Intel i5-7600K Kaby Lake Benchmarked - 10% Faster Than i5-6600K

>currytech
Just fuck off

>>57185149
>www.expreview.com = currytech

>>57185128
> 6-9% faster overall performance
> 8% faster clocks

nice IPC gains there, Intel

>>57185121
>increased clock

No shit. Goodluck with the Skylake OC edition

>>57185599
This. I hope Zen is decent so Intel gets some competition and starts releasing worth upgrading to CPUs again.

>>57185599
Intel wasn't promising increased IPC anyway, so. From what I recall they said they would do power usage optimization to make it take less power and that's it.

>>57189920
That's all they've been doing since Sandy Bridge. Sad.

>Higher clocks
>10% faster.

Wow, it's fucking nothing.
If you have bought a higher end CPU during the last few years, there won't be a reason for you to upgrade until the current CPU architecture and materials have run their course.
I bet the next worthwhile upgrade is going to be when they push photonic CPUs to the market or something.

File: dell_xps_13_kaby_lake_4k_10_bit_hevc_tears_of_steel_255_nits_battery_life-100685721-orig.png (13KB, 1491x796px) Image search: [Google]

13KB, 1491x796px

>>57189997
Improved battery life when playing videos.

>>57187717
>>57189938
>>57189900
people who want to upgrade their pcs often should just leave. You should be happy that desktops don't suddenly need a 20%+ gain in performance and that you can use your old cpu for recent programs. People should be more mad about often changes of sockets imo.
If intel released new chips with 5% gain but stayed on the same sockets so you don't have to suddenly sell everything out when you want a newer cpu it would be ok for me.

Guess those rumors about Intel having absolutely no fucking idea what they're doing is true.

>>57190013
Source on that graph? It simply looks like they introduced hardware decoding anyway, and nobody is going to buy Intel for their iGPU.

>>57190031
It's not that I want to upgrade, it's that I want them to keep making noticeable performance improvements like they've always been doing. I wouldn't care about socket changes if new architectures were faster, but they're not.

>>57190338
gaymers do not need performance improvement at this point to be honest. And offices/servers/workplaces need a better price/performance ratio at lower wattage because of power bills. So a 5% improvement is pretty good if you have reduced power usage, good enough that it will save you real money if you invest in 100+workstations with new architecture vs old at similar price.
I would love to see some good performance gains for games but for non-gaming desktops its really not required.

>>57190413
Well, that's just your opinion. Apart from games and servers, there's also encoding, compression, emulation... and performance improvements also imply that future low-end CPUs may be as powerful as the high-end ones from today, which massively improves bang for buck and makes cheap hardware give you a better experience.

>>57190299
Gamers won't, but if you want to build a small HTPC then it's more than adequate. As for the graph, not him but yeah there should be a major improvement. Skylake only has full hw decode for 8-bit HEVC. They didn't add support for 10-bit till Kabby.

>>57190489
>implying anyone does encoding/compression/emulation professionally on consumer grade products
as i said watts/performance is more important in this environments, since its lowering monthly costs.

>>57190031
I'd never use a newer CPU with an older chipset.

>>57190671
so if you had an old i3 and would want to upgrade to an new i5 you would also buy a new motherboard with the same specs just for fun? Ofc implying there are no major changes in other hardware.

>>57190299
That's pretty much what they did. The iGPU had 4k video playback. This is only really useful for laptops/tablets. So, it's still good for low end laptops or ultrabooks or office computers.

>>57190046
They know it is time to move on to new materials but they are milking their fabs.

>>57190720
The chipsets always have added features for new revisions.

Intel is shit now, its time AMD to shine

What did you expect? 50% performance gain?

>>57185121
10% clock improvement
10% performance gain

Intel IPC gains RIP

>>57185121
>currytech

>>57185128
>8% higher stock clock
>results in ~8% "faster" performance

Who else got memed? Literally just a fucking clockspeed increase.

>>57190031
>eople should be more mad about often changes of sockets imo.
>>57190031
>people who want to upgrade their pcs often should just leave
If you're upgrading your PC enough that socket changes affect you, congratulations, you should leave this thread by your own admission.

Also, kiddies with babby's first self-built PC tend to think they're going to keep parts 5-ever and upgrade for 20 years.

The majority of us that have been doing this for a long time have realized that just isn't going to happen. No one needs that full tower case, very few ever actually use that second PCIe 16x slot after they buy a single card (by the time they want an upgrade, buying an old ass card for SLI/CF to equal the performance of a mid-range brand new card feelsbadman), and few people actually upgrade their CPUs more than one generation, if at all.

I mean really, let's say you had a Sandy Bridge system way back in the beginning of 2011, you really want to keep that shit with Skylake/Kaby Lake in 2016/2017? lol No DDR4, poorer USB controller (and possibly not even a 3.0 controller if you got it early enough), no m.2 sockets, PCIe 2.0 only, etc.

A mobo is just a single component, and often isn't even the most expensive component, of a system. Why would you get so attached to a single component of a system yet so enthusiastic to replace the other (sometimes more expensive) parts?

>>57190013
Damn that's amazing battery life 10 hours screen on time wtf.

File: 1474496560896.jpg (85KB, 633x738px) Image search: [Google]

85KB, 633x738px

>>57185121
>>57185128
>7600k
>3.8-4.2GHz
>6600k
>3.5-3.9GHz

so basically they increased clock and called it an IPC gain?

>>57190013
>battery
>using laptops

basically a normie cpu

File: 1446270407379.gif (927KB, 250x230px) Image search: [Google]

927KB, 250x230px

>>57190013
>(Longer bars indicate better performance)

>>57191239

>No DDR4, poorer USB controller (and possibly not even a 3.0 controller if you got it early enough), no m.2 sockets, PCIe 2.0 only, etc.


Most of these things have no impact on practically anythings. Getting better USB speeds is probably the only thing that really has a tangible effect and only if your decides are USB 3.0 compatible.

But more to your point, I think people are simply reluctant to change their mobo simply because it's just really annoying thing to actually do. As in it's literally a hassle to remove all the fucking cable, unscrew the thing, handle it and so on.

File: Intel-Coffee-Lake-Roadmap.jpg (58KB, 635x255px) Image search: [Google]

58KB, 635x255px

>>57189920
>Intel wasn't promising increased IPC anyway, so. From what I recall they said they would do power usage optimization
>>57189938
>That's all they've been doing since Sandy Bridge. Sad.

Haswell was actually respectable in retrospect at least.
It's just Broadwell was a lackluster shrink and all the Lake series are garbage.

> Skylake - sub-5% IPC gains for non-AVX/accelerated stuff
> Kaby Lake - virtually 0% IPC gains
> Coffee Lake - hey, at least you'll get 6 cores finally, right?
> Cannonlake - glorious 10nm coming to your desktop in 2019 now!

>>57190013
I did the correct thing waiting to get a new laptop. I pretty much live in hotel rooms, airports, planes, trains and the backseat of a taxi nowadays. A light ultra portable laptop with great battery life will be awesome for entertainment and coding on the go.