too long moves

Matrix Games Forums

Forums Register Login Photo Gallery Member List Search Calendars FAQ

My Profile Inbox Address Book My Subscription My Forums Log Out

too long moves

View related threads: (in this forum | in all forums)

Logged in as: Guest

Users viewing this topic: none

Printable Version

All Forums >> [New Releases from Matrix Games] >> World in Flames >> Tech Support >> too long moves

Page: [1]

Message

<< Older Topic Newer Topic >>

too long moves - 10/12/2014 10:49:51 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

This is a thread I open in which I may post several posts for moves that take much too long.

The first one is incredible.

Pick a British submarine in Wonsan.
Move it in the Sea of Japan.
This takes about 30s to complete.

But note. This is a submarine, not a transport ; it has no ZOC (even in its hex), so leaving its hex cannot change the control of the hex it starts from or around and it cannot affect the supply status of any land unit around. It cannot cut supplies from the Japanese (because supply is decided by the presence or absence of convoys/transports.) It cannot bring supplies to any unit around the sea area it enters because it is not a transport. So in effect (and this is obvious from the start), it has no effect on supply.

Now, if I were not using limited oversea supply, it would exactly be the same, because a sub does not interrupt supply.

Last, but not least: even if that move could change anything about supply, the result should be immediate, because the sea area already contains units of that side and of the same kind...

So, what the heck is the game mulling about when moving that unit ????

Is it checking whether the sub is in supply at the start of the move ? But that's obvious : it is adjacent to a sea area that traces directly to Australia or India. It should not take that long!

EDIT: next move, load the TRS in Port Arthur with the Eng div and place it in the 3 box. Same thing. This is a TRS: it could change the supply status of allied units around. Could it ? no, there already are a TRS and convoy in the sea area : this new TRS brings nothing new.

EDIT: second such bug with save attached at post 7

Attachment (1)

< Message edited by yvesp -- 10/14/2014 6:20:21 PM >

Post #: 1

RE: too long moves - 10/13/2014 6:23:59 PM

Centuur

Posts: 8802
Joined: 6/3/2011
From: Hoorn (NED).
Status: offline

Yes, calculation time is very long for this game. That's not nice, because it means the supply routines have to be reviewed again...

Attachment (1)

_____________________________

Peter

(in reply to yvesp)

Post #: 2

RE: too long moves - 10/13/2014 6:42:07 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

Well... I know this.

The supply routine don't take the right path: they decide the supply status unit by unit rather than deciding how supply flows by zone.

This prevents giving an important information for beginners : which hex is in supply ? But well, things are how they are...

This also has the effect that whenever you move something that might change the supply status (that was not the case for a sub!), you have to check/find the new path to a primary source for all units (well, you can be smart and reduce the search to a handful of units, but I think this is where the current still algorithm fails)

I believe that the game doesn't keep in memory the supply path calculated for each unit, in particular the path to its secondary source (each source keeping the full path to the primary source) : this would tremendously accelerate the supply routines as they are. More to the point, in such an algorithm, each hex/sea area should keep in memory which units it supplies. I'm unsure it works like that now given the computation times I observe.

Furthermore, the oversea supply path search algorithm is a close parent to the resource transportation algorithm, and we all know how buggy this is! We will have hope on the supply side only and only when the resource transportation will work flowlessly in automated mode...

Almost one year since the game came out, and still so many problems.

< Message edited by yvesp -- 10/13/2014 7:44:31 PM >

(in reply to Centuur)

Post #: 3

RE: too long moves - 10/13/2014 6:54:10 PM

Shannon V. OKeets

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: yvesp

Well... I know this.

The supply routine don't take the right path: they decide the supply status unit by unit rather than deciding how supply flows by zone.

This prevents giving an important information for beginners : which hex is in supply ? But well, things are how they are...

This also has the effect that whenever you move something that might change the supply status (that was not the case for a sub!), you have to check/find the new path to a primary source for all units (well, you can be smart and reduce the search to a handful of units, but I think this is where the current still algorithm fails)

I believe that the game doesn't keep in memory the supply path calculated for each unit, in particular the path to its secondary source (each source keeping the full path to the primary source) : this would tremendously accelerate the supply routines as they are. More to the point, in such an algorithm, each hex/sea area should keep in memory which units it supplies. I'm unsure it works like that now given the computation times I observe.

Furthermore, the oversea supply path search algorithm is a close parent to the resource transportation algorithm, and we all know how buggy this is! We will have hope on the supply side only and only when the resource transportation will work flowlessly in automated mode...

Almost one year since the game came out, and still so many problems.

The program does record in memory all supply paths that succeeded. When supply needs to be recalculated, it checks them first to see if they are still valid.

The program also keeps track of which supply sources are supplying which units. That information is used if a supply source becomes dysfunctional. Then the program only has to find a new supply source for the units using the one that lost its ability to provide supply.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to yvesp)

Post #: 4

Slow Computer? - 10/13/2014 8:27:30 PM

Omnius

Posts: 833
Joined: 6/22/2012
From: Salinas, CA
Status: offline

yvesp,
Are you sure it's the game or do you have an old slow computer that makes moves take so long? I have no such issue moving units, I may wait 2 or 3 seconds but definitely nothing like 30. Only when it's doing a major recalculation of production do I see a 30 second wait.

Omnius

(in reply to yvesp)

Post #: 5

RE: Slow Computer? - 10/13/2014 8:39:14 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

quote:

ORIGINAL: Omnius

yvesp,
Are you sure it's the game or do you have an old slow computer that makes moves take so long? I have no such issue moving units, I may wait 2 or 3 seconds but definitely nothing like 30. Only when it's doing a major recalculation of production do I see a 30 second wait.

Omnius

For what it is worth, my computer has a windows performance index of 7.6.

Intel Core i7-2600K @ 3.4 Gh
RAM 8Go
Windows 7 - 64

It's not top notch, but its far from old (early 2013 I believe.)

And such problems I have only had when I reach the end game, and more precisely when I tackle Japan from a bit too close.

(in reply to Omnius)

Post #: 6

RE: too long moves - 10/13/2014 8:44:07 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

quote:

ORIGINAL: Shannon V. OKeets

The program does record in memory all supply paths that succeeded. When supply needs to be recalculated, it checks them first to see if they are still valid.

The program also keeps track of which supply sources are supplying which units. That information is used if a supply source becomes dysfunctional. Then the program only has to find a new supply source for the units using the one that lost its ability to provide supply.

So that's one possibility less.
In that case I wonder why some moves should last so long.

Still, it would be much faster that the hex/sea areas did also record the units that uses them : in this way, it would not be necessary to recheck many units, but only these few units that a given change might affect. But given the low number of units, rechecking even all units when they already have a valid path should not be long.

(in reply to Shannon V. OKeets)

Post #: 7

RE: too long moves - 10/14/2014 6:31:31 AM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

Move the Russian Koniev HQ from Kirin to the river Liao along the railway line. 30s. There are only about ten units that may have their supply affected.

Attachment (1)

< Message edited by yvesp -- 10/14/2014 7:32:23 AM >

(in reply to yvesp)

Post #: 8

RE: too long moves - 10/14/2014 5:35:07 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

Move the Chinese 2nd Cav Div into the hex NE.
This takes awfully long, even though a simple look at the the topology shows that this moves only put three unimportant hexes in supply.

Move the 5th Chinese Mot east. Same time lag ; but yes, at least this time one unit is put in supply.

Moregenerally, move the Chinese units into Japanese controlled hexes: every time, this takes an incredible time.

Attachment (1)

< Message edited by yvesp -- 10/14/2014 6:39:33 PM >

(in reply to yvesp)

Post #: 9

RE: too long moves - 10/14/2014 8:05:00 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

Move the Peking Warlord in Taiyuan.
This also takes eons, for a move which effect on suply is null.

Attachment (1)

(in reply to yvesp)

Post #: 10

Getting to Same Position - 10/14/2014 11:52:26 PM

Omnius

Posts: 833
Joined: 6/22/2012
From: Salinas, CA
Status: offline

yvesp,
If it's not your computer being slow then it's looking like near the end of a game there are a lot of things being calculated. I have a little older Apple 27" iMac which was top of the line in 2011 when I bought it. I'm definitely not seeing a slowdown like you are when making moves.

However I'm halfway through 1947 and am approaching Japan. I've liberated the Philippines and the Netherlands East Indies and will tackle Formosa next turn. Then the invasion of Japan will come after that. Even with so many units on the map it doesn't take so long to calculate moves, no more than a few seconds. Even production when it takes a long calculation doesn't take 30 seconds, maybe 15 to 20.

I posted a saved game file in the Mods and Scenarios section titled China Sea 1947 Showdown. If you want to play an awesome naval battle already set to go at the point of initiating combat then give it a try. See if you experience long waits making moves with my saved game. That would give us a good common baseline to compare speed of movement.

Omnius

(in reply to yvesp)

Post #: 11

Did the Test - 10/15/2014 12:05:27 AM

Omnius

Posts: 833
Joined: 6/22/2012
From: Salinas, CA
Status: offline

quote:

ORIGINAL: yvesp

Move the Russian Koniev HQ from Kirin to the river Liao along the railway line. 30s. There are only about ten units that may have their supply affected.

yvesp,
I did your test and moved the Koniev HQ 3 hexes to the hex adjacent to the Liao River. It took me 32 seconds to make the move. Most definitely longer than in my game. Is this perhaps due to the new 1.215 fixes for convoying and saving oil? I also got an error message upon opening the saved game about some Soviet unit transporting and getting something lost. Looks like an interesting end game situation.

Omnius

(in reply to yvesp)

Post #: 12

RE: Did the Test - 10/15/2014 5:20:51 AM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

quote:

ORIGINAL: Omnius

quote:

ORIGINAL: yvesp

Move the Russian Koniev HQ from Kirin to the river Liao along the railway line. 30s. There are only about ten units that may have their supply affected.

yvesp,
I did your test and moved the Koniev HQ 3 hexes to the hex adjacent to the Liao River. It took me 32 seconds to make the move. Most definitely longer than in my game. Is this perhaps due to the new 1.215 fixes for convoying and saving oil? I also got an error message upon opening the saved game about some Soviet unit transporting and getting something lost. Looks like an interesting end game situation.

Omnius

Yes, unfortunately.

As I said above, this looks like something is getting amiss when I approach japan from too close. This is not the first time I experience such lags : I had them (before the latest pass on supply ) in a previous game in the same situation. The recalculations took even longer then, and were quite worse in China than at any other place.

Note that I experience no such awfull lag on the German theater of the war (but there are so few German controlled hexes to capture...)

(in reply to Omnius)

Post #: 13

RE: Did the Test - 10/15/2014 5:02:13 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

Another long move, this time from an air unit which, like naval units, cannot change any supply status (under the current circumstances.)

Move the Chinese liberator into the China Sea.

Attachment (1)

(in reply to yvesp)

Post #: 14

RE: Did the Test - 10/16/2014 11:16:15 PM

Shannon V. OKeets

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: yvesp

Another long move, this time from an air unit which, like naval units, cannot change any supply status (under the current circumstances.)

Move the Chinese liberator into the China Sea.

Air units can change supply when playing with limited overseas supply. That's only under unusual circumstances: when a convoy (et al) is being used for supply and the convoy belongs to a major power which is neutral. For example, a US convoy can be used by the Commonwealth for supply through sea area, but not if an air/naval combat unit (belonging to a major power at war with the major power receiving the supply) is present. The rule is actually more complicated than that, but you should get the drift.

===

P.S., I just added code to fix the problem with subs moving out to sea causing supply to be recalculated. You'll see that in 1.2.2.x when it becomes a Public Beta.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to yvesp)

Post #: 15

RE: Did the Test - 10/17/2014 12:48:37 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

quote:

ORIGINAL: Shannon V. OKeets

quote:

ORIGINAL: yvesp

Another long move, this time from an air unit which, like naval units, cannot change any supply status (under the current circumstances.)

Move the Chinese liberator into the China Sea.

Air units can change supply when playing with limited overseas supply. That's only under unusual circumstances: when a convoy (et al) is being used for supply and the convoy belongs to a major power which is neutral. For example, a US convoy can be used by the Commonwealth for supply through sea area, but not if an air/naval combat unit (belonging to a major power at war with the major power receiving the supply) is present. The rule is actually more complicated than that, but you should get the drift.

===

P.S., I just added code to fix the problem with subs moving out to sea causing supply to be recalculated. You'll see that in 1.2.2.x when it becomes a Public Beta.

Yes I know the rule (actually, I always have to reread it precisely to get the exact terms...)
But under the circumstances, that particular rule should not fire :
a simple check on the war status should show that there are no neutral convoy, hence the rule doesn't apply and the move cannot change supply.

I also forgot to send a save (too late now, but you still could check using one of the saves in this thread) : when I railed US HQs near Vladivostok, they ended up OoS, even though the rail line extends all the way to Paris, a primary supply source for US units.
The fact that the USSR and USA don't cooperate is not relevant.
The only rules that would apply to limit this are:
- the USSR disagrees (but this is not a choice that is available in MWIF, so by default it should agree)
- the USSR is not at war with Germany (but that is not the case)

(in reply to Shannon V. OKeets)

Post #: 16

RE: Did the Test - 10/17/2014 1:49:43 PM

Orm

Posts: 22154
Joined: 5/3/2008
From: Sweden
Status: offline

I think that there still is a maximum of hexes/sea area that a railroad supply can pass through in MWIF. I cut in the latest quote from Steve, that I am aware of, and it is a year old now. Anyway. What I want to say is that a supply chain for USA from Vladivostok through USS to Europe is way past the limit and will therefore not work.

quote:

There is a limit on some of the major powers for how far they will search for a primary supply source/port in mainland Asia. There is none for Germany, Italy, China, but for the others I am limiting the search to no more than 40 hexes from the secondary supply source. For Japan it is 60 hexes.

I also suspect that this hex limit for supply is not just for Asia but applies all around the world.

_____________________________

Have a bit more patience with newbies. Of course some of them act dumb -- they're often students, for heaven's sake. - Terry Pratchett

(in reply to yvesp)

Post #: 17

RE: Did the Test - 10/17/2014 4:29:05 PM

yvesp

Posts: 2083
Joined: 9/12/2008
Status: offline

quote:

ORIGINAL: Orm

I think that there still is a maximum of hexes/sea area that a railroad supply can pass through in MWIF. I cut in the latest quote from Steve, that I am aware of, and it is a year old now. Anyway. What I want to say is that a supply chain for USA from Vladivostok through USS to Europe is way past the limit and will therefore not work.

quote:

There is a limit on some of the major powers for how far they will search for a primary supply source/port in mainland Asia. There is none for Germany, Italy, China, but for the others I am limiting the search to no more than 40 hexes from the secondary supply source. For Japan it is 60 hexes.

I also suspect that this hex limit for supply is not just for Asia but applies all around the world.

Oh.

So this is not a limitation of the rules per se, but of the implementation.

How annoying.

But now I understand better why the HQs were OoS...

Yves

(in reply to Orm)

Post #: 18

RE: Did the Test - 10/31/2014 10:24:15 PM

Shannon V. OKeets

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: yvesp

quote:

ORIGINAL: Orm

I think that there still is a maximum of hexes/sea area that a railroad supply can pass through in MWIF. I cut in the latest quote from Steve, that I am aware of, and it is a year old now. Anyway. What I want to say is that a supply chain for USA from Vladivostok through USS to Europe is way past the limit and will therefore not work.

quote:

There is a limit on some of the major powers for how far they will search for a primary supply source/port in mainland Asia. There is none for Germany, Italy, China, but for the others I am limiting the search to no more than 40 hexes from the secondary supply source. For Japan it is 60 hexes.

I also suspect that this hex limit for supply is not just for Asia but applies all around the world.

Oh.

So this is not a limitation of the rules per se, but of the implementation.

How annoying.

But now I understand better why the HQs were OoS...

Yves

I hated to do that but the calculation times for finding supply for secondary supply sources in Asia can be enormous. The rail network doesn't have to be contiguous. There can be breaks in the rail line that the supply path goes around (using the 4 Basic Path hexes available for a railroad supply path). That means the program has to check not only the rail lines themselves, but also branch off up to 4 hexes from every rail hex in the path.

[Personally, I find the US HQs in Vladivostok getting primary supply from Paris rather far fetched (pardon the pun).]

_____________________________

Steve

Perfection is an elusive goal.

(in reply to yvesp)

Post #: 19