Technical Support


uiWebPrevious1uiWebNext

#1 [en] 

I think most everyone has noticed by now that many of your favorite dig spots have moved.  Some of my favorites are now found 40 - 60m away when I am standing on the flags that have served me well for the past 5 years or so (post merge) .... and for the PR ones much longer than that.

Many mob spawns have also moved (i.e. najabs and kinchers to e alesser extent in void) some have moved more than others ... some seem to be right were we always remember them.  Arionasis took the time to send a detailed ticket response much of it being above my "nerd skill" level.

The short of it though, to my understanding, is that most of the game code was developed pre-2004 and I am sure all would agree, it's a good thing that the server is no longer running on circa 2003 equipment.  Along with the hardware, Operating Systems have changed and new programming tools have been developed to go along therewith.

Prior to patches being released, I imagine that when the game software is recompiled, the use of more modern / powerful / accurate compilers will result "changes" and the node / mob shifts are apparently the result of this.

Ari intends to do a detailed forum post, but in the meantime, wanted to have a place for players to report their observations and any suggestions etc. to assist other players in dealing with these shifts.

Also would invite anyone with a better understanding of the issue to join in, as i'm anxious to learn more.

---

#2 [en] 

My digging spot haven't moved at all, maybe you are just becoming crazy :-)
As for monsters I haven't noticed anything special, but I'm not hunting a lot atm

Newer compiler might produce different machine code, but the behaviour of the code is still the same, whether it's compiled with a 1990-era compiler or a 2017-one, thankfully. There is no such thing as tooling changing the monster spawn.

The two possibles explanations are someone in fact moving the spots (but this can only be done with a server reboot), or much more likely, confirmation bias

---

#3 [en] 

Lopyrech
Newer compiler might produce different machine code, but the behaviour of the code is still the same, whether it's compiled with a 1990-era compiler or a 2017-one, thankfully. There is no such thing as tooling changing the monster spawn.
32bit app has access to 387 fpmath, but 64bit apps are using sse emulation. Compiler switches again modify sse behaviour.

So 32bit gives different results than 64bit and 64bit gives different results based compiler switches. Its not like 2+2 = 3 all of the sudden, but ryzom does not use integer part much from those expressions.

Thats why weather is full integer math (hash) now.

---

Hello!

#4 [en] 

Ryzom E-Mail Support Responses

"I have experienced the same thing with my own characters. The placement of forage nodes are functionally generated and they are influenced by a number of compiler options. The reason for the stability over a long period is most likely due to the compiler options remaining constant. However, recently there have been a number of server patches which have used different compiler options, so we are experiencing these unusual offsets."

"It is likely that the compiler options used will stabilise again and we should be able to reply upon our forage spot markers. Obviously with all the recent changes to both the client and server code, I cannot give you any accurate time as to when this will happen."

"For those players who are popping every forage node in a 60m radius, they are unlikely to notice the movement of the forage spots. It is only when you are popping with no range or up to around 20m that it becomes seriously apparent."

From my experience, when i stand on a  flag I have been using since 2005, a flag which is next to a white dot on the grass that isn't big enough to fit 2 toons, and I cant get a pop or a message unless I move > 40m, that seems different at least to me.

Anxious to see effect of new patch.

Edited 3 times | Last edited by Fyrosfreddy (7 years ago)

---

#5 [en] 

Karu
Lopyrech
Newer compiler might produce different machine code, but the behaviour of the code is still the same, whether it's compiled with a 1990-era compiler or a 2017-one, thankfully. There is no such thing as tooling changing the monster spawn.
32bit app has access to 387 fpmath, but 64bit apps are using sse emulation. Compiler switches again modify sse behaviour.

So 32bit gives different results than 64bit and 64bit gives different results based compiler switches. Its not like 2+2 = 3 all of the sudden, but ryzom does not use integer part much from those expressions.

Thats why weather is full integer math (hash) now.

You're right, I was a little simplifying, but the difference only matters for very precise calculations like weather. The spawn algorithm is relatively easy, it's mostly a random position in polygon - so not much float calculation, it's maybe a .1 difference in position (as you said, we are not even close to 2+2=3), which is irrelevant (both for monster and node shift). Unless the spawn algorithm result in undefined behaviour (but as far as I've seen, it's not the case), it should not create a visible change.

So it's 100% confirmation bias and spot beeing quite big, or maybe ulukyn is playing with the spot for fun :-)

---

#6 [en] 

Well from 1st glance, it would appear that the mobs have shifted to their usual spots ... at least the jabs at PS did.    Dunno who fixed what ... but before shard went down things were "off" and 10 minutes later, all appeared right with the world again ... yeay.

Will follow up and see if the other observed shifts have gone back to normal.

Edited 2 times | Last edited by Fyrosfreddy (7 years ago)

---

#7 [en] 

I'm still seeing mobs, where I've never seen them before.
Like the Yubo spawn right at the Ranger path in Imperial Dunes, hope they stay there, they are so cute. :D

#8 [en] 

Hmmm... I've seen the Yubos there on and off for the last year (IRL) or more. I think it's a pattern on the Jena Year rather than the "short year."

---


Remembering Tyneetryk
Phaedreas Tears - 15 years old and first(*) of true neutral guilds in Atys.
(*) This statement is contested, but we are certainly the longest lasting.
<clowns | me & you | jokers>
uiWebPrevious1uiWebNext
 
Last visit Thursday, 28 March 08:49:56 UTC
P_:

powered by ryzom-api