|
Win8 Hetro Experie posted:malloc-like memory, what
|
#
?
Mar 13, 2013 15:11
|
|
|
|
| # ? Jun 11, 2024 18:30 |
|
|
maybe they mean an mmu or somethin?? also tef im just over in east bay if you want to have lunch, tho ive only been SAing regularly for like the last couple months (maybe thats a good thing)
|
|
#
?
Mar 13, 2013 15:17
|
|
|
iirc hennessy and patterson 4th ed had a bunch of stuff about the cell edit whoops wrong one http://www.amazon.com/Computer-Orga...tion+and+design i think GameCube fucked around with this message at 15:42 on Mar 13, 2013 |
|
#
?
Mar 13, 2013 15:35
|
|
|
Cybernetic Vermin posted:lol at people going on about computers of the future being all gpu and 128 stupid arm cores, intel will laugh all the way to the bank another decade at this rate Tilera 
|
|
#
?
Mar 13, 2013 16:06
|
|
|
Cybernetic Vermin posted:tech geeks like to imagine that it is worth it to build a restrictive and inconvenient programming model if it yields high performance in theory, but it never actually works out because the hours of competent programmer time you have always bottlenecks the whole thing anyway. Lol no. Programmer incompetence will always expand to fill the available clock cycles. It has nothing to do with whether the model is restrictive and inconvenient. We could have 100GHz single core machines and software would just be 50 times slower. I mean, look at how the web reacts to more people getting broadband, they just start shoving MBs of JS down to the client. CPU development should just stop, it gets us nowhere.
|
|
#
?
Mar 13, 2013 17:37
|
|
|
that is simultaneously stupid and correct, one of the main payoffs with increases in performance is increases in productivity, which is an advantage that restrictive models with high performance not only lack but invert
|
|
#
?
Mar 13, 2013 17:59
|
|
|
yeah web dev sure has been a big payoff for users
|
|
#
?
Mar 13, 2013 18:10
|
|
|
Cybernetic Vermin posted:that is simultaneously stupid and correct, one of the main payoffs with increases in performance is increases in productivity, which is an advantage that restrictive models with high performance not only lack but invert Yeah, we can spend our time productively learning JQuery and CoffeScript instead of coding.
|
|
#
?
Mar 13, 2013 18:14
|
|
|
jquery owns tho?
|
|
#
?
Mar 13, 2013 18:19
|
|
|
Zombywuf posted:Yeah, we can spend our time productively learning JQuery and CoffeScript instead of coding. well not everyone is an asperger math wizard, some people just want to make a fukin website, and jquery does a p good job of helping that happen
|
|
#
?
Mar 13, 2013 19:51
|
|
|
polpotpi posted:jquery owns tho? edit: i don't do webdev. i program computers.
|
|
#
?
Mar 13, 2013 19:53
|
|
|
Zombywuf posted:Yeah, we can spend our time productively learning JQuery and CoffeScript instead of coding. god forbid we have to learn soething
|
|
#
?
Mar 13, 2013 20:06
|
|
|
you know, continuous blocks of raw memory and pointers to read and write into it instead of all sorts of RGBA textures that you sample at linearly interpolated floating point texture coordinates
|
|
#
?
Mar 13, 2013 20:19
|
|
|
so a physical scheme that allows for programming languages to use a flat memory model. yeah, that's very nice to have.
|
|
#
?
Mar 13, 2013 21:08
|
|
|
Anybody signing up for Google code jam? I did, but I'm not really a programmer. I only write some code once in a while so if I get past the qualification round it'll be a small miracle.
|
|
#
?
Mar 13, 2013 21:48
|
|
|
Otto Skorzeny posted:so a physical scheme that allows for programming languages to use a flat memory model. yeah, that's very nice to have. when you're trying to mold your particular problem into the shape of a vertex shader function feeding data into a geometry shader function feeding data back into another vertex shader function which feeds data into a fragment shader function which finally writes the data as colors to a texture and that's just the first pass and it's not even given that you can read from a texture inside your vextex shader functions, I can sympathize if someone would rather just write the whole thing in C maybe OpenCL is the answer
|
|
#
?
Mar 13, 2013 22:01
|
|
|
Win8 Hetro Experie posted:when you're trying to mold your particular problem into the shape of a vertex shader function feeding data into a geometry shader function feeding data back into another vertex shader function which feeds data into a fragment shader function which finally writes the data as colors to a texture and that's just the first pass and it's not even given that you can read from a texture inside your vextex shader functions, I can sympathize if someone would rather just write the whole thing in C I used opencl once, it was pretty decent but a lot of boilerplate
|
|
#
?
Mar 13, 2013 22:23
|
|
|
Win8 Hetro Experie posted:when you're trying to mold your particular problem into the shape of a vertex shader function feeding data into a geometry shader function feeding data back into another vertex shader function which feeds data into a fragment shader function which finally writes the data as colors to a texture and that's just the first pass and it's not even given that you can read from a texture inside your vextex shader functions, I can sympathize if someone would rather just write the whole thing in C Isn't this the problem that Renderscript was supposed to solve? I mean, Android, but still.
|
|
#
?
Mar 13, 2013 22:25
|
|
|
i was agreeing with you
|
|
#
?
Mar 13, 2013 22:35
|
|
|
Zombywuf posted:Lol no. Programmer incompetence will always expand to fill the available clock cycles. It has nothing to do with whether the model is restrictive and inconvenient. We could have 100GHz single core machines and software would just be 50 times slower. I mean, look at how the web reacts to more people getting broadband, they just start shoving MBs of JS down to the client. CPU development should just stop, it gets us nowhere. you have identified the correlation and hosed the causation: we buy clock cycles specifically for programmers to fill them we buy bigger, faster computers so we can spend less money on jackasses typing away at keyboards and jabbering about "object oriented message passing scalability beep boop"
|
|
#
?
Mar 13, 2013 22:57
|
|
|
Reach true web scale today by running your application on Node.js and MongoDB.
|
|
#
?
Mar 13, 2013 23:04
|
|
|
gucci void main posted:Reach true web scale today by running your application on Node.js and MongoDB. its web scale.
|
|
#
?
Mar 13, 2013 23:05
|
|
|
Otto Skorzeny posted:i was agreeing with you oh. well I generally hate to fall on the less_restrictions && more_general == better_than side so in this case i'm half disagreeing with myself purely on principle
|
|
#
?
Mar 13, 2013 23:06
|
|
|
Win8 Hetro Experie posted:when you're trying to mold your particular problem into the shape of a vertex shader function feeding data into a geometry shader function feeding data back into another vertex shader function which feeds data into a fragment shader function which finally writes the data as colors to a texture and that's just the first pass and it's not even given that you can read from a texture inside your vextex shader functions, I can sympathize if someone would rather just write the whole thing in C one day, there will be people using an abstraction layer on top of opencl to write graphics code. they will drop down openCL when they want to get "closer to the metal". it is coming.
|
|
#
?
Mar 13, 2013 23:48
|
|
|
so long as man works in abstractions he will always be a slave to gods
|
|
#
?
Mar 13, 2013 23:56
|
|
|
PleasingFungus posted:one day, there will be people using an abstraction layer on top of opencl to write graphics code. they will drop down openCL when they want to get "closer to the metal". how soon
|
|
#
?
Mar 13, 2013 23:59
|
|
|
idris is cool
|
|
#
?
Mar 14, 2013 00:19
|
|
|
Messyass posted:Anybody signing up for Google code jam? I did, but I'm not really a programmer. I only write some code once in a while so if I get past the qualification round it'll be a small miracle. i got a t-shirt last year! thanks for reminding me to sign up
|
|
#
?
Mar 14, 2013 00:46
|
|
|
gucci void main posted:Reach true web scale today by running your application on Node.js and MongoDB. it works for trello!
|
|
#
?
Mar 14, 2013 02:10
|
|
|
I was looking at a book about "Web Apps" today because I'm making a database skin at work, and I looked at the ToC and saw both node.js and MongoDB called out by name, and then I looked at the author bio on the back and it talked about how duder had written a js library and spoke at RoR conferences and I safely put the book back on the shelf and exited the application. Why do people think that node.js is in any way desirable? Why in hell would anyone want to write server-side stuff in javascript, on V8 or otherwise?
|
|
#
?
Mar 14, 2013 03:31
|
|
|
cause web "developers"
|
|
#
?
Mar 14, 2013 03:31
|
|
|
quote:Node.js is currently used by a number of large companies including LinkedIn,[8][9] Microsoft,[10] Yahoo![11] and Walmart.[12] never mind, needs suited
|
|
#
?
Mar 14, 2013 03:32
|
|
|
Messyass posted:Anybody signing up for Google code jam? I'd rather die
|
|
#
?
Mar 14, 2013 03:32
|
|
|
Bream posted:I was looking at a book about "Web Apps" today because I'm making a database skin at work, and I looked at the ToC and saw both node.js and MongoDB called out by name, and then I looked at the author bio on the back and it talked about how duder had written a js library and spoke at RoR conferences and I safely put the book back on the shelf and exited the application. so I can not have to shift languages constantly when maintaining both ends of my application. same reason google made gwt or whatever that thing was that compiled java to javascript
|
|
#
?
Mar 14, 2013 03:34
|
|
|
Bream posted:I was looking at a book about "Web Apps" today because I'm making a database skin at work, and I looked at the ToC and saw both node.js and MongoDB called out by name, and then I looked at the author bio on the back and it talked about how duder had written a js library and spoke at RoR conferences and I safely put the book back on the shelf and exited the application. idk I've spoken at ruby conferences (about compilers and data structures) and node,js is reprehensible and mongo is trash for idiot retard babyes
|
|
#
?
Mar 14, 2013 03:34
|
|
|
my team just took ownership of a node app and once you scrape away the bad decisions the original author made its not bad. I'd never use it for actual websites, but for little applications its fine.
|
|
#
?
Mar 14, 2013 03:37
|
|
|
Cocoa Crispies posted:idk I've spoken at ruby conferences (about compilers and data structures) and node,js is reprehensible and mongo is trash for idiot retard babyes yeah but you've spoken at ruby conferences so your opinion is suspect
|
|
#
?
Mar 14, 2013 03:38
|
|
|
rotor posted:so I can not have to shift languages constantly when maintaining both ends of my application. same reason google made gwt or whatever that thing was that compiled java to javascript See now I think the stuff that compiles to JS is actually kinda cool, like emscripten. I also saw a book on coffeescript, but I decided I didn't feel like learning another language to solve a silly problem when it would be better for work if I could just do it faster. The coffeescript book had some serious webdev vibes to it as well, would that assessment be correct? I just wish we had widespread nacl already. 
|
|
#
?
Mar 14, 2013 03:40
|
|
|
Bream posted:See now I think the stuff that compiles to JS is actually kinda cool, like emscripten. I also saw a book on coffeescript, but I decided I didn't feel like learning another language to solve a silly problem when it would be better for work if I could just do it faster. The coffeescript book had some serious webdev vibes to it as well, would that assessment be correct? coffeescript is gay as heck. nacl seems neat, but also scary. if you want to run native apps in a browser, we already have an <object> tag, I suggest you use it.
|
|
#
?
Mar 14, 2013 03:43
|
|
|
|
| # ? Jun 11, 2024 18:30 |
|
|
note that 'gay as hell' is used here in the perjorative sense, not in the cool gay dude sense
|
|
#
?
Mar 14, 2013 03:45
|
|