Hi David,
I'm afraid this issue was not yet addressed. In 3.6 we have made few changes on the backend side to ensure review and changeset pages load faster, but we didn't manage to refactor review page to render diffs progressively I am afraid. Keep watching this issue please, so you will get notified when we manage to address this particular issue.
Hope that helps,
Piotr

> From: Sedlock David (LQKG IT RDS)
> Sent: 11 February 2015 15:04
> To: 'pswiecicki@atlassian.com'
> Subject: Crucible 3.6 release notes: "The diff calculation times for the review
> and changeset pages have been significantly reduced"
>
> Hi Piotr,
>
> I find the above statement in the release notes, but the issue that I have been
> following - https://jira.atlassian.com/browse/CRUC-4184 - is still unresolved
> and I can find no Jira issue in the Crucible release notes that addresses
> improvements in "diff calculation times". Please explain.
>
> Are the problems reported in CRUC-4184 fixed or not?
>
> Regards,
> David

This is a blocker for wide deployment in our case, performance on pre-submit diffs for files with 10k+ lines is so poor that our developers are refusing to use Crucible!

1.5 years later and I'm asking the same question: when?
-David

Hi Sten, thanks for the honest reply. We'll be following the issue and probably delaying our purchase till we find it resolved. If you want to experience what our evaluation users experience, go to https://support.atlassian.com/browse/FSH-10000, where I uploaded a video (https://support.atlassian.com/secure/attachment/499883/crucible_vs_fisheye_diff_performance.mp4).

BTW, in general no one here is saying anything negative about fisheye/crucible performance, except for this one case!

Hi adm.jira@lantiq.com,

We are currently working on improving the performance of Crucible and, while I cannot commit to a date as I do not want to build expectations and disappoint you later, I can ensure you that we are working hard to make Crucible faster and better at handling large changesets.

I am sorry to hear about the slowness that your users are experiencing and I can understand the resulting frustration. This particular issue (CRUC-4184) is not trivial because of the inline comments rendering which would require a change of the structure. On top of that Crucible has to deal with more context than FishEye in case of iterative reviews, which involves more processing around the diffs.

I want to be honest with you and this issue is not likely to be solved in the upcoming release but we're looking at other performance improvements that are going to make the product better.

Regards,

Sten Pittet
FishEye / Crucible Product Manager

A speedup of the display is crucial for us: we have users that will not tolerate the typical load time we are experiencing in the diff page in our evaluation of crucible.

It's very simple: if this isn't speeded up to be very comparable to the fisheye diff time, we won't buy.

We had originally opened up support case CRC-1935 on performance issues while working on reviews, and was pointed to this issue.

We have noticed that diffs take a long time to load, scrolling is slow, comments take a long time to create & post. The larger the review or the longer you are working in a review, the slower the performance. This has gradually gotten worse over the past few months; the performance in the past has not been this poor.

We haven't been seeing any memory errors on the server side, so I'd like to know what performance enhancements can be done on the underlying web server?