Maoni's WebLog

Hi Maoni,

I ran across the following statements:

"Allocating long lived objects at application initialization will ensure that the objects need not be compacted when garbage collection happens. When garbage collection happens, the GC compacts the memory to generate a contiguous space of free memory.

Allocating long lived objects at the application initialization has two benefits: ensuring that the objects survive the lifetime of application thereby reducing the number of Gen 2 collections and avoiding the need for GC to compact the memory"

I thought long lived objects would eventually be promoted to gen 2 if memory pressure occurs and that the only thing that would benefit early initialization is for object pooling ?

Can you comment on the statments ?

Thanks

Hi Maoni,

Thanks for the confirmation. When I first read the statements, I knew it was incorrect but I wanted to get some pro to verify it.

As for the source of the statement, I am more than happy to disclose to you *privately*. Do you proof-read blog entry before publishing to the public ?

I don't want to get myself in trouble :-))

At any rate, I can even attaching the 50+ pages of the document via email, etc. I think you would be suprised by the source. The problem with this document is that some of the statements are true !!!

Thanks

That's some public article. It was from some *big* company even though the article was kind of old (2003) and was a beta version. You could download it freely.

Because it is a public article and from a *big* company, I am just worried that it would make someone looks bad (even though it is a beta) ???

It is up to you.

Thanks