RE:(2) (2) [PATCH v3 01/14] fs: Move enum rw_hint into a new header file

[Daejun Park <daejun7.park@samsung.com>]

Hi Bart,

>On 10/31/23 23:39, Daejun Park wrote:
>>> On 10/30/23 04:11, Kanchan Joshi wrote:
>>>> On 10/18/2023 2:17 AM, Bart Van Assche wrote:
>>> Thanks for having taken a look at this patch series. Jens asked for data
>>> that shows that this patch series improves performance. Is this
>>> something Samsung can help with?
>> 
>> We analyzed the NAND block erase counter with and without stream separation
>> through a long-term workload in F2FS.
>> The analysis showed that the erase counter is reduced by approximately 40%
>> with stream seperation.
>> Long-term workload is a scenario where erase and write are repeated by
>> stream after performing precondition fill for each temperature of F2FS.
>
>Hi Daejun,
>
>Thank you for having shared this data. This is very helpful. Since I'm
>not familiar with the erase counter: does the above data perhaps mean
>that write amplification is reduced by 40% in the workload that has been
>examined?

WAF is not only caused by GC. It is also caused by other reasons.
During device GC, the valid pages in the victim block are migrated, and a
lower erase counter means that the effective GC is performed by selecting
a victim block with a small number of invalid pages.
Thus, it can be said that the WAF can be decreased about 40% by selecting
fewer victim blocks during device GC.

Thanks,

Daejun

>
>Thanks,
>
>Bart.