tag:blogger.com,1999:blog-34926058.post4393021190919060667..comments2020-11-05T19:55:50.110-08:00Comments on DSO Guy: Ehcache At 2 Billion...Steve Harrishttp://www.blogger.com/profile/17342689039838434111noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-34926058.post-34872995238597724152011-05-23T11:44:20.941-07:002011-05-23T11:44:20.941-07:00Thanks, these (1.3M) number is pretty consistent w...Thanks, these (1.3M) number is pretty consistent with my own (690K, but for less powerful server). The most interesting part of all of this is the fact that pure on-heap works faster (900K) than BigMemory for caches up to 30G in size, but there are occasional GC pauses up to 70 secs.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-34926058.post-9139174315529939632011-04-12T20:17:58.972-07:002011-04-12T20:17:58.972-07:00Hi, when I said symmetric read/writes I mean that ...Hi, when I said symmetric read/writes I mean that the performance of read operations and write operations are the same. In other words, in my testing, access to off heap has the same performance for both reads and writes so the mix has almost no impact.Steve Harrishttps://www.blogger.com/profile/17342689039838434111noreply@blogger.comtag:blogger.com,1999:blog-34926058.post-57506539580431300062011-04-12T16:34:01.077-07:002011-04-12T16:34:01.077-07:00Symmetric read/write means write after read or 50-...Symmetric read/write means write after read or 50-50 read/writes?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-34926058.post-2472628780274507672011-02-09T16:10:54.477-08:002011-02-09T16:10:54.477-08:00Hello,
These tests were done using Ehcache core ee...Hello,<br />These tests were done using Ehcache core ee 2.4 beta2 (enterprise version of ehcache standalone) using BigMemory (One process and all in memory). The pounder I used is open source and highly configurable to try what ever one wants. It's linked to the blog.<br />Hope this helps<br />Cheers,<br />SteveSteve Harrishttps://www.blogger.com/profile/17342689039838434111noreply@blogger.comtag:blogger.com,1999:blog-34926058.post-5796867054971509352011-02-09T15:30:01.862-08:002011-02-09T15:30:01.862-08:00Was this a test with the open-source Ehcache versi...Was this a test with the open-source Ehcache version or DX/EX? It looks like these tests were all done within a single JVM with an in-memory Ehcache instance, is that correct? Are there metrics that compare performance for given cache sizes across the different versions/strategies?Zack Radickhttps://www.blogger.com/profile/15449103183158790776noreply@blogger.com