RFR: 8341334: CDS: Parallel relocation [v9]

[Thomas Stuefe]

On Wed, 6 Nov 2024 12:49:42 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

>> In Leyden performance investigations, we have figured that `ArchiveRelocationMode=0` is saving 5..7 ms on HelloWorld startup. Mainline defaults to `ARM=1`, _losing_ as much. `ARM=0` was switched to `ARM=1` with [JDK-8294323](https://github.com/openjdk/jdk/commit/be70bc1c58eaec876aa1ab36eacba90b901ac9b8), which was delivered to JDK 17+ in in Apr 2023.
>> 
>> Profiling shows we spend time mem-faulting the memory loading the RO/RW regions, about 15 MB total. 15 MB in 5ms amounts to >4GB/sec, close to the single-threaded limits. I suspect the impact is larger if we relocate larger Metaspace, e.g. after dumping a CDS archive from a large application. There is little we can do to make the actual relocation part faster: the overwhelming majority of samples is on kernel side.
>> 
>> This PR implements the infrastructure for fast archive workers, and leverages it to perform parallel core regions relocation. The RW/RO regions this code traverses is large, and we can squeeze more performance by parallelizing it. 
>> 
>> (I'll put some performance data in the comments)
>> 
>> Additional testing:
>>  - [x] Linux x86_64 server fastdebug, `runtime/cds`
>>  - [ ] Linux AArch64 server fastdebug, `all`
>
> Aleksey Shipilev has updated the pull request incrementally with two additional commits since the last revision:
> 
>  - ArchiveWorkers lifecycle: share it globally
>  - Add sanity assert

> > I can rewrite this to make `ArchiveWorkers` globally accessible, let me see..
> 
> Done. It also resolves the "let's not do this on dump path" wrinkle :)
> 
> Updated performance data here: [#21302 (comment)](https://github.com/openjdk/jdk/pull/21302#issuecomment-2388349450)

Yes, I like this better too. Tying the thread pool to a utility class life cycle felt awkward.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/21302#issuecomment-2459680700