#avr | Logs for 2013-02-11

Back
[04:08:05] <Guest5084> hi
[04:12:57] <Guest5084> is someone using her arm?
[04:15:42] <Guest5084> arm really shows high computation performance and is still low power...how is this achieved?
[04:32:17] <tld> Guest5084: What do you mean? Do you mean how does ARM do it, or do you mean how you can use ARM?
[04:32:50] <Guest5084> if you need more performance u cant just increase the cpu frequency
[04:33:03] <Guest5084> because heat dissipation
[04:34:36] <tld> I still don't understand your question.
[04:35:08] <Guest5084> how can arm increase the performance on a single core processor and still reduce the heat dissipation?
[04:35:41] <tld> that's a huge topic.
[04:36:03] <tld> two important factors are low-power design (low power == low heat)
[04:36:07] <tld> and being "smarter"
[04:36:17] <Guest5084> how about multicore?
[04:36:20] <tld> ARMs are clever 32bits CPUs, while AVRs are dump 8-bit micros.
[04:36:50] <tld> multicore is one way to try to be smarter, but you can get heat-issues with too many cores as well.
[04:37:01] <Guest5084> why?
[04:37:09] <Guest5084> they usually run with lower cpu frequency?
[04:38:56] <Guest5084> there are a lot of smart phones with multicore these days
[04:39:09] <tld> not automatically lower cpu frequency
[04:39:15] <tld> and you have twice as many cores generating heat.
[04:39:49] <Guest5084> yeah but the heat is distributed
[04:40:14] <Guest5084> so that u dont need some sort of active cooling :)
[04:44:31] <tld> well, that all depends really
[04:44:43] <tld> and distributing it doesn't make it less.
[04:45:31] <Guest5084> to get the same performance out of a single core the heat dissipation is higher at a point
[04:47:54] <jacekowski> that's why you get thermal throttling
[04:58:15] <Guest5084> jacekowski, ?
[04:59:56] <Guest5084> is the ps3 cell processor multicore?
[05:13:28] <jacekowski> Guest5084: yes
[05:13:37] <jacekowski> 8 core or something like that
[05:13:47] <Guest5084> and ps4?
[05:14:08] <jacekowski> dunno
[05:14:09] <Guest5084> what i read is that they dont use a cell processor
[05:14:30] <jacekowski> ps3 is 7+1 with one of the cores permanently disabled (it's just there to increase production yields)
[05:16:16] <Guest5084> http://www.omnigamer.com/why-the-ps4-need-a-next-generation-cell-cpu/
[05:17:14] <Guest5084> ost industry sources are convinced that Sony’s next generation console will not have a Cell based CPU. In my opinion if these rumors are true, this will be a huge mistake!
[05:18:23] <Guest5084> The PS4 will back the same CPU as the Xbox 720 ... single core?
[05:19:11] <jacekowski> there xbox 360 had multicore cpu
[05:19:21] <jacekowski> so i don't see a reason why 720 wouldn't go that route
[05:19:40] <Richard_Cavell> Is there any word on PS4 CPU?
[05:19:48] <Guest5084> http://www.techradar.com/news/gaming/consoles/ps4-release-date-news-and-rumours-937822
[05:19:50] <Guest5084> here
[05:20:24] <jacekowski> rumors
[05:20:43] <RikusW> they should put an AVR inside :-P :-D
[05:20:45] <tld> a lot of things have changed since PS3.
[05:20:49] <Guest5084> Xbox World reported that the Xbox 720 will pack a 16-core CPU,
[05:21:15] <Guest5084> why dont they use a GPU with more cores?
[05:21:20] <tld> PS3 is *singlecore* from one perspective, with multiple smaller/weaker accelector-thingies.
[05:21:42] <tld> You could view a PS4 based on a single or multi-core CPU and a powerful multi-core GPU as similar.
[05:21:47] <jacekowski> Guest5084: GPUs have been multicore for ages
[05:22:03] <jacekowski> Guest5084: GPUs have multiple stream units and multiple vector units
[05:22:06] <Guest5084> why do they need to many cores in the CPU?
[05:22:49] <jacekowski> Guest5084: to provide more processing power
[05:23:04] <Guest5084> cant they use juse the GPU for most stuff?
[05:23:10] <jacekowski> no
[05:23:13] <Guest5084> why not?
[05:23:30] <jacekowski> because GPUs are slow for things other than graphics
[05:23:45] <Guest5084> other thinks like?
[05:23:49] <jacekowski> AI
[05:23:52] * tld thinks Guest5084 is a North-Korean spy, out to steal out technologies to help their nuke program.
[05:23:59] <Guest5084> haha
[05:25:02] <jacekowski> http://www.vgleaks.com/world-exclusive-ps4-in-deep-first-specs/
[05:25:07] <jacekowski> 4 core (2 core pairs) 3.2 GHz AMD x86 (Steamroller)
[05:25:30] <Richard_Cavell> Interesting stuff
[05:25:31] <Guest5084> why do u need so high processor power for the artificial intelligence?
[05:26:17] <Richard_Cavell> The consoles have always tended to be behind a build-your-own PC
[05:26:18] <jacekowski> you need the CPU for a lot of other things
[05:26:53] <jacekowski> physics simulations
[05:26:57] <jacekowski> ai
[05:27:07] <jacekowski> loading stuff to gpu
[05:27:16] <Guest5084> update: http://www.vgleaks.com/world-exclusive-orbis-unveiled/
[05:28:13] <crazy_imp> x86 and steamroller?! don't think so
[05:28:26] <jacekowski> why not
[05:28:35] <jacekowski> xbox is x86
[05:29:20] <Richard_Cavell> If one of the next consoles were x86 and the other was not, the one that was not would get locked out
[05:29:27] <Richard_Cavell> Developers would go x86 only
[05:29:40] <Richard_Cavell> plus there's the face that x86 is the only bloody technology that can run bleeding edge games
[05:30:20] <jacekowski> xbox was x86 from the start
[05:30:26] <jacekowski> and ps3 went cell way
[05:30:51] <Richard_Cavell> Cell was a bit of a let down
[05:30:53] <jacekowski> and besides, compilers hide most of the difference
[05:31:07] <Richard_Cavell> It was supposed to work as a kind of streaming CPU, where you use the 8 processors in sequence
[05:31:24] <Richard_Cavell> Cell was a waste of bloody time
[05:31:37] <jacekowski> not really
[05:35:50] <jacekowski> they've gave everyone multicore cpu when no game developer knew how to use it
[05:38:33] <Richard_Cavell> that's probably true
[05:38:45] <Richard_Cavell> But Bioshock had like a dozen threads
[05:44:39] <jacekowski> yes, and not really
[05:44:47] <jacekowski> one 2 threads done most of the works
[06:32:54] <Malinuss> Tom_itx, fyi: your programmer works ;)
[07:00:56] <Malinuss> "The default
[07:00:57] <Malinuss> clock source setting is therefore the Internal RC Oscillator running at 9.6 MHz with longest startup time and an initial system clock prescaling of 8."
[07:01:11] <Malinuss> so it's actually running at 1.2MHz, right?
[12:17:18] <Malinuss> haha I can't stop being impressed by the ease of atmel studio. Even changing the fuses, is a gui click and select..
[12:18:20] <prpplague> Malinuss: hehe if you like that kind of interface, hehe
[12:20:17] <sabesto> use debug->start without debugging for fast programming
[12:20:25] <sabesto> Malinuss: ^
[12:21:08] <Malinuss> prpplague, all the fuses are also logically named.. like external_64ms ... seriously, if someone is claiming to not like that kind of interface, they are lying to either others, or themself, or both ;D
[12:21:20] <OndraSter__> indeed
[12:21:29] <OndraSter__> that is because it is based on visual studi
[12:21:30] <OndraSter__> o
[12:21:35] <OndraSter__> which set some level of usability :)
[12:21:39] <OndraSter__> very high
[12:23:04] <Malinuss> sabesto, oh thanks. Right now I'm only programming attinys, so even if I fill up the flash, it would propably take less then a blink of an eye to finish ;D
[12:23:04] <prpplague> Malinuss: i can't stand the atmel studio, but then i can't stand most IDE's of any type
[12:23:45] <sabesto> Malinuss: by fast i mean not so many clicks
[12:24:31] <Malinuss> sabesto, but I just click f5 - one click?
[12:24:47] <sabesto> or keep 2 programs running with programming window open in one/use avrdude in cmd
[12:25:05] <sabesto> no, there is no shortcut
[12:25:27] <Malinuss> I know, but if I click f5, it just compiles and programs in one click?
[12:25:47] <Malinuss> prpplague, really? So not even vim or anything like that?
[12:25:48] <sabesto> f5 starts debugging
[12:25:58] <sabesto> f7 builds
[12:26:17] <sabesto> if you click start without debugging its builds and programs
[12:26:22] <sabesto> *it
[12:27:22] <Malinuss> sabesto, if you click debug, it will actually build and program, too :)
[12:27:57] <prpplague> Malinuss: i do use vim, but not in an IDE style, and vim is far from being a GUI style IDE
[12:28:28] <Malinuss> prpplague, let me guess - your desktop is build up of 4 consoles next to each other. Am I right ? ;P
[12:28:49] <prpplague> Malinuss: pretty close, hehe
[12:29:09] <Malinuss> prpplague, is arch by any chance your OS of choice?
[12:29:48] <Malinuss> *"os"
[12:30:18] <prpplague> Malinuss: don't really have a preferred distro, but i normally use ubuntu 10.04 for most applications
[12:30:49] <sabesto> Malinuss: if the uC and programmer supports it
[12:31:35] <Malinuss> sabesto, well mine doesn't have on-board debugging ( I use mkII copy)... Neverless, if I click f5, it both compiled/builds and programms
[12:31:55] <sabesto> hm, i get error messages
[12:32:00] <Malinuss> really?
[12:32:39] <Malinuss> oh wait. I have my configuration on "Release"?
[12:32:46] <Malinuss> maybe that's why sabesto ?
[12:33:03] <sabesto> dunno, i dont have anything handy to test
[12:33:21] <Malinuss> prpplague, :O the last one I would have guess! Guess I shouldn't persue my dream of being a fortune-teller.. sigh
[12:34:34] <prpplague> Malinuss: i only use ubuntu because that is the distro approved for use in the office and that is what the other internal developers are using, i have no preference other than it be a decent linux distro
[14:17:01] <bobby44> :)
[14:20:19] <wakko> how does avr studio build system know if it has to use gcc or g++ ?
[15:00:27] <w|zzy> it's smart.
[15:00:46] <w|zzy> you select a c++ or c project.
[15:30:56] <Malinuss> heh
[15:42:38] <RikusW> http://www.hypirion.com/musings/without-conditionals
[16:16:42] <creep> h
[16:19:21] <OndraSter__> t
[18:47:26] <HeartOfAce> if I plug a piezo speaker to my MCU, do I need to plug a voltage amplifier? does the voltage amplifier increase the volume?
[18:48:12] <Casper> the higher the voltage, the higher the volume
[18:48:25] <Casper> HeartOfAce: do you know what is sound?
[18:49:11] <HeartOfAce> I see
[18:49:16] <HeartOfAce> vibrations?
[18:49:27] <HeartOfAce> at a certain frequency
[18:50:00] <Casper> :)
[18:50:06] <HeartOfAce> yeah, makes sense, thanks
[18:50:23] <Casper> a piezo is simply a material that disform itself depending on the voltage applied on it
[18:50:36] <Casper> the higher the voltage, the more disformed it become
[18:50:55] <Casper> the higher it disform, the higher the amplitude
[18:51:01] <Casper> exacly like a standard speaker
[18:51:30] <HeartOfAce> makes sense :D
[18:53:03] <Casper> now, going back to electonics...
[18:53:15] <Casper> there is 2 things to consider: voltage and current
[18:53:28] <Casper> the avr provide 5V and about 40mA max
[18:53:44] <Casper> will that be enought for your needs? that is what you'll have to see :D
[18:57:13] <HeartOfAce> I guess :P, thanks
[19:01:51] <creep> a piezo can tweet loud at 2mA current ;>
[19:02:10] <creep> best when driven at resonance frequency
[19:02:32] <Casper> 12V sine at resonant frequency... :D
[19:03:00] <HeartOfAce> can you guys hear about my idea, and suggest an AVR? I've done research, and I have some selected, but I'm very new to all of this :D
[19:03:11] <creep> hmm... waveform does not matter at resonance
[19:03:27] <creep> the piezo will resonate according to since function
[19:03:50] <creep> *sine
[19:04:17] <Casper> HeartOfAce: sure
[19:04:27] <creep> HeartOfAce<< 3-5V is enough for piezo drive... 12V will be louder
[19:05:11] <Honeyclaw> HeartOfAce: also, Atmel's site has a parametric part search feature to help you select the right AVR if you know how much memory or what peripherals you'll need
[19:05:29] <creep> HeartOfAce<< i prefer atmega48 for small stuff
[19:05:58] <creep> but you may want an attiny if you only need 1-2 ports
[19:06:14] <HeartOfAce> oh, it's pretty big project :(
[19:06:27] <HeartOfAce> it's for a group project for a microprocessor interfacing course
[19:06:43] <creep> hm... then atmega48 is 32 pin tqfp
[19:07:03] <Honeyclaw> atmega328 is nice because there is a lot of info on the web about it (same chip as some arduino boards)
[19:07:44] <HeartOfAce> well here's our project, it's a laser tag system (gun + vest), with bluetooth connected to an android phone.
[19:08:11] <Casper> but there is another thing that have to be considered: a well programmed code will work with minimal change on almost all avr that contain the function used (ex serial port)
[19:08:15] <creep> Honeyclaw<< it is the same as atmega48 just 32k flash right ?
[19:08:34] <Casper> so the code for one can be used for another with very minimal change
[19:08:42] <Casper> often a simple search and replace
[19:08:46] <HeartOfAce> the system has: the MCU, IR LED, IR Receiver, Speaker, Trigger Button, GPS, Bluetooth, Battery Charger/Gauge, and some LEDs
[19:09:45] <HeartOfAce> I think we need 2 UART, so that's why I picked the 324p, but I saw I can program that, so then the 328p would work fine
[19:25:35] <HeartOfAce> any thoughts :P?
[19:27:01] <creep> none
[19:27:51] <Horologium> atmega328 should work.
[19:28:38] <Casper> you can even have software uart in case you don't have enought hardware and have some ressources to spare, but better use hardware
[19:30:18] <creep> what do you think of this stuff ? :) http://www.bis0uhr.de/projekte/schwebekugel/english.php
[19:30:33] <creep> a magnet "levitator"
[19:32:02] <Horologium> it pulls the ball up till the hall sensor gives a certain reading then stabilizes it..
[19:32:06] <Horologium> what about it?
[19:32:11] <Horologium> it's been waaay done.
[19:32:13] <Horologium> lots.
[19:32:53] <creep> :)
[19:33:01] <creep> it is new to me
[19:33:38] <creep> i'd like a magnet float above a plate more
[19:33:58] <creep> but that is more complicated
[19:36:40] <creep> "http://www.ordersmax.com/p-6156880931-%E3%80%90%E3%80%91-10.html"
[19:37:01] <creep> seen a youtube video starting it, it has very little stability
[19:38:19] <creep> but sure is cool having some stuff floating over my desk
[19:44:18] <dioxide> they gain lots of stability if you spin them
[19:44:42] <HeartOfAce> does anyone know the difference between: https://www.sparkfun.com/products/10269 and https://www.sparkfun.com/products/10393 I think I'm retarded, cuz they look the same to me.
[19:46:10] <Horologium> creep, that schematic lends itself to adjustment...
[19:46:11] <creep> dioxide<< what is its secret ?:) do you have one?
[19:46:17] <dioxide> theyre supposed to look the same. the mate silver was designed to clone the bluesmirf, but be arduino friendly
[19:46:19] <Horologium> in fact, it has an adjustment.
[19:46:53] <HeartOfAce> ah, great, thanks
[19:47:05] <dioxide> creep: i dont know the technicalities, i just know magnetic levitation works better when its spinning
[19:47:20] <Horologium> spin it and get some nifty eddy currents in the metal.
[19:47:26] <creep> problem is preventing it from flipping over...
[19:47:31] <dioxide> you gain the gyroscope effect as well
[19:47:46] <dioxide> spinning prevents flipping
[19:47:58] <creep> so it drops if it stops ? ;/
[19:48:08] <dioxide> it flips if it stops, probably
[19:48:25] <creep> then it needs some development
[19:48:53] <dioxide> you can make the spinning continuous with a little ingenuity
[19:53:24] <Autorun> someone programming in BitCloud?
[19:57:07] <Horologium> never heard of it.
[20:13:21] <creep> it would be possible to capacitively levitate a sphere right ?
[20:13:37] <Casper> no
[20:13:54] <creep> capacitor charging creates repulsive force between plates
[20:14:06] <Casper> static electricity yes...
[20:14:25] <Casper> need the same polarity however
[20:14:29] <Casper> and will dissipate...
[20:14:44] <Casper> that's why they use magnets
[20:14:53] <creep> i mean periodically recharging the capacitor with ac current
[20:15:21] <creep> but for larger distances it would probably need high voltage
[20:17:01] <creep> and the base plate should be like a bowl
[20:18:55] <Horologium> look at superconductive magnetic levitation if you want to levitate something from the bottom up.
[20:19:32] <creep> i saw that but it needs a superconductor
[20:20:31] <creep> i don't store superconductors on my table top yet ;<
[20:20:47] <Horologium> so invent a room temp superconductor.
[20:20:52] <creep> looking forward to use some carbon nanotubes
[20:20:56] <Horologium> or cryogenic desk.
[20:44:58] <HeartOfAce> the PA atmegas are better than the P atmegas, right?
[20:45:10] <HeartOfAce> like, from what I found, they just figured out how to reduce the current consumption or something
[21:44:27] <Casper> HeartOfAce: "better" is always debatable
[21:45:31] <HeartOfAce> how so?
[21:45:36] <HeartOfAce> I just can't seem to find what it means
[21:50:41] <Casper> new newer version have some issues compared to the older one, but some issues of the old is gone
[21:50:48] <Casper> so it really depend on the design
[21:54:29] <HeartOfAce> ah
[21:54:36] <HeartOfAce> I thought they were functionally the same
[21:55:29] <Casper> yes
[21:55:31] <Casper> on paper
[21:55:46] <Casper> in practice, they react slightly differently
[21:57:03] <Casper> everything have bugs, including the hardware
[21:57:11] <creep> Pa should be newer
[21:57:27] <creep> lower voltage stuff
[21:57:42] <creep> lower voltage= lower power required
[21:57:57] <creep> well, smaller transistor are more sensitive
[21:58:11] <HeartOfAce> I see
[21:58:28] <creep> a 80486 or pentium would work more if they throw an A-bomb around you for example than a new core i7
[22:04:55] <hexapod> any sonsivri member here?
[22:06:02] <Casper> sonsiwhat?
[22:09:13] <hexapod> forget it
[22:09:40] <hexapod> any crack for eagle 6.4?
[22:10:40] <Tom_itx> not here
[22:11:15] <creep> haha
[22:13:40] <hexapod> I just got one right now