Hep beraber LINUX Sunucularımizin performansini olcelim (advanced benchmark)

09-08-2010, 17:34:28

Merhabalar : Yaz aylarinda hazir yogunlugum yokken size ne tur bir katkida bulunabilirim diye dusunurken , hepimizin bir yerde merak ettigi dedicated'lerimizin yada aldiginiz VDS-VPS hizmetlerinizin performans olcumu geldi.

Kullanacagimiz benchmark yazilimi 2 turlu calisiyor 1.cisi "cpu/ram" kombinasyonu ile birlikte 2.cisi cpu/ram/storage kombinasyonu ile sunucunuzun performansini olcuyor , biz genelde 1. kombinasyonu kullanip storage 'i (hardisk) gormezden gelecegiz cunki cogu sunucu SAS hdd'ler kullanmadigi icin skorlar I/0 wait performansi yuzunden dusuk cikacaktir.(moraller bozulmasin).

Benchmark yazilimin kurulumuna gecelim , kurulumu amator duzeyde kullanicilara hitaben adim adim anlatacagim ;

wget http://byte-unixbench.googlecode.com/files/unixbench-5.1.2.tar.gz

Cogu sunucuda "lbXext" paketi kurulu degildir bu pakete ihtiyacimiz var kuruyoruz ,

yum -y install libXext-devel

Daha sonra tar.gz 'yi acalim

tar zxvf unixbench-5.1.2.tar.gz

Klasorun icine girelim ,

cd unixbench-5.1.2/

Benchmarki baslatalim asagidaki baslatma kodu HDD 'nizi gormezden gelerek cpu/ram kombinasyonu ile performansa baslayacaktir ;

./Run dhry2reg whetstone-double syscall pipe context1 spawn execl shell1 shell8 shell16

Yaklasik sunucunuza gore 10-15 dk sure zarfinda devam edecektir , bittiginde asagidaki sekilde bir sonuc gelir , testi uyguladigim sunucumun ozellikleri ;

2x5620 Quadcore 16 cekirdek - 12gb ram

Benchmark sonucu ;

Benchmark Run: Sun Aug 08 2010 16:17:11 - 16:41:03
16 CPUs in system; running 16 parallel copies of tests

Dhrystone 2 using register variables      106397851.9 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    31403.7 MWIPS (10.2 s, 7 samples)
Execl Throughput                              43601.4 lps   (30.0 s, 2 samples)
Pipe Throughput                             8759333.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                2837054.6 lps   (10.0 s, 7 samples)
Process Creation                             151069.7 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  58710.6 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  3882.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   7898.7 lpm   (60.0 s, 2 samples)
System Call Overhead                        7166185.2 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0  106397851.9   9117.2
Double-Precision Whetstone                       55.0      31403.7   5709.8
Execl Throughput                                 43.0      43601.4  10139.9
Pipe Throughput                               12440.0    8759333.4   7041.3
Pipe-based Context Switching                   4000.0    2837054.6   7092.6
Process Creation                                126.0     151069.7  11989.7
Shell Scripts (1 concurrent)                     42.4      58710.6  13846.8
Shell Scripts (16 concurrent)                     ---       3882.0      ---
Shell Scripts (8 concurrent)                      6.0       7898.7  13164.6
System Call Overhead                          15000.0    7166185.2   4777.5
                                                                                   ========
System Benchmarks Index Score (Partial Only)                         8664.6

Sonucta goruldugu gibi 16 cekirdegin tumu kullanilarak yapilan bencmark testinde sunucumun skoru : "8864.6"

Beki bu testlerin size nasil faydasi olacak? ; Herkez hizmet aldigi vps 'sinin ozelliklerini yada dedicated'inin ozelliklerini buraya yazarak benchmark testinde bulunursa ileriki zamanlarda benchmark arsivimiz bu konu altinda genisler ve yeni bir sunucuya gecis yapilacaginda bu arsive bakilarak gecmek istedigi dedicated/vps benchmark skoruna ve kendi eski sunucusunun skoruna bakarak % de kac bir verim alacagini kestirebilir.

Umarim isinize yarar faydali bir paylasim olmustur.

Saygilarimla

Çağdaş Aydoğdu

09-08-2010, 17:46:19

#2

Serverfield_Sezer

Xeon X3430 HT'siz 4 çekirdek.

Alıntı

Benchmark Run: Mon Aug 09 2010 18:15:11 - 18:51:39
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 35926603.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 8548.4 MWIPS (10.2 s, 7 samples)
Execl Throughput 12221.7 lps (29.7 s, 2 samples)
Pipe Throughput 2165325.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 691587.5 lps (10.0 s, 7 samples)
Process Creation 36343.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 13454.0 lpm (60.0 s, 2 samples)
Shell Scripts (16 concurrent) 1099.4 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 2172.0 lpm (60.4 s, 2 samples)
System Call Overhead 2497501.1 lps (10.0 s, 7 samples)

System Benchmarks Partial Index BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 35926603.5 3078.5
Double-Precision Whetstone 55.0 8548.4 1554.3
Execl Throughput 43.0 12221.7 2842.3
Pipe Throughput 12440.0 2165325.6 1740.6
Pipe-based Context Switching 4000.0 691587.5 1729.0
Process Creation 126.0 36343.0 2884.4
Shell Scripts (1 concurrent) 42.4 13454.0 3173.1
Shell Scripts (16 concurrent) --- 1099.4 ---
Shell Scripts (8 concurrent) 6.0 2172.0 3620.0
System Call Overhead 15000.0 2497501.1 1665.0
========
System Benchmarks Index Score (Partial Only) 2358.5

09-08-2010, 17:54:14

#3

KaliteNet

Kimlik doğrulama veya yönetimden onay bekliyor.

Intel Core2 Quad Processor Q9650 (12M Cache, 3.00 GHz, 1333 MHz FSB)

Alıntı

------------------------------------------------------------------------
Benchmark Run: Mon Aug 09 2010 19:20:27 - 20:02:17
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 12384314.0 lps (9.9 s, 7 samples)
Double-Precision Whetstone 2618.7 MWIPS (9.7 s, 7 samples)
Execl Throughput 4307.3 lps (29.3 s, 2 samples)
Pipe Throughput 892336.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 289558.8 lps (10.0 s, 7 samples)
Process Creation 13763.6 lps (30.5 s, 2 samples)
Shell Scripts (1 concurrent) 5002.4 lpm (60.4 s, 2 samples)
Shell Scripts (16 concurrent) 663.9 lpm (59.6 s, 2 samples)
Shell Scripts (8 concurrent) 1216.9 lpm (60.0 s, 2 samples)
System Call Overhead 767806.2 lps (10.0 s, 7 samples)

System Benchmarks Partial Index BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 12384314.0 1061.2
Double-Precision Whetstone 55.0 2618.7 476.1
Execl Throughput 43.0 4307.3 1001.7
Pipe Throughput 12440.0 892336.9 717.3
Pipe-based Context Switching 4000.0 289558.8 723.9
Process Creation 126.0 13763.6 1092.3
Shell Scripts (1 concurrent) 42.4 5002.4 1179.8
Shell Scripts (16 concurrent) --- 663.9 ---
Shell Scripts (8 concurrent) 6.0 1216.9 2028.1
System Call Overhead 15000.0 767806.2 511.9
========
System Benchmarks Index Score (Partial Only) 890.4

------------------------------------------------------------------------
Benchmark Run: Mon Aug 09 2010 20:02:17 - 20:29:05
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 37321991.0 lps (10.1 s, 7 samples)
Double-Precision Whetstone 10368.5 MWIPS (10.0 s, 7 samples)
Execl Throughput 7764.0 lps (29.4 s, 2 samples)
Pipe Throughput 1790221.7 lps (10.0 s, 7 samples)
Pipe-based Context Switching 809064.8 lps (10.0 s, 7 samples)
Process Creation 22767.4 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 7629.9 lpm (60.2 s, 2 samples)
Shell Scripts (16 concurrent) 618.7 lpm (60.4 s, 2 samples)
Shell Scripts (8 concurrent) 1272.4 lpm (60.1 s, 2 samples)
System Call Overhead 1378646.5 lps (10.0 s, 7 samples)

System Benchmarks Partial Index BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 37321991.0 3198.1
Double-Precision Whetstone 55.0 10368.5 1885.2
Execl Throughput 43.0 7764.0 1805.6
Pipe Throughput 12440.0 1790221.7 1439.1
Pipe-based Context Switching 4000.0 809064.8 2022.7
Process Creation 126.0 22767.4 1806.9
Shell Scripts (1 concurrent) 42.4 7629.9 1799.5
Shell Scripts (16 concurrent) --- 618.7 ---
Shell Scripts (8 concurrent) 6.0 1272.4 2120.7
System Call Overhead 15000.0 1378646.5 919.1
========
System Benchmarks Index Score (Partial Only) 1802.5

09-08-2010, 18:00:02

#4

PcMaKeR

Güzel fikir. Bakalım ne sonuçlar çıkacak.

09-08-2010, 18:20:47

#5

Kosanbit

Ben daha çok http://www.cpubenchmark.net/high_end_cpus.html bu listeyi merak ediyorum. Pek çoğumuz bu listeyi baz alıyoruz ama bana ciddi tutarsızlıklar var gibi geliyor.i7 işlemciler, pek çok üst düzey Xeon işlemciden daha falza puan almış ve bununda ötesinde bazı düşük alt modeller, üst modellerinden daha fazla puan almış. Örnek vermek gerekirse,
Intel Xeon X5650 @ 2.67GHz 7,765 puan alırken, bir alt modeli olan Intel Xeon X5660 @ 2.80GHz 7,015 puan alıyor. Bu tablodaki testler neye göre yapılıyor ve ne kadar sağlıklı? Ben açıkçası pek de sağlıklı olduğunu düşünmüyorum şu örnekten sonra.

09-08-2010, 18:39:52

#6

abacus

Kosanbit adlı üyeden alıntı: mesajı görüntüle

Ben daha çok http://www.cpubenchmark.net/high_end_cpus.html bu listeyi merak ediyorum. Pek çoğumuz bu listeyi baz alıyoruz ama bana ciddi tutarsızlıklar var gibi geliyor.i7 işlemciler, pek çok üst düzey Xeon işlemciden daha falza puan almış ve bununda ötesinde bazı düşük alt modeller, üst modellerinden daha fazla puan almış. Örnek vermek gerekirse,
Intel Xeon X5650 @ 2.67GHz 7,765 puan alırken, bir alt modeli olan Intel Xeon X5660 @ 2.80GHz 7,015 puan alıyor. Bu tablodaki testler neye göre yapılıyor ve ne kadar sağlıklı? Ben açıkçası pek de sağlıklı olduğunu düşünmüyorum şu örnekten sonra.

Merhabalar : Benchmark testleri hic bir zaman %100 sonuclar vermez fakat yeni bir sunucu yatirimi yapacaginizda size az bucuk fikirler verebilir , mesela dun yaptigim bir kac islemci testinden size bir kac enstantene sunabilirim ;

2X5430 2.66GHz (2 fiziksel islemcili)Quadcore 8 cekirdek - 12GB DDR2 RAM

Benchmark sonucu :

------------------------------------------------------------------------
Benchmark Run: Sun Aug 08 2010 17:13:52 - 17:38:05
8 CPUs in system; running 8 parallel copies of tests

Dhrystone 2 using register variables       90366849.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    19096.4 MWIPS (10.1 s, 7 samples)
Execl Throughput                              20090.9 lps   (30.0 s, 2 samples)
Pipe Throughput                             4018098.2 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                1546855.1 lps   (10.0 s, 7 samples)
Process Creation                              63629.9 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  30394.6 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  2400.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   4762.2 lpm   (60.0 s, 2 samples)
System Call Overhead                        2203295.7 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   90366849.6   7743.5
Double-Precision Whetstone                       55.0      19096.4   3472.1
Execl Throughput                                 43.0      20090.9   4672.3
Pipe Throughput                               12440.0    4018098.2   3230.0
Pipe-based Context Switching                   4000.0    1546855.1   3867.1
Process Creation                                126.0      63629.9   5050.0
Shell Scripts (1 concurrent)                     42.4      30394.6   7168.5
Shell Scripts (16 concurrent)                     ---       2400.0      ---
Shell Scripts (8 concurrent)                      6.0       4762.2   7937.0
System Call Overhead                          15000.0    2203295.7   1468.9
                                                                                ========
System Benchmarks Index Score (Partial Only)                         4433.8

Simdi tek fiziksel islemcili Xeon 3460 HT 2.80GHz 8 cekirdek - 4GB RAM DDR3

Bencmark sonucu :

------------------------------------------------------------------------
Benchmark Run: Sun Aug 08 2010 18:52:16 - 19:16:06
8 CPUs in system; running 8 parallel copies of tests

Dhrystone 2 using register variables       59716676.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    18199.7 MWIPS (10.2 s, 7 samples)
Execl Throughput                              28119.0 lps   (30.0 s, 2 samples)
Pipe Throughput                             4506101.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                1492913.5 lps   (10.0 s, 7 samples)
Process Creation                              91171.5 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  34835.1 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  2521.7 lpm   (60.1 s, 2 samples)
Shell Scripts (8 concurrent)                   5051.0 lpm   (60.0 s, 2 samples)
System Call Overhead                        6452069.4 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   59716676.6   5117.1
Double-Precision Whetstone                       55.0      18199.7   3309.0
Execl Throughput                                 43.0      28119.0   6539.3
Pipe Throughput                               12440.0    4506101.1   3622.3
Pipe-based Context Switching                   4000.0    1492913.5   3732.3
Process Creation                                126.0      91171.5   7235.8
Shell Scripts (1 concurrent)                     42.4      34835.1   8215.8
Shell Scripts (16 concurrent)                     ---       2521.7      ---
Shell Scripts (8 concurrent)                      6.0       5051.0   8418.3
System Call Overhead                          15000.0    6452069.4   4301.4
                                                                                  ========
System Benchmarks Index Score (Partial Only)                         5286.1

Gordugunuz gibi cok enteresan sekilde 2 islemcili 5430 2.66GHZ - 12gb ram DDR3 benchmark skoru : "4433.8" fakat Tek islemcili Xeon 3460 2.80GHZ Hyper-threading 8 cekirdek benchmark puani "5286.1".

Bu 2 sunucuyu yurtdisinda kiralamaya kalksaniz 2x5430 , 3460 'dan en asagi 100$ daha pahali , mesela bu testleri gordukten sonra 2X5430 sahipleri varsa 3460 'a gecmeleri onlara daha iyi performansi daha ucuza almalarini saglayacak.Tabiki Database uygulamarinda cache on plana cikacagi icin 2X5430 'un cache'si daha (2x12mb) fazla ve o tur bir operasyon icin daha uygun.

O yuzden bu tarz testler yapilmali ve arsiv olusturulmali.

iyi calismalar.

09-08-2010, 18:47:52

#7

Kosanbit

abacus adlı üyeden alıntı: mesajı görüntüle

Merhabalar : Benchmark testleri hic bir zaman %100 sonuclar vermez fakat yeni bir sunucu yatirimi yapacaginizda size az bucuk fikirler verebilir , mesela dun yaptigim bir kac islemci testinden size bir kac enstantene sunabilirim ;

2X5430 2.66GHz (2 fiziksel islemcili)Quadcore 8 cekirdek - 12GB DDR2 RAM

Benchmark sonucu :

------------------------------------------------------------------------
Benchmark Run: Sun Aug 08 2010 17:13:52 - 17:38:05
8 CPUs in system; running 8 parallel copies of tests

Dhrystone 2 using register variables       90366849.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    19096.4 MWIPS (10.1 s, 7 samples)
Execl Throughput                              20090.9 lps   (30.0 s, 2 samples)
Pipe Throughput                             4018098.2 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                1546855.1 lps   (10.0 s, 7 samples)
Process Creation                              63629.9 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  30394.6 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  2400.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   4762.2 lpm   (60.0 s, 2 samples)
System Call Overhead                        2203295.7 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   90366849.6   7743.5
Double-Precision Whetstone                       55.0      19096.4   3472.1
Execl Throughput                                 43.0      20090.9   4672.3
Pipe Throughput                               12440.0    4018098.2   3230.0
Pipe-based Context Switching                   4000.0    1546855.1   3867.1
Process Creation                                126.0      63629.9   5050.0
Shell Scripts (1 concurrent)                     42.4      30394.6   7168.5
Shell Scripts (16 concurrent)                     ---       2400.0      ---
Shell Scripts (8 concurrent)                      6.0       4762.2   7937.0
System Call Overhead                          15000.0    2203295.7   1468.9
                                                                                ========
System Benchmarks Index Score (Partial Only)                         4433.8

Simdi tek fiziksel islemcili Xeon 3460 HT 2.80GHz 8 cekirdek - 4GB RAM DDR3

Bencmark sonucu :

------------------------------------------------------------------------
Benchmark Run: Sun Aug 08 2010 18:52:16 - 19:16:06
8 CPUs in system; running 8 parallel copies of tests

Dhrystone 2 using register variables       59716676.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    18199.7 MWIPS (10.2 s, 7 samples)
Execl Throughput                              28119.0 lps   (30.0 s, 2 samples)
Pipe Throughput                             4506101.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                1492913.5 lps   (10.0 s, 7 samples)
Process Creation                              91171.5 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  34835.1 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  2521.7 lpm   (60.1 s, 2 samples)
Shell Scripts (8 concurrent)                   5051.0 lpm   (60.0 s, 2 samples)
System Call Overhead                        6452069.4 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   59716676.6   5117.1
Double-Precision Whetstone                       55.0      18199.7   3309.0
Execl Throughput                                 43.0      28119.0   6539.3
Pipe Throughput                               12440.0    4506101.1   3622.3
Pipe-based Context Switching                   4000.0    1492913.5   3732.3
Process Creation                                126.0      91171.5   7235.8
Shell Scripts (1 concurrent)                     42.4      34835.1   8215.8
Shell Scripts (16 concurrent)                     ---       2521.7      ---
Shell Scripts (8 concurrent)                      6.0       5051.0   8418.3
System Call Overhead                          15000.0    6452069.4   4301.4
                                                                                  ========
System Benchmarks Index Score (Partial Only)                         5286.1

Gordugunuz gibi cok enteresan sekilde 2 islemcili 5430 2.66GHZ - 12gb ram DDR3 benchmark skoru : "4433.8" fakat Tek islemcili Xeon 3460 2.80GHZ Hyper-threading 8 cekirdek benchmark puani "5286.1".

Bu 2 sunucuyu yurtdisinda kiralamaya kalksaniz 2x5430 , 3460 'dan en asagi 100$ daha pahali , mesela bu testleri gordukten sonra 2X5430 sahipleri varsa 3460 'a gecmeleri onlara daha iyi performansi daha ucuza almalarini saglayacak.Tabiki Database uygulamarinda cache on plana cikacagi icin 2X5430 'un cache'si daha (2x12mb) fazla ve o tur bir operasyon icin daha uygun.

O yuzden bu tarz testler yapilmali ve arsiv olusturulmali.

iyi calismalar.

Kişisel olarak E5620 ve 55** serisini karşılaştıracak olursanız, pratikte gördüğünüz farkları söylemeniz mümkün mü? 55 serisinden cache olarak fazla olduğunu biliyorum ancak testler yine de kafamı karıştırıyor (: Tabi iki neslide çift işlemci olarak değerlendirelim.

09-08-2010, 18:53:55

#8

abacus

Merhahabalar : Dual 5620 - 12gb ram DDR3 bencmark testini konuyu acarken yapmistim ve skoru yukarida yaziyor.

İsteginiz uzere , Dual 5520 - 12gb RAM DDR3 benchmark sonucu :

Benchmark Run: Sun Aug 08 2010 17:18:16 - 17:42:08
16 CPUs in system; running 16 parallel copies of tests

Dhrystone 2 using register variables       99365861.5 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    29766.9 MWIPS (10.2 s, 7 samples)
Execl Throughput                              39451.2 lps   (30.0 s, 2 samples)
Pipe Throughput                             7489176.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                2476935.3 lps   (10.0 s, 7 samples)
Process Creation                             133218.5 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  56335.8 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                  4110.2 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   8197.6 lpm   (60.0 s, 2 samples)
System Call Overhead                        6140801.5 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   99365861.5   8514.6
Double-Precision Whetstone                       55.0      29766.9   5412.2
Execl Throughput                                 43.0      39451.2   9174.7
Pipe Throughput                               12440.0    7489176.1   6020.2
Pipe-based Context Switching                   4000.0    2476935.3   6192.3
Process Creation                                126.0     133218.5  10572.9
Shell Scripts (1 concurrent)                     42.4      56335.8  13286.7
Shell Scripts (16 concurrent)                     ---       4110.2      ---
Shell Scripts (8 concurrent)                      6.0       8197.6  13662.7
System Call Overhead                          15000.0    6140801.5   4093.9
                                                                                  ========
System Benchmarks Index Score (Partial Only)                         7928.8

Gordugunuz uzere DUAL 5520 skoru "7928.8" , Dual 5620 skoru (en yukarida var ) "8664.6" , bu demek oluyorki yeni cikan islemcilere hemen yapismamak lazim ve bir takim testlere tabi tutmak lazim , ufak performans farklari icin degmez.

iyi calismalar.

09-08-2010, 19:00:21

#9

Kosanbit

Tabiki pratikteki yansımalarıda önemli. Teoride fazla fark yok gibi gözüksede pratikte nasıl değişir sonuç, onu da test etmek lazım