[
  {
    "Id": "588888",
    "ThreadId": "251820",
    "Html": "\r\n<p>Hi,</p>\r\n<p>I tested the performance of DotNetZip.GZipStream with CompressionLevel &quot;BestSpeed&quot; and native zlib.</p>\r\n<p>The size of compressed file is the same, but zlib needs 14s and DotNetZip 34s. &nbsp;</p>\r\n<p>Is there any possibility to speed up performance?</p>\r\n<p>Thanks in advance.</p>\r\n",
    "PostedDate": "2011-03-30T08:42:18.143-07:00",
    "UserRole": null,
    "MarkedAsAnswerDate": null
  },
  {
    "Id": "588938",
    "ThreadId": "251820",
    "Html": "<p>Well it depends on the test.&nbsp; Sometimes in perf tests you want to measure a 1-time transaction - \"how long will it take to do X once?\"&nbsp; Other times, in particular when testing for server scenarios, you want to test more-or-less continuous performance.&nbsp; \"How long will it take to do X 10,000 times?\"&nbsp; When testing this, the test needs to perform a \"warm up\" period before the measuring interval.&nbsp; In a one-time test, DotNetZip will likely lag a native implementation by some significant amount. I never measured DotNetZip against zlib, but your quoted figures don't surprise me.&nbsp;&nbsp;</p>\r\n<p>One thing I noticed is that .NET assemblies (including DotNetZip) compiled with the \"debug\" option are significantly slower than those compiled without \"debug\".&nbsp; Check which version of DotNetZip you are using, and also check your own .NET application.&nbsp; This is one sure way to speed up a .NET assembly: compile it as an optimized binary.</p>\r\n<p>As for other optimizations - I don't have any good ideas.&nbsp;</p>\r\n<p>&nbsp;</p>",
    "PostedDate": "2011-03-30T09:56:19.143-07:00",
    "UserRole": null,
    "MarkedAsAnswerDate": null
  },
  {
    "Id": "589314",
    "ThreadId": "251820",
    "Html": "<p>Pity! I already used the \"Optimized code\" option and release version of my assembly and DotNetZip.</p>",
    "PostedDate": "2011-03-31T01:30:00.243-07:00",
    "UserRole": null,
    "MarkedAsAnswerDate": null
  }
]