Apple iMac 27inch 5K Retina Display: i5 vs. i7 >> Performance Comparison via Oracle WebLogic

| June 8, 2016 | 0 Comments

iMac_i7_vs_i5_fevzi_korkutata_weblogic_java_server_loaderYes, the result is obvious and you can say that there is no need to compare them. But, let’s do a deep Middleware analyse and calculate price/cost/performance results.

First of all I have decided to buy a new all in one PC. Normally I use Microsoft Windows PC for my daily business work and quite satisfied with Windows 10. But this PC would be my second PC with the purpose of using at home for my personal researches and developments. So I decided to use MAC OS X at this time. Then visited Apple Store and checked iMac 27inch 5K all in one PC. There are different configuration offers for this PC. At the beginning I thought “This would be my second PC and there is no need to buy the best configuration”. Apple Genius told me, If you want to buy a customised configuration you have to wait about one month” After all, I bought the beginner and cheaper configuration which was “3.2 GHz i5, 1TB Serial ATA Drive, 8GB Memory”.

And then my iMac 27inch 5K Display performance test journey had been started. After the first JSL performance result I changed my mind immediately and gave back “iMac i5″ to Apple store because of weak performance. We can say that WebLogic and JSL (Java Server Loader) performance results on “iMac i5″ changed my opinion for using i5. Now I am using “iMac 4.0 GHz i7″ and this one has a remarkable performance test results and I am really satisfied with the current performance. Yes, I’ve paid additional cost, but I am totally happy with the result and it’s profitable for long term usage. If you’re deciding to buy a new PC think twice and don’t waste your money for long term. If you’re planing to use iMac for your business (I’m using for Java IDE’s, Application Servers, SOA Tools..) the i5 choice is not at your league.

And now I am going to share all the details about performance tests and the results, WebLogic configurations on iMac, Java Server Loader (Runs on WebLogic), benchmark results, iMac configurations…

I have also prepared a video tutorial for these WebLogic and JSL performance tests on iMac. Here is the URL: https://youtu.be/d0rEk9-AZ0k

 

iMac Configurations: 27-inch iMac with Retina 5K display: i5 3.2 GHz >> vs. << i7 4.0 GHz

27-inch iMac with Retina 5K display: 3.2GHz i5 vs. 4.0GHz i7

27-inch iMac with Retina 5K display: 3.2GHz i5 vs. 4.0GHz i7

 

Summary of iMac Hardware Configurations:

Configuration CPU Disk Memory Price
Configuration-1: 27-inch iMac with Retina 5K Display - i5 3.2GHz quad-core Intel Core i5, Turbo Boost up to 3.6GHz 1TB Serial ATA Drive @ 7200 rpm 8GB 1867MHz DDR3 SDRAM – two 4GB $1,799.00
Configuration-2: 27-inch iMac with Retina 5K Display: i7 4.0GHz quad-core Intel Core i7, Turbo Boost up to 4.2GHz 512GB Flash Storage (SSD) 8GB 1867MHz DDR3 SDRAM – two 4GB $2,749.00

 

iMac Performance Test Environment and Applied Tools & Technologies:

Performance Test - Infrastructure

Performance Test – Infrastructure

Software & Tools Version
Oracle WebLogic Application Server 12c (12.2.1)
Java (JDK) 1.8.0_92
WLSDM: WebLogic Smart Dashboard & Monitoring 2.1.8
JSL (Java Server Loader) 0.1.9

 

About the Performance Test Environment for Comparing iMac:

The tests were very realistic because I was planing to buy which iMac hardware I would buy by applying Java Server Loader (JSL) results. I had setup and installed Middleware infrastructure by using Java, WebLogic, WLSDM and JSL technologies. I did these setups for the both iMac PCs. Firstly installed infrastructure and configured WebLogic environment. After all, I’ve deployed JSL(Java Server Loader) for executing hardware performance tests.

 

About JSL (Java Server Loader):

We developed this application for comparing servers, their hardware specs and reporting benchmark reports. JSL is able to run CPU, Memory, Disk, Network, Database performance tests and produce benchmark results. Its TPS results are very reasonable and we like to support our clients for their hardware purchase POCs and supporting them with a scientific view for their decisions.

Now I am going to share technical performance results for the iMac, then I will compare both hardware configurations with a summary table at the end. This method would be more scientific for buying a new hardware :)

Java Server Loader (JSL)

Java Server Loader (JSL)

 

CPU Results: Configuration-1 (27-inch iMac with Retina 5K Display - i5)

Thread Count Thread Per Second Execution Time (ms) Average Thread Time (ms) System CPU Usage (%) JVM CPU Usage (%) Memory Usage (MB) Processor Count
4 0.3889158969372873 10285 10193 55.55144001991602 44.937749054473784 145.25428867340088 4
8 0.39331366764995085 20340 20151 70.91128498135228 58.09528775995481 159.54904079437256 4
12 0.40602266959905264 29555 29349 76.6487136862797 67.66585737673971 175.4078884124756 4
16 0.41385375442952843 38661 38481 81.26672398407277 72.51261931328169 194.96190071105957 4
20 0.4129842240026431 48428 48160 84.04216005055666 75.38408903015842 219.06800651550293 4
24 0.4132942999827794 58070 57603 86.3014415239784 77.18316451328062 252.26954746246338 4
28 0.4132719329318692 67752 67425 88.31892112792674 80.28353104376379 285.0838289260864 4
32 0.41561680130919293 76994 76617 89.33946596337809 81.36712617351365 142.56557941436768 4
36 0.4179097547101912 86143 85748 90.78845456534586 83.9443933696554 145.9521360397339 4
40 0.41752345959938625 95803 95140 91.11640914814872 84.2060627014152 184.3931369781494 4

 

CPU Results: Configuration-2 (27-inch iMac with Retina 5K Display - i7)

Thread Count Thread Per Second Execution Time (ms) Average Thread Time (ms) System CPU Usage (%) JVM CPU Usage (%) Memory Usage (MB) Processor Count
4 0.5009392611145899 7985 7882 30.834053565019946 22.03705235492332 260.72339820861816 8
8 0.8823205029226866 9067 8964 49.69822068090664 45.555005430471276 290.6070919036865 8
12 0.9252833680314597 12969 12903 58.05911989742495 54.38671717218727 306.2667694091797 8
16 0.8597990219786125 18609 18491 65.72167966144768 60.27468743333701 140.9364891052246 8
20 0.8779631255487269 22780 22615 69.77619051737048 64.38297764980659 184.09584617614746 8
24 0.9117848187827673 26322 26197 73.81918297681555 70.06224502447583 218.4161605834961 8
28 0.914017105177254 30634 30383 76.22551270441382 73.27573622211764 183.80804252624512 8
32 0.9067724567866251 35290 35108 79.72376170633343 75.22209955326613 110.73560237884521 8
36 0.9073037955542115 39678 39502 80.28803945178394 77.36507817639252 150.24032402038574 8
40 0.9033219665319212 44281 43801 82.1382915942199 79.79517763873085 255.5305633544922 8

 

IMPORTANT DIFFERENCE: Java sees 8 processor for Configuration-2:i7 and the number of virtual processor is 4 for Configuration-1:i5 (Check JSL’s benchmark results for “processor count” columns above)

 

TPS (Transaction Per Second) Chart:

iMac Configuration-2 (i7) CPU Results

iMac Configuration-2 (i7) CPU Results

 

CPU Chart:

CPU Usage Chart: iMac i7

CPU Usage Chart: iMac i7

 

Let’s compare best CPU TPS results;

Configuration Thread Best TPS (Thread Per Second)
iMac i5 3.2 GHz 36 0.4179097547101912
iMac i7 4.0 GHz 28 0.914017105177254

Ultimate CPU Result: i7 is 2.187 (0.9140171/0.417909) times better than i5 !!!

I’ve done lots of hardware performance tests on both iMac configurations and prepared below summary result table. Let’s check the result table. I gave 1 point to the lower results and calculated second one’s point according to the lower TPS and assigned its multiplier (factor).

 

Overall Performance Results (Summary):

Configuration Hardware Test Type Specs Thread Best TPS (Thread Per Second) COMPARISON RESULT (p=point)
iMac i5 3.2 GHz CPU 3.2 GHz i5 36 0.4179097547101912 1p
iMac i7 4.0 GHz CPU 4.0 GHz i7 28 0.914017105177254 2.18p
iMac i5 3.2 GHz Memory 8 GB 9 2.8239723878255400 1p
iMac i7 4.0 GHz Memory 8 GB 9 5.22344747533372 1.84p
iMac i5 3.2 GHz Disk (I/O) 1 TB Serial ATA 12 16.71309192200557 1p
iMac i5 3.2 GHz Disk (I/O) 512GB Flash Storage (SSD) 36 116.88311688311600 6.99p
iMac i5 3.2 GHz Network (Socket) 10/100/1000BASE-T Gigabit Ethernet (RJ-45 connector) 100 0.6500217757294869 1p
iMac i7 4.0 GHz Network (Socket) 10/100/1000BASE-T Gigabit Ethernet (RJ-45 connector) 100 1.1926485145562753 1.83p

 

Overall Performance Points and Results:

Configuration Total Performance Point Price Result
iMac i5 3.2 GHz 4p $1.799 x Power
iMac i7 4.0 GHz 12.84p $2.749 3.21x Power

As you see, iMac-i7 is x3.21 time powerful than iMac-i5 and the overall result and multiplier value (factor) really makes sense. According to the performance results, the price can be “$1799 x 3.21 = $5774″, but you are going to pay about $2.749 and for the long term usage you’ll save money and time.

My main purpose for using this iMac is “Oracle JDeveloper, Eclipse, WebLogic, JBoss, Tomcat, SOA technologies, preparing tutorial and educational documents… etc. That’s why I am highly recommending an “i7 over 3.2GHz, SSD Disk (must), minimum 8gb Memory” PC hardware to developers and middleware integrators.

If you want to see additional benchmark results for your hardware, please contact to me and give details about your environment. I am able to create any hardware performance tests, create benchmarks and generate fancy reports by using Java Server Loader.

Below YouTube tutorial is the summary screencast for my performance tests and also I explained some technical details about the JSL stress test methodologies.

LinkedInTwitterGoogle+FacebooktumblrEmail
    Share admineer resources...

Tags:

Category: Common, Development, Java Server Loader, Middleware, WebLogic

M.Fevzi Korkutata

About the Author ()

M.Fevzi Korkutata: Deep level technical consultant... Oracle ACE Associate in Middleware & SOA expertise. His specialty is large scale and mission critical live production systems and like to work as DevOps. He knows all kind of application servers and its environments. Software product development, operation management, design, implementation, integration... etc. Korkutata working as "Application and Infrastructure Architect"... Likes to work and communicate internationally, stay connected :)

Leave a Reply

Your email address will not be published. Required fields are marked *