, SP_STR3, SP_STR4, SP_STR5, Sp_str 6)}) DUR2: = Agi. Timestest (Max, func () {bf: = bytes. Newbufferstring ("select") bf. WriteString (SP_STR1) bf. WriteString ("") bf. WriteString ("from") bf. WriteString (SP_STR2) bf. WriteString ("") bf. WriteString ("WHERE") bf. WriteString (SP_STR3) bf. WriteString ("") bf. WriteString ("ORDER by") Bf. WriteString (SP_STR4) bf. WriteString ("") bf. WriteString ("GROUP by") Bf. WriteString (SP_STR5) bf. WriteString ("") bf. WriteString ("LIMIT") bf. WriteString (sp_sTR6)}) Dur3: = Agi. Timestest (Max, func () {bf: = "Select" + Sp_str1 + "from" + sp_str2 + "WHERE" + SP_STR3 + "O Rder by "+ SP_STR4 +" GROUP by "+ SP_STR5 +" LIMIT "+ SP_STR6 If Len (BF) > 0 {}} ) Dur4: = Agi. Timestest (Max, func () {bf: = strings.            Join ([]string {"Select", Sp_str1, "from", Sp_str2, "WHERE", Sp_str3, "ORDER by", SP_STR4, "GROUP by", SP_STR5, "LIMIT", SP_STR6,}, "") If Len (BF) > 0 {}}) fmt. Println (Dur1. Seconds ()) fmt. Println (DUR2. Seconds ()) fmt. Println (DUR3. Seconds ()) fmt. Println (DUR4. Seconds ())}func main () {//Test_2_args ()//Test_4_args ()//Test_6_args ()}

The specific code is here: Http://git.oschina.net/janpoem/go-benchmarks/blob/master/str_concat.go

Timestest, the call is this: Benchmark.go, his function is to execute the specified number of anonymous functions, depending on how many times you pass in. At first, there is some concern about the performance of anonymous functions, but compared to the execution time of bare code and anonymous function, the difference is not much, it is ignored.

This test is just a test of the execution time and does not include CPU and memory condition monitoring.

Er, the extra said, @ Sweet Potato Blog markdown Editor is not too wonderful, table, code, what format is not supported, too difficult to use.