so /g/, do you think computer performance is slowly grinding to a halt?
>>52351960
Not enough competition. Wait until Zen is released to see a surge. Even if Zen isn't good, Intel is going to want to have new CPU's.
>>52351960
Why would you think that? Every year the new shit becomes obsolete, if anything we're going 2fast
>>52352006
I'm just talking about how moore's law can only allow so much compactness in components to allow for better performance that the wall will finally be hit. Manufacturers will have to deliberately slow down performance progress while waiting for something better to replace the current transistor technology.
>>52352006
Well, the computer I build in 2009 can still run every newly released videogame on the highest settings
Ergonomics have continued to improve
Affordable IPS monitors
Longer battery life
and SSDs are still getting cheaper/bigger year after year
Memory bandwidth still rising
The most important issue, is still largely unsolved though: Free as in Freedom firmware
There are only 2 motherboards support reasonably powerful CPUs (Opteron)
>>52352109
I'm using a laptop from 2007 right now that performs about as well as a smartphone from 2 years ago. Can't say the same if this was in the 90s, where your Commodore 64 felt ancient compared to your 386 PC even though they were no more than 4 years apart.
>>52352168
I meant *80's, sorry for the typo
>>52352109
lmfao not at 60fps 1080p
>>52351960
for consumers, yeah, because the more powerful they get the more retarded software developers become
>>52352245
Which explains the ludicrous amount of RAM web browsers use
>>52352274
yeah although id blame that more on the way websites are developed
you got all these frameworks to make programs out of documents and let stupid designers realize their ideas on their own.
it's like architects pretending to be carpenters and making some insane monstrosity that only looks good from the right angle
>>52352507
every designer cuck should read it at least once before they start to work on a project
http://motherfuckingwebsite.com/
>>52354350
I make my sites look just like modern websites but with simple HTML and CSS. Runs like butter on anything, keeps my customers happy and autists that will never visit the sites happy as well.
It surprises me that shit like facebook and tumblr load 8MB of bullshit scripts for a simple static page.
Computers are fast as blistering hell. Software has gotten lazy due to lack of restrictions in system resources. Blame that.
>>52351960
Depends.
Consumer performance has gone downhill.
But server architectures and CPU's have been growing at relatively steady pace.
For example when nehalem architecture was released backed in '08 you could get up to 16 DDR3 Dim slots you could get like 128 gb of RAM (Correct me if i'm not mistaken) and now with teh advent of DDR4 and haswell and other architectures you can basicly get 512 gb of RAM per CPU.
It's a steady growth just on RAM capacity alone not speaking of PCIE and peripherals.
>>52354482
they probably track everything possible to track for marketing research, like where people move their mouse to get an idea if they were more interested in one ad than another. or what they should post more of in their feed to make sure they see more advertisements.
facebook is so big i doubt their website is slow out of pure laziness, there must be something they are actually doing with the resources besides wasting them on inefficient algorithms.
>>52351960
for most everyday jobs it has been since core 2
>>52352006
*Maybe* in gaymes, but for most home use cases you can still use 10 year old shit just fine, fuck I'm shitposting from a Q6600 right now that feels no different from my sandy bridge or nehalem systems in terms of practical performance.
>>52352168
Shit was different back then, the Commodore 64 was obsolete before the first one rolled off the assembly line, likewise with early PCs until the 386.
>>52351960
No, GPUs have been consistently increasing in raw compute power, the problems are in the CPU space where with x86/ARM, chip designers are hitting a brick wall when it comes to perf/watt since the prediction/reorder logic in the modern out-of-order superscalars is literally consuming 90% of the die space and power and they've almost sneezed every last drop of improvement from them. The only real solution is to adopt a different approach to CPU architectures that lends itself better to the increasing demand in performance and power efficiency, the Mill comes to mind as a possible alternative
>>52352002
this is what retards say when forced to accept the laws of physics
>>52351960
That picture is so indie
>>52355173
>Mill
>ever seeing consumer adoption
I really don't see that happening.
>>52355519
intel has enough money to pay congress to change those laws
