These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE Information Portal

 
  • Topic is locked indefinitely.
 

Crius Issues

First post First post
Author
Dareth Astrar
Astrar Logistics and Engineering
#881 - 2014-07-29 08:17:19 UTC
CCP Nullarbor wrote:
Ophelia Valentine wrote:
Echo Mande wrote:
When I try to start a job at a POS with a 'new' BPO type (for example an auguror or LSE; in anycase something I haven't used that session) the input/output selections appear to default to the lowest number hangar partition you have access to and not the hangar partition (or maybe can; haven't tried that) that the blueprint is actually in.

I really hope that this is not intended behavior.


In the previous system you could configure the science/industry window to use the last used input/output hangars for new jobs. Being able to do that again would be very nice since my jobs almost always take their input from one hangar and output to another, neither of which is the first hangar.


It will remember the input / output in your client between sessions, per type at a facility. The first time you pick a blueprint of that type you need to select, after that all of the settings are remembered (runs, output, decryptors etc)


So now we have to set these more times then before, for every type of product BPO we have? That's really very boring, un-fun, and surely more storage work for you as well in the back end, or is this yet more data dumped in our beloved cache, so if we move elsewhere (on to the laptop from the desktop say), we have to go through this process again?

I'm sorry, but I would rather have a user choose a set of defaults, as for most of us we've set it up as the same hangers every time based on the previous UI's convenience/working. It was quicker, and it worked. Why change it? Smile
Dareth Astrar
Astrar Logistics and Engineering
#882 - 2014-07-29 08:32:55 UTC  |  Edited by: Dareth Astrar
Firvain wrote:
Miyazaki Hetoshi wrote:
Not sure if this is a bug, or working as intended;

Copy times have increased by a factor of 5

Light Missile Launcher BPO

Pre-Crius: 20 Copies, 300 runs - 2 days 17 hours.
Post Crius: 20 Copies, 200 runs - 15 days


intended, you only need 1 run for max run invention bpc


Yes, but you need far more if you are actually manufacturing from a BPC as they had stated in some of the blogs as being your way to minimize your risk with production at a POS. Those kinds of times are massively impractical if you are copying to build, and the reduction of runs also impacts the number of character slots being taken up, as you will need to schedule more slot-to-bpc time to produce the same amount that you could have in the previous release.

I understand there is always going to be a difference here, but the thought has obviously been applied to one situation and the other has been forgotten. The build from copies is just as important to me as having copies to Invent from. Perhaps the 1 run per BPC for invention was just too generous an item in some cases, even though I know we are benefiting from it as well, it is causing a bigger problem from the build perspective relating to copies and runs allowed on those copies.

As a bare minimum, the max runs shouldn't just be thought of for invention, but at the bare minimum how an average player could build from this BPC, would it take less then a day? In which case perhaps the runs should be increased to approximate a days run in a Level 5 building character.
Dareth Astrar
Astrar Logistics and Engineering
#883 - 2014-07-29 08:48:44 UTC  |  Edited by: Dareth Astrar
CCP RubberBAND wrote:
I will raise this with the team, I know performance is an ongoing concern and we are investigating smart ways to improve the experience without reducing the access to the all the blueprints you and your corporation own. Keep the feedback coming, we are listening.


Something other then loading everything from the outset would be great! Applying the last set filters before retrieving data, or even if on the first few characters entered into the filter you did a query that resulted in a reduced dataset, that would probably still be a far swifter response time.

Fine, whilst I understand that doesn't meet your newbie user interaction experience, please don't forget many of us that have been around a long time are constantly inconvenienced by the massive delays, and sometimes even the Eve client going non-responsive completely, and windows warning the application has stopped responding.

Perhaps give us the choice of how to do these things, do we want it to load all by default (checked initially) and unchecked for those of us that know what blueprints we are going to use to put a job on, and entering a filter based on the text would be quicker. Smile

The downside is these are all things that weren't such issues in the old UI, so what changed in the data access code that has caused such problems with performance, that would be a place to start looking.

I had previously mentioned that during the SiSi testing, granted I may have forgotten to mention that that 20-25 second delay often experienced was on an i7 system with 32Gb of ram and twin graphics cards. So it's was a fairly significant impact on performance. Smile
Angie Chatter
#884 - 2014-07-29 09:06:51 UTC  |  Edited by: Angie Chatter
Alundil wrote:
Angie Chatter wrote:
CCP RubberBAND wrote:

A video of this would be extremely helpful. Seeing what kind of jumping/skipping you are experiencing would help us tackle the issue.


This simply happens with BP that cant be "naturally" sorted, because they are all the same. Like component, T2, T3 BP all with the same name, ME/TE, runs remaining. Simply put 50 exactly same BP or BPC in the same division, than sort by ME/TE and now install 30 jobs.

Now watch the frustrating result.


CCP RubberBAND wrote:

The gray blueprints should insert into the browser, but this should be predictable based on the sorting of the blueprints.


It is simply not "predictable" if all BP are the same and cant be "naturally" sorted, now after each new job the list resorts however it feels, which results in a random mix of used "gray" and usable BP.

So if the list can't be naturally sorted, plz always put the "gray" items at the end. This should at least work until we get a filter to remove those entries entirely.

thx

In that instance it is probably sorting by item id in the database. Which is, of course, not exposed in the client so it looks like a random sort. Just a guess.


Not really, this would at least result in some parts of the list keeping its order. What happens in practice is that a BP u selected right before the list refresh at top 2 position, now suddenly is not even visible anymore and gets to a 20+ position. It seems the resulting list "sorting" order completely changes every time the list is refreshed, if no real sort criteria is there.

Simply put the system can not handle "sorting" identically entries properly, without "randomly" changing the order after each new refresh/job. The problem is we only have name, ME/TE, Runs, Division as sort criteria. Those are all the same for me normally, if i use the name filter to only get the BPC i want. The ME/TE, Runs, div is always the same, so the sorting needs to work in those situation and produce a stable list.

So plz fix this annoying behaviour, so by default first sort by internal ID and than by the selected criteria and if they are all the same the list should stay fixed after each refresh. So a BPC that was at list pos 3, should stay at pos 3 after a refresh, as long as all BPC are the same. "Unusable" items should also always move to the bottom of the list, if all sort criteria are the same. If i setup 20+ jobs i do not want to scroll down for each new job, only because all those gray entries take up all the upper screen pos list spaces.
Dareth Astrar
Astrar Logistics and Engineering
#885 - 2014-07-29 09:15:12 UTC
McBorsk wrote:
Any chance we could do this?

100 veld = 1 compressed veld
95 Concentrated veld = 1 compressed veld
90 Dense veld = 1 compressed veld

That would give us one big compressed veld market instead of three.


That's a very sensible way of looking at this problem and a nice idea for a solution. It will give them a starting block for a discussion at least, and greatly simplify the compressed mineral markets if we only have 1/3rd the types to review.
Dareth Astrar
Astrar Logistics and Engineering
#886 - 2014-07-29 09:17:40 UTC
Irin Fidard wrote:
I have about 2000 BPO and somewhere arround 25000 Copies...Yes I aware that its because the huge amount of data the client needs to load. but I think it would help allready when it would load "Jobs" instead of "Blueprint" everytime the UI opens.


Another sensible suggestion, as we mostly look at the jobs in progress/needing delivery first most of the time, and then look to installing additional jobs afterwards.
Dareth Astrar
Astrar Logistics and Engineering
#887 - 2014-07-29 09:19:17 UTC
Koenaika wrote:

Obviously you have to hire the crew for your starbase assembly line from the local area, so if demand is high they have to be paid more to live on a dirty space station.


How many thousands of years in the future and you think you still be employing people rather then everything being 100% automated? Really? Smile
Dareth Astrar
Astrar Logistics and Engineering
#888 - 2014-07-29 09:23:33 UTC  |  Edited by: Dareth Astrar
CCP RubberBAND wrote:
Maxpie wrote:



Perhaps I wasn't clear. I'm asking about the unlocking and locking process which currently must be done one print at a time. This is time consuming and tedious, not to mention downright physically painful when you have hundreds of prints locked down that you now need to move to be useful.

If you mean to say that there is an existing solution to this problem, please share it.


If you are referring to corporation blueprint lockdown (I.e. voting) then there are no plans for this team to address this for the Crius release. Though corporation voting in general will need to be looked at as part of the Corporation and Alliance changes planned down the line.

EDIT: Neither are there plans to look at or address the way in which just locking blueprints in containers works at the moment.


No, he was fairly clear first time. He is talking about the process of having to move all these locked blueprints to a different location, and having to lock them all up again. The current process for doing this is manual, and he is talking about physical pain of actually having to unlock them and do it all over again.

I think what this raises is a sensible point about if a corp wants to move these locked items, they should be able to move them in a big batch without it being a gigantic pain. i.e. a better way to manage the controlled resource if moving from, for example say, station to a pos with the recent changes.
Dareth Astrar
Astrar Logistics and Engineering
#889 - 2014-07-29 09:47:09 UTC
Serinas Setzuni wrote:
Bug with delivering manufactured ships into arrays:

They are using their assembled volumes instead of their packaged volumes on delivery, causing an 'out of space' error where there shouldn't be any issue (The jobs are 280k m3 and 300k m3 delivery, and there's 1.9m m3 remaining in the ship array) It is erroneously stating that they are over 3m m3 in size.

Petitioned and bug reported. Other items do not appear to have this issue so far, only ships. Would be nice to have more storage ability in the ship assembly arrays as well.

Idea: Dragging a packaged ship to a ship maintenance array automatically assembles it if there's enough room for the assembled ship there? Or a question box asking if you want it assembled?


This previously allowed the jobs to be delivered to the ship array consuming more space then was available, but you were simply not able to add anything else to the hangers until you had emptied the delivered goods.

I don't see why there was a need to change. This has always been an issue with larger output items.
Dareth Astrar
Astrar Logistics and Engineering
#890 - 2014-07-29 09:59:54 UTC  |  Edited by: Dareth Astrar
Veinnail wrote:
Aluka 7th wrote:
I managed to find quiet system. On star map is shows that in last 24hours only 9 R/D jobs were installed and that index is 0.00.
BUT Industry index (red bar) in my facility rose to half and cost already increased 5 fold.
Since Crius, I have installed 6 jobs that lasted 16h and 4 jobs that lasted 7hours and that got index to half red?!
Little steep IMHO.
Is that intended? Star map shows low activity but facility cost being at half red already?!


They need to implement a threshold where the base costs stay "understandable" untill the system gets heavily loaded.

this system is built as an isk sink in its current state.


Really, major manufacturers don't move their Factories around Countries every month. That would really test Toyota's JiT and other manufacturing processes to the extreme! Smile

For those of us that were into efficient production, loading all available labs and factories constantly, this is going to be a constant pain to consider relocation all the time, and extremely unproductive when they do.

There needs to be a more gradual increase to elements, the current implementation appears to be far too rapidly adjusting, which would mean any efficient use of a system with even 2+ characters would probably get a big and noticeable change in a few days.
Dareth Astrar
Astrar Logistics and Engineering
#891 - 2014-07-29 10:04:47 UTC
Veinnail wrote:
Ezeus wrote:
When you announced the changes on the BPO ME and PE you basically said that existing bpos that were researched would stay the same or get better, well that is not the case.

Just tried to start a job with my archon BPO which was ME 5 and wasting only 1 drone bay,

Now its me 9 and wasting
1 cap engine
1 power generator
1 battery
1 construction parts
1 jump drive
1 armor plates
1 corporate hangar
1 ship maintenance

To me this BPO actually got worst than what it was

this BPO is now not competitive on the market due to the patch.

the last line,
this is a widespread concern.

ofcourse we're expected to adjust our current stocks value and push on ~dealwithit?~

rounding up damaged the market.


Whilst I agree, the maths buy which those little changes make it non-competitive has perhaps some other faults. If you aren't adding time and costs in for the production of those components, and adding a profit for each of them, then I hate to break it to you the time taken to make a profit that recovers your initial investment in the blueprints and the associated research will likely be in the many thousands of days.

I had posted some maths regarding the current state of Orca production and Jita prices, assuming all components completed instantly and ship assembly was instant, so that we didn't take into account market fluctuations (or more realistically entropy, the constant downward influence in eve prices prior to this patch), and the RoI was silly given the ship profit margins. I think the isk / factory hour was something like 1k. I think this was as part of the Freighter changes a few months back, but obviously that methodology probably no longer is accurate.
Dareth Astrar
Astrar Logistics and Engineering
#892 - 2014-07-29 10:09:45 UTC  |  Edited by: Dareth Astrar
CCP Nullarbor wrote:
Rapscallion Jones wrote:
I'm begging for a Dev reply, this was originally asked in the feedback forum, but no one's answering over there.

With regards to jobs in progress pre-Crius, I figured out (with much trepidation) that I could take down the labs and tower without interrupting the jobs in progress. Clear, concise language on this in the patch notes would have been greatly appreciated.

However, I have a new source of anxiety. I'd like to move those BPOs with the jobs in progress to my new corp home. Will doing so interfere with the pre-Crius job completion? Logically I'd think not, but I don't like to assume anything with the way things have gone so far.

Please provide a simple yes or no, not a description of how it will work. The previous answers on this topic have been unnecessarily descriptive without providing a concise answer.


We have been fixing some issues related to pre-crius jobs installed where the blueprint was in a station and the job was installed in a POS. These were likely causing some confusion but look to be all fixed up now.

If you have a blueprint currently visible in your hangar, you can do whatever you like with it as it should no longer be associated with any jobs in progress.

If you had a POS structure with jobs installed using remotely accessed blueprints in a station, all these active jobs should now be moved to the station in system. In this case you may safely unanchor the POS structure EXCEPT if you have an active super capital job. Super capital producers in this situation have already been contacted about this particular scenario and our GMs are handling those cases.

Is this description clear? I am happy to answer any other specific questions if not.


May I ask what you are intending to do about jobs that were installed pre patch that have 30-50 days on them for 1 level of TE gain, possibly from zero or low figures, on capital items. Obviously with the changed curve now, surely the end time for these items should have been adjusted accordingly, but all still remain in that long period.

There have been a number of questions along this lines which haven't actually been answered as yet.

Elisabeth Jane wrote:
Question: ME jobs installed before patch are running a very long time now, some 26+ days, if I would do the job now, it would be only 2 days +/-. is this going to get fixed or am I out the install cost and need to re-run the jobs or wait 26 +/- days?


Aluka 7th wrote:
I think BPO in research is converted wrong, what should I do?

I have put BPO in station into TE research from 0 to 1 on July 14th and it should exit Aug. 2nd. so 10 more days out of 18 days that are required. On exit from research, TE -1 should be converted to TE -10% in new system based on dev blogs BUT in outcome window in industry it says -2%.

If that is true then its better to stop research and put it again to -2% because that requires in new system only 5 hours. Which makes no sense. What should I do? Stop and put it into research to -2% which lasts only 5 hours or wait 10days and again get -2%.

Please DEV, will it exit as TE-10% after 10days and industry window shows wrong or what can I expect?!
Should I wait 10days then petition?


I would have thought the script could have called a stored-proc to convert the current values (ME or TE) and end expected value, and adjust the End Time accordingly.
Dareth Astrar
Astrar Logistics and Engineering
#893 - 2014-07-29 10:25:31 UTC
Mara Pahrdi wrote:
Zumatra Huri wrote:
Note: what exactly means when ccp say: Crius has been successfully deployed on July 22nd???

Reeading as "We are not going to roll it back, no matter how bad it actually is".

Anyone else suffering from terrible performance issues up to the client getting unresponsive and even crash when trying to install a blueprint?


Yep, had all of that. Whilst I'm sure they are probably looking into the performance issues, I'm unsure if they've logged much on the crash front. I've only had it happen once that the client completely stopped and died, but if you get any useful reproduction steps for them, I'm sure they'd be happy to receive them.
Dareth Astrar
Astrar Logistics and Engineering
#894 - 2014-07-29 10:30:21 UTC
Show Info window with the Required Input Materials could use some localization to give thousands seperator, to make it a lot easier and clearer when reading items information.

I was sure we used to have this, didn't we?
Dareth Astrar
Astrar Logistics and Engineering
#895 - 2014-07-29 10:37:00 UTC
CCP Arrow wrote:

Dareth Astrar wrote:
Right-clicking on a blueprint and > Move to > [Choose Corp Hanger] no longer exists!
....Please don't say move it from inventory, as you obviously can't see the runs remaining.


When we added the capability of seeing blueprints inside containers in the blueprint browser, the 'Move to' became a very tricky option to keep because then the dropdown would also have to show all the containers within all the divisions and it became a massive unusable and unpredictable dropdown list.

In an effort to solve this we added the ability to drag blueprints from the blueprint browser to the appropriate location in the inventory window, picking any destination you want, containers too, instead of just corp division roots like before. This way you still use the blueprint browser to pick the blueprint to move based on stats you have available there like runs but can drag it directly to any location you want in the inventory instead of using a limited options dropdown. So it is a different method but we hope the added capability it provides will make it worth your while to use instead.

Let me know how it works out for you Dareth and thanks for your feedback.

o7


Hi Arrow,


A bit of mixed results. The screen seems to regularly refresh itself, so selections not used within 10-15 seconds seem to reset.

Group select some items (click bottom on, then shift-click the top one, then go and CTRL+click another to select it, then doing the same to one more will reset the shift-click selected items, so in effect you only have two items selected now).

Also after the move the UI goes all loading swirly for 22-25 seconds, during which time I cannot interact further.

When you are selecting a group of items, it still changes the top panels displayed blueprint. This might have something to do with the items being reset, but it really should ignore SHIFT/CTRL/ALT type key pressed in accordance with clicks, otherwise that's causing both a bit of lag and also possible problems attempting to group select things.

To be honest this still took me significantly more time to interact with then I would like.

Please give it a go on say a couple of hundred BP of mixed ME/TE levels, and try to sort out those that have been suitable perfected and those that still need work, and move those that no longer require research from a pos array, and order them some way so they are mixed up so that you have to use combinations of Shift and Ctrl clicking, and you will see what I mean.

I hope that is helpful. Smile
Dareth Astrar
Astrar Logistics and Engineering
#896 - 2014-07-29 10:47:39 UTC
Veinnail wrote:
Your cost index is too steep.

also, it is a great intel tool for finding where supercapitals are being built. thanks for eliminating the need to scout.

edit:
Quote:
The Market is not ready at the moment. Please try again later.


I have to admit Veinnail, one I hadn't even considered. That's a very good point, yet another intel disclosing item which hasn't required your enemies to do any form of in-game investigation.

I had thought when I mentioned they would have to give out more information via API to accommodate for their change in calculations it was going to be cause for concern on the load applied to the API servers, but this one related to Cost-Index adjustments and increases is an obvious major intel disclosure issue. Now it's out there, I doubt they'll be able to take it back.

Obviously moving that kind of production constantly isn't possible, ohh dear this is a rather major game play issue affected.
suid0
Pandemic Horde Inc.
Pandemic Horde
#897 - 2014-07-29 10:53:36 UTC
Dareth Astrar wrote:
CCP RubberBAND wrote:
I will raise this with the team, I know performance is an ongoing concern and we are investigating smart ways to improve the experience without reducing the access to the all the blueprints you and your corporation own. Keep the feedback coming, we are listening.


Something other then loading everything from the outset would be great! Applying the last set filters before retrieving data, or even if on the first few characters entered into the filter you did a query that resulted in a reduced dataset, that would probably still be a far swifter response time.


^^ this... Ideally I'd like to see it not load anything on UI load and have an additional dropdown with Public/Corp/Alliance.

After selecting pub/corp/alliance the job location dropdown is then populated with the arrays available and print counts. Then when you select an array it will populate the blueprint window with the contents in that array/station/hangar.

To maintain the "I'm new to industry and super unorganised, please show me everything I own" theme you could still have an option in the dropdowns for 'all available'.

Currently the new interface is a lot less clicks (if you exclude all the reclicking due to interface lag etc) but takes much longer to actually start your full batch of jobs.

the entire enemy support fleet is dead except for one interdictor a titan could easily finish off with drones  - Commander Ted

CCP Greyscale
C C P
C C P Alliance
#898 - 2014-07-29 10:57:52 UTC
Uff. Update to previous statement: blueprint data will likely now go out on Thursday due to "technical reasons" (I think it involves branch structure). Sorry about the delay. This may mean that we fix T1 materials in T2 items at the same time, though, assuming I can make everything work today.
Dareth Astrar
Astrar Logistics and Engineering
#899 - 2014-07-29 11:00:12 UTC
CCP Greyscale wrote:
Hashi Lebwohl wrote:
Small rounding issue.

Making 1320 Quantum Microprocessors using a 10/20 Blueprint in a station required 7129 Phenolic Composites

Formula I've been using is this:

=MAX(CEILING(Base_units_required*(1-ME/100)*Product_units*Location_bonus,1),Product_units)

= 6 * (1 - 0.1) * 1320 * 1

= 7128

I do not see how the client has rounded up to 7129




Floating point math is basically magic.

(Nullarbor is fixing this.)


If double or float are used yes, but this is precisely why Decimal types came into existence because accounting/financial rounding has long been an issue with float/single/double types (base 2 math based types) in computing.

Decimal may take a little extra cpu time, but the end figures are always correct from base 10 maths perspective.

Might be worth making those adjustments to used types to get the correct figures. This might resolve many of the other mentioned rounding issues experienced by players. Smile
zeahg
Caldari Provisions
Caldari State
#900 - 2014-07-29 11:02:38 UTC
ESS Units are not showing the % increase of return. (The % increase does not appear to be working at all)