Author |
Thread Statistics | Show CCP posts - 9 post(s) |
|

CCP Habakuk
C C P C C P Alliance
1300

|
Posted - 2015.04.29 13:00:40 -
[1] - Quote
Hi,
For the next few days we are testing our Brain in a Box branch on Singularity - Singularity will be updated in the next few hours for this. Tomorrow on Thursday we will be running a mass test at 17:00 EVE-time.
Disclaimer: We have still a bunch of known bugs (and probably even more not known bugs) and we are still several months from release. Your ride on Singularity can be a bit bumpy (but most known bugs should be quite hidden).
Why this test?
- We want to validate our internal performance measurements and see if we missed any performance problems (quite likely) - both for CPU and memory performance
- We want to get many bug reports from you about bugs, which we missed. :)
What has changed? This current test includes Brain in a Box (BIAB), a rewrite of Dogma attributes and a rewrite of skills.
Brain in a Box: On Tranquility all your skills are recalculated each time you jump to a different system and then the skills are re-applied to your ship. With this change all your skill-effects are being pre-calculated on a different server node and then this pre-calculated brain is applied when needed. This improves the performance of jumping through gates and many other session changes (docking, undocking, switching ships in space, being killed).
Rewrite of Dogma attributes: Dogma attributes are nearly all of the values on your ships, modules and similar. Many of those you can see in the show info window in the "Attributes" tab or in your HUD (like your current shield). This rewrite changes how these attributes are stored and calculated on the server and the client. This change was done to allow the implementation of Brain in a Box. If you want to know more about this, feel free to watch the recording of our fanfest presentation from this year. As part of this we also rewrote most of how fleet bonuses work internally and we also had to re-implement the system of heat and overloading modules.
Skill rewrite: Together with BIAB we also moved all skill training to different server nodes. We also changed how trained skills are stored: On TQ all your skills are just inventory items, which are located within your character (=brain). After these changes they are now stored separately as skills. For this we had to change all related systems like skill training, skill queues, implants, attribute respecs, and similar - everything should behave the same, but the code behind it changed quite a bit. These changes should also fix some old issues with skill training (example: skill queue getting stuck, when the skill completes exactly while jumping between systems)
Known issues: The current test build does not include the latests fixes for the Mosaic release - everything from the last 3 weeks is missing. Please ignore any related bugs. BIAB bugs: We have currently about 30 bugs in our defect tracking system. Most of them are very hidden, but here the more visible ones:
- The skill training indicator in the Neocom is not correct
- Various e-war notification are not displayed in the HUD or the overview
- The bonus of control towers is not applied correctly
- Some skill bonuses are not applied correctly (but please send bug reports about this, if you find anything!)
- Fleet bonuses are wrong / lost in some cases (please send bug reports about specific cases)
- Show info window does not show correct modified values for drones
- Everything with missiles and drones is still a bit in flux - but please send bug reports about it
What shall we test? Just for the case that you really want to help with testing, here some areas: 
- Skill training
- All various skill effects (on ships, modules, completely different features, ...)
- Implants
- Ship, ship modules, fitting
- Session changes like jumping, docking, ...
- Fleets
FAQ:
- When will it be deployed to TQ? Not soon.
We are still several months away from being ready to ship. Our internal plans are aiming for a release at the end of summer 2015, but this can change.
- Will this fix all our lag? No. It will help quite a bit for everything to do with jumping and similar - but in our current version there are also some cases which are slightly slower (for example shooting missiles). We working on improving this, but BIAB is not magic, which makes all our lag problems go away.
- Will this change feature x? Very likely no. We have so far only one very minor feature change for fleets (the wing commander receives now fleet bonuses from the fleet commander) and we are not planning any further changes. Further changes might come in later releases, but mostly from other teams.
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|

CCP Habakuk
C C P C C P Alliance
1301

|
Posted - 2015.04.29 13:29:08 -
[2] - Quote
Masao Kurata wrote:Oh god, I'm not sure I wanted to know some of those Dogma implementation details, I would never have guessed it was that bad.
Can we rely on statistics displayed in the client matching their actual performance in game or does that need to be tested too? e.g. if the interface says I have 83% thermal resistance and this matches what I should have on TQ, do I still need to test that I am indeed only receiving 17% of the raw damage from a thermal source?
This is one of the big problems actually. W spend quite some time testing that the information is correct in all places (on the server, in the client internally and in the displayed UI). A good example are the values on drones, which are currently not updated with skills (even on TQ they are only correct in the drone bay and not in space). But: If you are just helping in the mass test - feel free to ignore these details. If you want to go a bit further with testing: Send us details in bug reports if you assume that something does not match. We have then further tools to poke at the values directly.
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|

CCP Habakuk
C C P C C P Alliance
1301

|
Posted - 2015.04.29 13:30:51 -
[3] - Quote
Nalha Saldana wrote:Holy hell, is it just me or is this huge? This is not only a huge blob of text, but the code changes are really quite massive - and bring all kinds of bugs with them.
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|

CCP Habakuk
C C P C C P Alliance
1305

|
Posted - 2015.04.29 14:37:17 -
[4] - Quote
Vincent Athena wrote:It's nice to see this work coming along, it's been discussed for years. But I am left wondering; is this the end when it comes to reducing lag and getting more players into one system?
I know one solar system runs on one core. Code like BIAB moves some of the work of a solar system to a different core. But is that going to be it? Or are there any plans to truly spread the work of running a solar system across multiple cores in a way that allows ever more cores to be added as more players enter a solar system?
Also:
Any idea how high the Jita cap will be after BIAB is live? How big a fleet fight can get before TiDi kicks in? How big a fleet fight can get before TiDi reaches 10%? I am pretty sure that this is not the end.
For everything else: A bit too early to tell. This mass test will help us a bit to make some estimations.
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|

CCP Habakuk
C C P C C P Alliance
1305

|
Posted - 2015.04.29 14:38:34 -
[5] - Quote
bassie12bf1 wrote:Would this reduce TiDi/serverload when massive fleets jump gates at the same time? This is exactly the use case which should benefit most from our changes (at least as far as I know).
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|

CCP Habakuk
C C P C C P Alliance
1325

|
Posted - 2015.05.05 10:53:48 -
[6] - Quote
Thank you all for your help with testing. 
Singularity has now been reverted to be running code very close to TQ. All skill training data of the last few days is gone as part of this (but /copyskills should work again).
CCP Habakuk | EVE Quality Assurance | Team Gridlock
Bug reporting | Mass Testing
|
|
|
|