FOSS gamedev and creative worlds

Posted: **Fri Nov 07, 2014 15:53**

At first, I thought this was an issue with Freeminer, but when I tried the latest development version of Minetest, I found out that this game is the originator of my CPU being overworked. Why is this a problem again, just like it was once last year? Is there a massive memory leak that needs to be fixed?

Posted: **Fri Nov 07, 2014 17:51**

DO visit IRC.

Posted: **Fri Nov 07, 2014 23:24**

Inocudom wrote:At first, I thought this was an issue with Freeminer, but when I tried the latest development version of Minetest, I found out that this game is the originator of my CPU being overworked. Why is this a problem again, just like it was once last year? Is there a massive memory leak that needs to be fixed?

Apparently Minetest_Game has been creating tons of CPU spikes, but other games such as notetopia, Voxelgarden or BFD don't use lots of CPU for some reason. Nobody knows why.

Posted: **Mon Nov 10, 2014 18:28**

Inocudom wrote:At first, I thought this was an issue with Freeminer, but when I tried the latest development version of Minetest, I found out that this game is the originator of my CPU being overworked. Why is this a problem again, just like it was once last year? Is there a massive memory leak that needs to be fixed?

What's your FPS limit? Minetest is mostly CPU-dependent for rendering.

Posted: **Mon Nov 10, 2014 18:39**

I can't see any difference with version 0.4.9-dev and current version, playing with base_game-game.

Posted: **Mon Nov 10, 2014 21:59**

Calinou wrote:
Inocudom wrote:At first, I thought this was an issue with Freeminer, but when I tried the latest development version of Minetest, I found out that this game is the originator of my CPU being overworked. Why is this a problem again, just like it was once last year? Is there a massive memory leak that needs to be fixed?

What's your FPS limit? Minetest is mostly CPU-dependent for rendering.

60 frames-per-second.

Posted: **Tue Nov 11, 2014 10:28**

Could you possibly give us some real data, not your claims, Inocudom?
And by the "real data" I mean:
- your HW specifications;
- an output of "ps -aux" with and without Minetest running;
- a record of your CPU fan sound with and without Minetest running;
- log of Minetest of the time when you did the measurements with ps and recorded the audio.

One more time I'd like to say that Acer Aspire One handles Minetest just fine.
Please, bring in some evidence or shut the *%!№@#%* up.
Really, you've claimed this for over a quarter having absolutely NOTHING to prove your words.

People here and there grow up concerned about the impression the newbies get of Minetest.
And you ruining not only that but the impression of Freeminer as well.

Almost every your statement can fit 1 of the 3 categories (depending on the message it convey):
1st message: It's so-o-o-o heavy/cold/resource consuming/unstable/bad/hot/inevitable/place-a-thousand-more-reasons-to-feel-empty-and-go-whininig.
2nd message: It was very good/wise/nice/insert-any-word-of-false-appreciacion of <Username> to <insert a deed>.
3rd message: would you please look into <insert a pile of links> to make it work with <insert a pile of links>.

Guess what? The first it the largest one.
Or so it seems. (Which makes the real statistics useless - people make conclusions about what they *think* a thing is, not about the thing itself.)
Of course there are multiple reposts - both asked and not, both asked with irony/sarcasm and w/o it. Those aren't your words, but those of other people.

It's good you're trying to help, but telling that 'YOUR cpu revs' is not helping. How a cpu can make ANY sounds in the first place? Ok, I'm trying to bite you for the usage of a metaphor, so you can ignore the last question.

Bring us some helpful data!

Posted: **Tue Nov 11, 2014 14:37**

Inocudom wrote: ...

I don't know of the history regarding this, but thanks for the report. Can you provide some soild data that can be looked at? Also, please include your game settings etc.l.

4aiman wrote: ...

Wow. Maybe (just maybe, and it depends on how you read Inocudom's post) Inocudom could have been less "accusatory" (but I didn't read it that way). But the developers rely on feedback from users -- it's impossible to test every computer configuration. So, I agree that data is required, but surely you could have asked for that without being so inflammatory. Being agressive and defensive is usually not constructive for either party.

I've found that by assuming that the user is correct is most often the best course of action -- they would not raise an issue or voice their opinion if they weren't experiencing something astray or problematic. If it's a problem with their configuration or system specs or something they're doing wrong, then so be it and that will be discovered upon looking into things.

If I got a response like yours I'd never report an issue or concern ever again, and that's not good for anyone. And the thing is there was nothing to get defensive or angry about. If there is a mistake in the source code, then *shrug* it will be fixed. If there was or is no mistake or nothing can be done, then *shrug* again.

Posted: **Tue Nov 11, 2014 16:26**

People need to understand the situation that I am curently living in. You see, I have to be very quiet during the vast majority of the hours of my days, and having my CPU fan scream too much makes that impossible.

My CPU is an AMD Athlon II X4 645 Quad-Core Processor and I have 4 gigabytes of RAM. My configuration file for Freeminer/Minetest is as follows:

Your phone or window isn't wide enough to display the code box. If it's a phone, try rotating it to landscape mode.

Code: Select all: name = Inocudom server_dedicated = false mg_flags = trees, caves, v6_biome_blend, dungeons, v6_jungles disable_fire = true hud_map = true hud_map_tracking = 1 hud_map_above = false hud_map_alpha = 150 hud_map_scale = 2 hud_map_width = 300 hud_map_height = 300 fullscreen = true screenW = 1360 screenH = 768 enable_vbo = true enable_damage = false creative_mode = true address = digitalaudioconcepts.com anisotropic_filter = true aux1_descends = false doubletap_jump = false enable_bumpmapping = true enable_parallax_occlusion = true enable_shaders = true enable_waving_leaves = true enable_waving_plants = true enable_waving_water = true fast_move = false fixed_map_seed = generate_normalmaps = true keymap_backward = KEY_KEY_K keymap_chat = KEY_KEY_H keymap_cmd = / keymap_console = KEY_F10 keymap_drop = KEY_KEY_P keymap_fastmove = KEY_KEY_G keymap_forward = KEY_KEY_I keymap_freemove = KEY_KEY_F keymap_inventory = KEY_KEY_U keymap_jump = KEY_SPACE keymap_left = KEY_KEY_J keymap_noclip = KEY_KEY_X keymap_print_debug_stacks = KEY_KEY_E keymap_rangeselect = KEY_KEY_R keymap_right = KEY_KEY_L keymap_sneak = KEY_KEY_N keymap_special1 = KEY_KEY_O mainmenu_last_selected_world = 1 maintab_LAST = multiplayer menu_last_game = minetime mg_name = v5 mip_map = true preload_item_visuals = false public_serverlist = true remote_port = 30000 selected_world_path = /home/derek/.freeminer/worlds/GreatMagicWorld viewing_range_nodes_min = 155 bilinear_filter = true free_move = false noclip = true sound_volume = 1 trilinear_filter = true fps_max = 60 wieldview_node_tiles = true wieldview_update_time = 2 farmesh = 3 main_menu_tab = singleplayer weather = false main_menu_last_game_idx = 2 disable_wieldlight = false 3d_armor_update_time = 1 mouse_sensitivity = 0.2 mainmenu_last_selected_TP = 1 texture_path = lottarmor_node_tiles = true lottarmor_update_time = 2 #more_threads = 0 enable_node_highlighting = true keymap_zoom = KEY_KEY_Y flowing_sand_disable = 0 classes_default_class = human random_messages_interval = 120 mg_params = {"layers":[{"name":"default:stone","thickness":20},{"height_max":-1000,"name":"default:stone_with_coal"},{"height_max":-50,"height_min":-3000,"name":"default:water_source"},{"height_max":-50,"height_min":-3000,"name":"default:stone"},{"height_max":50,"height_min":-500,"name":"default:dirt"},{"name":"default:desert_stone","thickness":3},{"height_max":-2000,"name":"default:stone_with_iron"},{"height_max":-1,"name":"default:gravel"},{"height_max":-3000,"name":"default:stone_with_copper"},{"height_max":-1,"name":"default:stone"},{"height_max":-5000,"name":"default:stone_with_gold"},{"height_max":-3000,"name":"default:lava_source"},{"height_max":-7000,"name":"default:stone_with_diamond"},{"height_max":-5000,"name":"default:obsidian"},{"name":"default:stone","thickness":20},{"height_max":-10000,"name":"default:stone_with_mese"},{"height_max":-1000,"height_min":-20000,"name":"default:air","thickness":20},{"height_max":-20000,"name":"default:lava_source","thickness":20},{"height_max":-15000,"name":"default:mese"},{"name":"default:air"}]} liquid_real = true

Posted: **Tue Nov 11, 2014 17:21**

I'm not angry a bit, Zeno...
And I know that he IS very useful, cause he always amongst the first ones to complain about an hypothetical issue.
In fact that is the very reason why I'm not angry.

But I AM irritated by his wining.
Why do I call it so?
Because complaining while providing no data is just that - wining.

Just look what did he post: His config of minetest and listed his CPU/RAM w/o ANY (!!!) real speed in GHz or type of a memory (DDR-what?).
How does that comply with anything that will prove the ISSUE even exist?
Where's CPU load? Where's noise sound? Where's his VIDEO CARD info???
Do I really have to google what type of memory does his hypothetical MB can hold?

So, excuse me, but I'm NOT offensive/inflaming. He needs help, needs understanding his situation but can't see neither polite nor impolite request. IDK what may stir him to share.

I'm all ears and willingness to help, but it seems to me that he either do not want to provide data or IDK what.
All I can do with his info is to ask to lower the viewing_range_nodes_min.
I can't even suggest him to uncheck the "shaders" tick of the "smooth lighting" one, cause there's NO info on that. He may use a different minetest.conf w/o even knowing it! (If the game supplies one.)

What subgame does make his CPU to revv (sic!) anyway? I'm failing to have the answer since June.

^^^^^^
That's why I am irritated. And this is NOT the first time I tell him what data we need. And once again I see his config :(
But angry? Nope. Not at all.

Posted: **Tue Nov 11, 2014 20:41**

This is a review of the exact same CPU that I have:
http://www.techpowerup.com/reviews/AMD/Athlon_II_X4_645/13.html

My video card is an EVGA NVidia GeForce GT 610 with 2 gigabytes of RAM. You can read more about it below:
http://www.amazon.com/EVGA-GeForce-2048MB-Graphics-02G-P3-2619-KR/dp/B00847TPH0

My computer is an Acer Aspire AX1420G-U5832. You can read more about it below:
http://www.geeks.com/details.asp?invtid=AX1420G-U5832-FB-R

The graphics chip that my computer came with is an NVidia GeForce 6150SE nForce 430. You can read more about it below:
http://www.game-debate.com/hardware/?gid=780&graphics=GeForce%206150%20SE%20nForce%20430

Posted: **Wed Nov 12, 2014 05:12**

4aiman wrote:I'm not angry a bit, Zeno...
[...]
But angry? Nope. Not at all.[/b]

Ok :) I'm not sure if I was reading an old version of your post or what happened, but it seems better now that I re-read it. Weird, but *shrug*

Posted: **Wed Nov 12, 2014 12:17**

paramat wrote:I ordered the Sedatech mini-PC posted on page 2 in 8Gb RAM 1000Gb HDD flavour, i can do without a dedicated graphics card =) If or when it gets here i will probably need help with installing Ubuntu from DVD disk :/ So this thread has been useful i learnt a lot.

How well is that computer running Minetest these days, paramat? If it is running it well, can you post a link to it?

Posted: **Wed Nov 12, 2014 15:06**

Inocodom, I have made some improvements relating to CPU performance for the client and will make a pull request tomorrow.

Posted: **Wed Nov 12, 2014 16:12**

Zeno wrote:Inocodom, I have made some improvements relating to CPU performance for the client and will make a pull request tomorrow.

Thank you, Zeno. I hope that your pull request is accepted.

Posted: **Thu Nov 13, 2014 02:48**

Zeno wrote:
4aiman wrote:I'm not angry a bit, Zeno...
[...]
But angry? Nope. Not at all.[/b]

Ok :) I'm not sure if I was reading an old version of your post or what happened, but it seems better now that I re-read it. Weird, but *shrug*

As you can see, I haven't edited that post even once :)
Also, while your commit makes the game faster, any error in get_node will crash the whole MT if it'll get merged. Needs testing, I suppose?

A build? Anyone?

@Inocudom:
If that pull request won't help then I'm still waiting for ps -aux output. I hope it does.
Also I've made a bug report today - you may want to check that in order to see what's the difference between "formspecs are buggy" and the stuff that I've posted.
I've tried to supply examples and all that I've found to be related.

You also may want to check this: Magichet system requirements. Note, that for EVERY dug block there's an entity in Magichet. It's a part of a game. So after a while that CPU of mine has to do LOTS of calculations.

While I admit that CPU fan start to spin a bit more faster, I must say that it's perfectly normal when using IGA on a such low-end HW. But even then, my HDD experience a LOT more "pressure" than the CPU when it comes to logging.
Maybe all you need is a change of a thermo paste/glue and a new cooler?

Posted: **Thu Nov 13, 2014 03:12**

4aiman wrote:Also, while your commit makes the game faster, any error in get_node will crash the whole MT if it'll get merged. Needs testing, I suppose?

That's not true -- it won't crash. The exceptions have simply been replaced with a different mechanism (the error checking is still happening). Before the change the exceptions were being used in place of a condition (I.e. the "errors" are valid and 50% of the time they were being caught and used to excecute a different branch; i.e. they were being used as a conditional). This meant that >95% of the time spent in getNode() was unwinding the exception stack etc etc. If you look at the changes more closely you'll see that the "errors" are still being dealt with :)

Examples for handling player movement

+ Spoiler

Edit: If you know of a situation where it would crash I'd like to know, of course :)

Posted: **Thu Nov 13, 2014 04:31**

Oh, now I see...
Maybe it was a bad idea to look at any code when it's 4AM )))))))

How about a speed-up for the SetNode?
FM did smth to speed it up.

Posted: **Thu Nov 13, 2014 16:20**

I think Minetest needs Multithread support... For the moment a P4 (Pentium4) runs Minetest better than my i5 2500k ...

Posted: **Thu Nov 13, 2014 16:53**

Minetest IS multithreaded, it just needs to utilise it more. (which is probably what you meant)

Posted: **Thu Nov 13, 2014 17:58**

4aiman wrote:Oh, now I see...
Maybe it was a bad idea to look at any code when it's 4AM )))))))

How about a speed-up for the SetNode?
FM did smth to speed it up.

Something like this one?

Posted: **Fri Nov 14, 2014 06:47**

Krock wrote:
4aiman wrote:Oh, now I see...
Maybe it was a bad idea to look at any code when it's 4AM )))))))

How about a speed-up for the SetNode?
FM did smth to speed it up.

Something like this one?

Have you even read the discussion, Krock?
That IS the commit we're talking about...
Which is about getNnode, collisions etc, and not about minetest.set_node().
You may want to look at this if you still think that FM has removed that 'try' wrap ;)

Posted: **Fri Nov 14, 2014 22:13**

So is this issue fixed in later versions? I noticed that my processor is getting some really heavy work on it whenever Minetest is started. Anyhow, what is the status?

+ "Offtopic @4aiman:"

Posted: **Sat Nov 15, 2014 03:17**

I have tried a few versions of Minetest and am having the same issue. I have a quad core processor and 6 gigs of ram. I am running Ubuntu Gnome 14.10. I have the cpu ram and temp stats in the top bar. Without minetest running and doing outer things such as watching videos or surfing the web or even using Gimp the cpu seems fine and the ram andd temp stays low. When I play Minetest it starts out fine but gradually increases until the ram is at 95 - 100%. The cpu doesn't go too high but the temp increases by 50%. The frame rate drops to 3 - 5 fps.
I tried different video drivers but no change. I disabled all apps and features I could and no change.
Once I shut down Minetest the ram drops down to around 12 - 14 %. When I start Minetest up again it gradually goes up again.
Hope this helps. If you need more info let me know.

Posted: **Sat Nov 15, 2014 09:19**

rubenwardy wrote:Minetest IS multithreaded, it just needs to utilise it more. (which is probably what you meant)

Only map generation can be (partially) multi-threaded, not mods, not stuff like entity management. Increasing num_emerge_threads may cause issues with some caves being cut out early, also concurrency often happens so a lot of processing power is lost.

Posted: **Sat Nov 15, 2014 13:27**

I've been experiencing something similar in MT 0.4.10 when using mapgen v7.

Your phone or window isn't wide enough to display the code box. If it's a phone, try rotating it to landscape mode.

Code: Select all: 00:24:59: INFO[main]: Pointing at [node under=-1267,7,2735 above=-1267,8,2735] 00:24:59: INFO[EmergeThread0]: Map::getNodeMetadata(): Need to emerge (-84,-8,177) 00:24:59: VERBOSE[main]: OpenALSoundManager::maintain(): 2 playing sounds, 22 sound names loaded 00:24:59: VERBOSE[main]: OpenALSoundManager::maintain(): deleting 1 playing sounds 00:24:59: INFO[main]: OpenALSoundManager: Creating playing sound 00:25:00: INFO[EmergeThread0]: Map::getNodeMetadata(): Need to emerge (-83,-9,177) 00:25:00: INFO[EmergeThread0]: WARNING: Map::getNodeMetadata(): Block not found 00:25:00: INFO[EmergeThread0]: Map::getNodeMetadata(): Need to emerge (-83,-9,177) 00:25:00: INFO[EmergeThread0]: WARNING: Map::getNodeMetadata(): Block not found ... 00:51:05: INFO[EmergeThread0]: Map::getNodeMetadata(): Need to emerge (-84,-2048,177) 00:51:05: INFO[EmergeThread0]: WARNING: Map::getNodeMetadata(): Block not found 00:51:05: INFO[EmergeThread0]: Map::getNodeMetadata(): Need to emerge (-84,-2048,177) 00:51:05: INFO[EmergeThread0]: WARNING: Map::getNodeMetadata(): Block not found

So far I haven't noticed it happening in v6 or singlenode.
Is this the issue that hmmm (kwolekr) has been working on in dev?

I'm running on Debian testing / XFCE4 / ext4
(which also has separate issues with Thunar/tumblerd I discovered in the process)
on an oldish Dell Inspiron laptop
CPU: Intel(R) Celeron(R) CPU B800 @ 1.50GHz
Generic Intel graphics card
Mem: 3gB

Posted: **Sat Nov 15, 2014 14:16**

Memory usage < 10% but top reports CPU usage of up to 175%, which goes down to just over 100% when I attempt to exit the game. Minetest hangs and I kill it at this point, so I have no idea whether the Map::getNodeMetadata() process ever completes.

Posted: **Sat Nov 15, 2014 17:42**

4Aiman: You kinda forgot the best one:
4: "Here's this item that has no bearing on Minetest and that maybe 6 people will use (if we're lucky), let's spend lots of time and effort to implement it/what do you think?"

This is also the best bug thread ever.

EDIT: typo

Posted: **Sun Nov 16, 2014 16:57**

This isn't a laughing matter, guys. In fact, this is very, very, very serious. I should also mention that every issue that affects Minetest affects Freeminer as well if proller doesn't make commits that fix such issues.

Posted: **Sun Nov 16, 2014 17:07**

I can't see any useful information in this thread. What do you mean by working the CPU too hard? Low FPS or high % CPU usage?

I don't see how this is very very very serious, it isn't a blocker. It is a problem, but isn't a serious problem.

FOSS gamedev and creative worlds

Minetest Is Working The CPU Too Hard These Days

Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days

Re: Minetest Is Working The CPU Too Hard These Days