Archive: 18 posts

---

Hi guys, I may as well do my own benchmarks, but maybe some of you have already knowledge on this: Do sensors with a big radius take more thermo space? (e.g. 5000 max, vs. 5) Do sensors who output closeness take more thermo space, than those who just output closest tag's output value? Do sensors that have custom angles and no-detect zones take more thermo space? Or may be less? I'm guessing it really doesn't matter, LBP2 may be using the same code path for every single scenario, than again, maybe it is optimized for some common scenarios. Oh, btw, do disabled chips save thermo / reduce local complexity?

2011-07-07 22:43:00 Author: hesido Posts: 166

---

No. not really

2011-07-07 23:15:00 Author: zzmorg82 Posts: 948

---

Disabled chips do reduce complexity, compared to its logic spread over one chip and simply interrupted. They might also save thermo, as in logic heavy levels in Create you can see the thermometer go up as you expand more chips. Whether that means that in preview or play mode (when neither a stack of chips or a single chip with all logic on board is shown) there is also a difference, I don't think so.

2011-07-08 00:24:00 Author: Antikris Posts: 1340

---

It's odd that open microchips cost more than closed microchips on the thermo. It makes me wonder how they calculate it, since something must be changing when you open/close it. Perhaps it's just the graphical representation of the wires and logic tools that has an impact...?

2011-07-08 01:21:00 Author: SSTAGG1 Posts: 1136

---

I thought it had to do with wire pathing. If the game is working out the wiring on a circuit board it does add to thermo. But luckily it's temporary, I think of it like using an emitter to cheat thermo. I try to node everything I can. As for tags... Hmmm. I can say that the first iteration of the BTTF DeLorean I built was more complex then subsequent versions as I upgraded the logic to be wireless. Didn't really do any experiments on HOW much thermo went down, but it did. I think that I have most iterations of the car. So I can jump back and check it out. It is something I was thinking of for a while... I'm rebuilding and making the logic more efficient for what I hope will be the final time, so it's been on my mind. Especially because I'm also putting notes on everything this time and am hoping that they don't hurt final thermo too much!

2011-07-08 02:39:00 Author: Mr_Fusion Posts: 1799

---

It's odd that open microchips cost more than closed microchips on the thermo. It makes me wonder how they calculate it, since something must be changing when you open/close it. Perhaps it's just the graphical representation of the wires and logic tools that has an impact...? I too think it is the graphical representation is the driving force behind that thermo change upon open / close. I think we need to test the turning on and off, and maybe Mr_fusion can make a comparison test for thermo load on comparable chips with wiring and signaling.

2011-07-08 08:03:00 Author: hesido Posts: 166

---

You shouldn't see any thermo improvement for switching chips on / off, but you will get a performance increase (as none of the gadgets inside the chip need to be evaluated). The changes to sensor settings also shouldn't save / cost you any thermo either, but some may well be more efficient. Edit: These effects are actually quite common - no change to thermo, but you still get a change in performance... The thermo isn't there to give an indication of how well the engine can handle the performance of your level, it's pretty much just a counting up of how much "stuff" you have, which is why open microchips are more costly than closed ones - additional data is required to handle an open microchip than a closed one.

2011-07-08 08:26:00 Author: rtm223 Posts: 6497

---

As far as I know, changing the settings of the tag sensor should not really affect thermo, especially if they are placed in a closed microchip.

2011-07-08 08:38:00 Author: sp0ngyraver Posts: 407

---

Thanks I had been wondering about this too.

2011-07-08 11:46:00 Author: Ayneh Posts: 2454

---

I too think it is the graphical representation is the driving force behind that thermo change upon open / close. I think we need to test the turning on and off, and maybe Mr_fusion can make a comparison test for thermo load on comparable chips with wiring and signaling. Disappointingly (for this anyway) I went wireless much sooner than I thought, so while subsequent versions do use progressively less thermo I cant say 100% how much it is because of wireless signals. Versus less logic components. However after a quick (and extraordinarily scientific) test I will say that wireless seems to use substantially less thermo (remember; when I say substantial I'm talking on a logic scale, so its, pretty small...). I set up 40 basic logic components wired across to another microchip and had 10 of them in a stage. A battery on one chip a transmitter tag on the other. All up 400 circuits (with 800) logic components = about a half full bulb of the thermo (by bulb I mean before the the first line). If I twisted the chips/moved them around It didn't matter how the wires were pathed it stayed the same. Opening the circuit boards effectively doubled the thermo hit (we're still talking about below the first line, but pushing it). Turning off the chips (with a selector) didn't effect thermo. Neither did setting the wiring to invisible. Replacing the setup with wireless added another piece to the logic with the receiver. This time the battery and transmitter were on one chip the receiver on the other. (again 400 of them, but now 1200 components)... Which is where it started to get interesting... as I deleted each wire that ran to the edge of the MC and replaced it with the tag the thermo progressively dropped. The same setup used about half the thermo of the wired version. BUT the interesting part is it uses much, much more when it is all open... I was expecting about a 1/3 more (the extra logic component), but it may have been slightly more (hard to tell) but it did get above the first line and close to the second. so I'm not sure... but I'd say that the range does have some thermo to it. However, moving the sensors in and out of range had no impact on thermo. The only other interesting things I found is that maybe (a big maybe), just maybe a tag (set to on) uses more thermo than a battery (at 100%) when on an open microchip. This was before adding the sensors, so this was interesting, as the tags themselves not track range, but maybe the colour selection costs thermo. Aaaand the other thing is that chips with heaps of outgoing wires to another chip benefit from being placed side by side on another microchip, but it still just isn't as good a wireless. So basic conclusion its the outgoing wire ports on your logic that most impacts the thermo. Hiding wires has seemingly no effect, but hiding the microchips inside other chips so the outgoing ports are not visible does help a little. Wireless seems to be the most efficient when INSIDE a microchip. Switching off a chip has no effect on thermo. So rtm223, the master of logic is right, like usual! If you have a cycling light with 10 outputs on the MC making this wireless and having the sensors visible (in create mode) is worse than just having the wires coming out of the chip. It looks as if it all comes down to what needs to be rendered and a larger MC with outgoing ports needs more graphics. SO make sure you always tighten up those graphics before you finish a level. I want to add that this was pretty interesting and I may delve deeper into it but it was a pretty casual test to begin with. I'm sure there are others that can prove me wrong on most points but hopefully others can confirm my findings?

2011-07-08 13:00:00 Author: Mr_Fusion Posts: 1799

---

Thanks for the insight, guys. I'll also do tests of my own and let you know here. Like RTM says, and what I've been also wondering, is also the performance side of things aside from just thermo, so this may need to be tested using complex logic to push not only the thermo but the performance.

2011-07-08 13:25:00 Author: hesido Posts: 166

---

Wow, that kinda sucks. In stead of the thermo warning us it's fooling us. Makes me wonder, is the thermo just an output of the game engine's load?

2011-07-08 21:43:00 Author: Rogar Posts: 2284

---

Wow, that kinda sucks. In stead of the thermo warning us it's fooling us. Makes me wonder, is the thermo just an output of the game engine's load? I have stated before in other threads about thermo issues, but it seems like the thermo needs to be patched to work properly.

2011-07-08 23:14:00 Author: sp0ngyraver Posts: 407

---

I have stated before in other threads about thermo issues, but it seems like the thermo needs to be patched to work properly. Out of curiousity, what's not working "properly" about it atm?

2011-07-09 00:48:00 Author: rtm223 Posts: 6497

---

Out of curiousity, what's not working "properly" about it atm? Well for one. It CAN be tricked, so that you may put more content in the level than the thermo would normally allow, causing gameplay problems. Basically, the thermo is a bit deceptive. When a thermo is 3/4 full, sometimes no more than one player will be allowed in because the extra sackboys cause the level to overheat, yet some levels have a full thermo and can easily fit 4 sackboys. It's little things like this that can actually make a BIG difference in your creation. Imagine having to delete 1/4 of your level just because it can't be played in multiplayer, that would kinda suck.

2011-07-09 01:30:00 Author: sp0ngyraver Posts: 407

---

I have always considered Thermo the amount that the PS3 has to render graphically, which is why there is a blue Thermo line for music... the amount that the PS3 has to sequence. You do get spikes when a second player drops in as it adjusts to the costume that they are wearing. But it loads into another part of the game engine once it's settled down which is why it normally returns to the original level. I think a sackbot is the same but three players and three sackbots is SIX costumes to render. So a sackbot costs a little more than initially anticipated. If you build a stage with a DCS set to auto-enter and once activated it spawns the sackbot I don't think it's really a cheat, it just a way of giving the PS3 a break to work out what needs to be shown on screen, what belongs to the stage render cap and what belongs to reserved game processing stuff not to do with the level. I think that having four idle sackbots all kitted out in costume and then having four players load in before the get into the invisible chair is where the issues may come from. The may not be visible but the game is anticipating that because they are in the stage somewhere that they need to be drawn and ready for use at all times. Which is, the more I think about it, why some people may be experiencing the issue where they can't join once the level has started; There is no gap in what needs to be on screen so the PS3 spikes thermo and the safety measure comes into play. (the safety measure being the 'THE LEVEL HAS OVER HEATED' message) Taking this further: Using emitters to 'trick' the thermo no longer feels like the right term. The emitters to me now feels more like a compression tool. The thermo will spike when the stage needs to decompress that part of information. (i.e. the emitter emitting). As a creator it is up to us to decide what information needs to be on screen at any one time and when to remove it to allow the PS3 space to render what could potentially be on screen. The more of this you do the longer the stage will take to load for the end user which is why I don't think it's 'tricking' the game, because you can't really hide the initial load time. (I still have never successfully loaded Hansel and Gretelbot). Again, I want to stress I don't know how correct I am so feel free to correct me.

2011-07-09 02:43:00 Author: Mr_Fusion Posts: 1799

---

That was never really trickery, though, was it? https://lbpcentral.lbp-hub.com/index.php?t=23793-How-To-Seriously-Cheat-The-Thermo is trickery. In the first case you're reducing the thermo because there's actually less for the game engine to worry about, in the second case you reduce thermo because the thermo is fooled into forgetting about some stuff.

2011-07-09 12:19:00 Author: Rogar Posts: 2284

---

That was never really trickery, though, was it? https://lbpcentral.lbp-hub.com/index.php?t=23793-How-To-Seriously-Cheat-The-Thermo is trickery. In the first case you're reducing the thermo because there's actually less for the game engine to worry about, in the second case you reduce thermo because the thermo is fooled into forgetting about some stuff. That is exactly what i meant when I said it could be fooled, the fact that you can load more than the game would normally allow could also mean there are other flaws to the thermo. I think we can at least agree there are many subtle nuances to the thermo that we don't 100% understand. If the thermo doesn't need patching, we could at least use some more info as to the subtle details we do not completely understand yet.

2011-07-09 12:25:00 Author: sp0ngyraver Posts: 407

Posts: 1077139    Threads: 69970    Members: 9661    Archive-Date: 2019-01-19

Aus dem Archiv wurden alle persönlichen Daten wie Name, Anschrift, Email etc. - aber auch sämtliche Inhalte wie z.B. persönliche Nachrichten - entfernt.
Die Nutzung dieser Webseite erfolgt ohne Speicherung personenbezogener Daten. Es werden keinerlei Cookies, Logs, 3rd-Party-Plugins etc. verwendet.