The tests are slightly unfair as my disk optimizations (I previously used RAID 0 - but when installing everything I messed up, forgot to set my RAID in the BIOS beforehand and Windows doesn't really like you changing from AHCI to RAID after it has been installed/configured). So I forwent RAIDing my drives and keep a USER drive and a WORK drive. My point is that read and write times will be slower. That being said, some things to note:
Large system compilation went from an average of 58 seconds on the Intel build to 72 seconds on the AMD build. No surprise as single thread execution is known to be slower in some cases.
The flip side is WinZip decompresses the WAR files MUCH more quickly in the AMD build. It is so fast I don't have time to do my usual little things (closing my selection windows for the war, emptying the trash, stuff like that). I click "Expand to" and *BOOM*... done. The AMD CPU is practically made for this sort of work tho'...
Starting JBOSS (after a FRESH deploy) the first time on my AMD build took 13 seconds. That's not bad... I've only done it once so far... taking into account slightly slower disk read times that's not too different from the 9 to 10 seconds it took on the Intel build. Again, all of this doesn't surprise me either, since Java doesn't seem to be overly multi-thread friendly (the Java app itself, not its capabilities).
More soon!
1 comment:
Today, a fresh build start of JBoss took 12 seconds... So that will be right about the average. Still amazes me that Winzip performs a LOT better even without RAID.
Post a Comment