[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y ] [Search | Home]
4Archive logo
CPU Upgrades
If images are not shown try to refresh the page. If you like this website, please disable any AdBlock software!

You are currently reading a thread in /g/ - Technology

Thread replies: 63
Thread images: 7
File: L_00004550.jpg (54 KB, 524x488) Image search: [iqdb] [SauceNao] [Google]
L_00004550.jpg
54 KB, 524x488
Why is CPU development seemingly so slow? I've had an i7-2600 in my machine for 5 years now, and every time I think about upgrading I'm told that there's still nothing out there that's worth swapping my motherboard and RAM for. Meanwhile, a 5 year difference in GPU releases presents a massive performance difference.

What's the deal? I do a lot of simulation, data grinding, and video rendering on my PC and I'm looking for a significant performance boost. Is my 2600 really not holding me back? Is there anything currently worth upgrading to / waiting for this year?
>>
>>52400609

wait for zen
>>
>>52400712
so wait for amd's catch up to intel's haswell from 2013?

zen won't even be worth it.
>>
I've had an i7 [email protected] since 2008 and still see no reason to upgrade.

I might get a 775 6 core Xeon to pop in it at some point
>>
>>52400743
1366 i mean
>>
I have an i7 870 @ 4Ghz. Thinking about upgrading to an i5 4690k. Would it be worth it?
>>
>>52400609
What happened to 1155? Seems like they barely made any CPU's for it and gave up.
>>
>>52400609
If competition was present we would have 8 real core i7 in place of 4 core today.
>>
>>52400609
shitty devs have hard time implementing existing instructions and / or threading. and yes stick to your OC'd 2600 unless you're willing to go full retard with 2011-3 socket
>>
>>52400609
>Why is CPU development seemingly so slow?
by now any significant improvements require huge RnD investments
>>
>>52400759
unlikely
>>
>>52400759
why would you go to an i5? get 4790k if you're that desperate, at least it has more cache on top of everything else
>>
>>52400777
there is 10 core i7 today, only it costs $1500. I'm not even talking about 16 core server monsters
>>
Intel has become complacent. AMD needs another Sledgehammer to push them forward.
>>
You guys sure like giving Intel your mom's money
>>
moore's law

we are at the tail end of x86's improvement curve because the laws of physics are catching up. They can't make it any smaller or faster because they are running out of ways to do it.
>>
>>52400833
amds need to fucking die
ftfy
>>
>>52400759
no
>>
>>52400875
This guy knows what he's talking about

We need a drastically new architecture to really have any performance increase because we've stretched x86 as far as it can go. Also I could be wrong but I did read that we can barely shrink it anymore because the electricity jumps channel to channel as they are so close together
>>
>>52400875
SiGe transistors will go down to 7nm or so, there's a ways to go.
>>
>>52400967
You sound like a fucking marketing dev. Intel is sitting on oodles of patents and designs that will keep them afloat no matter what AMD throws at them.

Moore's Law is only "dead" if you read articles from moron editors. What you're seeing is an issue of resting on laurels, talent poaching and sabotage on one side and the other not doing the same quite as well as the other.
>>
>>52400907
That's literally the most retarded thing I have heard yet. Give me one good reason why they need to die. Plus it wouldnt happen either way cause competition is always needed otherwise Intel can just do jackshit and get away with it.
>>
>>52401049
They've been doing jackshit besides die shrinks since Nehalem
>>
>>52401121
GNAA RLY?
>>
File: 1437107326064.png (64 KB, 320x276) Image search: [iqdb] [SauceNao] [Google]
1437107326064.png
64 KB, 320x276
>Moore's Law no longer holds
>increases in processing speed cease
>programmers have to O P T I M I Z E again
>a new Golden Age
Reminder that shit like this can be done on an 8088 CGA
https://www.youtube.com/watch?v=yHXx3orN35Y
>>
>>52401121
Yea cause amd 's been sucking dick. If zen doesn't at least hold a bit of a fight they are fucked but with all the shit I've been hearing it should be a good year for amd. It's mostly rumors though so take it with a grain of salt.
>>
Look at this shit.
Almost half of it is a useless shitty embedded GPU.
It's not even an option to buy the cpu without it.
>>
i have a i7 3930k since 2 year and is thinking to upgrade is worth it and to which?
>>
>>52401191
No need for upgrade.
>>
I have a Q6600. Is it worth upgrading it?
>>
>>52401268
Depends what you wanna do with it.
>>
>>52400743
I upgraded from an i7 920 to a i5 6600. Blows the i7 away. Just upgrade. And get an M.2.
>>
File: 1268453211722.jpg (30 KB, 298x403) Image search: [iqdb] [SauceNao] [Google]
1268453211722.jpg
30 KB, 298x403
It's basically like this:

AMD is pure and utter shit.
AMD is Intel's only competitor.
Intel utilizes the fact that AMD is pure and utter cow dung.
Intel slows down development and sells marginal performance increases for over 10 times their value, slowly.
Intel is helped by the fact that it is based in Shitsrael, the nation of profiteering culture, and rip-off culture.


When people tell you to wait for something, dismiss all of it unless the rumor is related to graphene.
Graphene is the real CPU advancement we are all waiting for. CPU core numbers are just beating around the bush and bullshitting into a cesspool.
>>
>>52401322
>Intel slows down development and sells marginal performance increases for over 10 times their value, slowly.

Money is in mobile (i.e. laptops) and not really in desktops any more, so for the last 5 years Intel has been working on power efficiency.
>>
Yo, I have a 3570k it's still great. However my dims on my mobo are dying one by one. I'm slightly reluctant to buy a second hand mobo. What do?
>>
>>52401186
This desu senpai. Intel is too busy working on their useless embedded GPUs
>>
>>52400609
The performance increase trends definitely have decreased with each generation.

But they are significant enough to warrant an upgrade.

The thing that did change over the years is that more people became poor, so that would explain why others tell you not to upgrade.

But seriously though, the latest graphical cards require the latest cpu's because otherwise they would be bottlenecked.

The difference between your cpu and the latest one would probably be many, many frames for vidya.
>>
>>52401375
The format of devices is irrelevant to porting from silicone to graphene anon. I know you want a jerkfest about device format, but that other thread has pretty much crapped all over that argument.
>>
>>52401295
How so? What does it do better? All I do is code haskel and browse 4chan.
>>
>>52401417
The embedded GPUs are fast enough for 99.99% of all applications.
>>
>>52401029
Moore's Law IS technically dead if you can't keep doing die shrinks.

Optimization of processor design has nothing to do with Moore's Law.
>>
I buy AMD so Intel doesn't get too complacent. I don't understand why people circlejerk so much over Intel vs. AMD. People need to be buying both (or more, if there were any other competitors) so they don't get lazy and/or start monopolizing prices...
>>
>>52401268
Do you often encode video or extensively use encrypted filesystems?
>>
>>52402544

/g/ has no fucking clue how cpus work in general so anons just buy whatever falcon tells them to.
>>
>>52402599
I buy what /g/ tells me to
>>
>>52402612

Which is always Intel because /g/ is ignorant as fuck about cpus. Want vt-d? woops you got jewed! Conversely all vishera chips support AMD's version of it iirc.
>>
>>52400609
I regularly swap between a C2D P8600, C2Q Q6600, a Xeon E5507 and an i7 2620M and I notice no real practical difference between the three anymore even for VMs.
>>
>>52402650
>buying shitty consumer chips
You deserved to get jewed. If you really gave a shit about VMs you wouldn't be running them on your gaymer shitbox anyway.
>>
>>52401186
Just buy a more expensive LGA 2011 cpu. I'm pretty sure those don't gave any gpu on them.
>>
>>52403097
Why would he do that for a single-socket system when he can just buy a regular LGA 1151 Xeon that's probably even cheaper than its i7/i5 equivalent?
>>
>>52403143
No idea, I'm just saying words.
>>
>>52400609
Why spend money on R&D when AMD are so far behind?
>>
>>52400967
>electricity jumps from channel to channel
The fuck? That's not it at all, the problem is quantum tunneling which makes transistors unstable.
>>
>>52402512
Yes it does. Moore's Law is literally a statement of how the amount of transistors on a processor die duplicates every ~18 months. That has everything to do with optimization.
>>
>>52403221
>duplicates
I think you mean doubles.

Also, making the same thing smaller isn't really optimizing. Making the architecture more efficient so you don't need as many transistors to do the same thing is optimization.
>>
>>52400743
>4.4ghz
d0 or c0? I got my hands on a d0 a while back and found a nice and easy 4.2ghz stable (58% OC). At the time, people seemed to have trouble with their c0s below 4ghz, but found greater success above. These days you'd be hard pressed to find an ivy or haswell chip that can reach 50% safely, let alone within a reasonable temp range.
Nehalem was truly a greater time for cpus.
>>
>>52403279
Yes, that's what I meant, no idea where my brain was when I wrote that.

Depends on the type of optimization, finding a way of fitting more things into a confined space can be considered as optimization.
>>
File: De_Meiling.png (206 KB, 384x538) Image search: [iqdb] [SauceNao] [Google]
De_Meiling.png
206 KB, 384x538
>>52400609
because cpu manufacturers are lazy

top cpu stats 2010:
8 cores 4.0ghz

2016: 8 cores 4.5ghz

that means you got like maybe a 10% performance increase in the past like 8 years of cpu development
>>
>>52400743
Other than more power efficiency, newer instructions, more pci lanes, more io speed, ddr4, fixed bugs and better nsa backdoors in the microcode, yeah you have no reason to upgrade.
>>
>>52400742
This. Zen will suck.
>>
File: 1426347915033.gif (2 MB, 260x260) Image search: [iqdb] [SauceNao] [Google]
1426347915033.gif
2 MB, 260x260
>>52403451
>more pci lanes
Not true.

>ddr4
While improvement is nice, the average user won't notice the difference between ddr4 and ddr2vl 800mhz.

>fixed bugs
HAHAHAHAHA
>>
>>52403545
off yourself
>>
>>52404300
he's right tho
Thread replies: 63
Thread images: 7
Thread DB ID: 404696



[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the shown content originated from that site. This means that 4Archive shows their content, archived. If you need information for a Poster - contact them.
If a post contains personal/copyrighted/illegal content, then use the post's [Report] link! If a post is not removed within 24h contact me at [email protected] with the post's information.