TheHig
[H]ard|Gawd
- Joined
- Apr 9, 2016
- Messages
- 1,346
But I used the wink face.. Shit man! Lol.10% isn't great but at 10% per generation, we have double the performance in 7 generations...
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
But I used the wink face.. Shit man! Lol.10% isn't great but at 10% per generation, we have double the performance in 7 generations...
But I used the wink face.. Shit man! Lol.
That would certainly fit in with the lazy kind of IPC gains that intel has provided with each very minor architecture revision they put out with each "new" generation.
I don't quite get this attitude. We all know, from many manufacturers, that dropping node size is costly and difficult and with very little IPC change. If Intel decided to never release another chip until they could provide a very large increase you would all be bitching that they haven't released a new chip in ages. Because of this slowdown they have had to start adding new features surrounding power states, as an example.
Zen 14nm Now
Zen 14nm+ 2018
Zen2 7nm 2019-2020
Zen3 7nm+ 2020-2021
Dont expect much besides the process to be the driver. Zen2 is still 4 cores per CCX as well.
To be honest, clock speed is pretty much the only reason anyone still cares about Intels offerings at the moment. If an 8 core Ryzen or a 16 core Threadripper could get up to 4.6GHz without too much heat, most people wouldn’t give a crap about Intel anymore.
And THAT is the big issue, right? Their speed is low and they put out too much heat. Pretty much like every AMD chip, whether its a graphics card or a cpu for nearly a decade
as it should deliver higher clocks and lower heat
http://www.cpu-monkey.com/en/compare_cpu-intel_core_i7_8700k-763-vs-intel_core_i7_7700k-664
From this the result is:
Cinebench 11.5 ST
- Core i7-7700K: 2.36
- Core i7-8700K: 2.57 (+8.9%)
Cinebench 11.5 MT
- Core i7-7700K: 10.4
- Core i7-8700K 16.72 (+60.8%)
Cinebench R15 ST
- Core i7-7700K: 200
- Core i7-8700K: 218 (+9%)
Cinebench R15 MT
- Core i7-7700K: 995
- Core i7-8700K: 1523 (+53%)
Passmark CPU Mark
- Core i7-7700K: 13.098
- Core i7-8700K: 20.020 (+52.9%)
Geekbench 3 64-bit Single-Core
- Core i7-7700K: 5130
- Core i7-8700K: 5605 (+9.25%)
Geekbench 3 64-bit Multi-Core
- Core i7-7700K: 18.702
- Core i7-8700K: 28.581 (+52.8%)
http://www.cpu-monkey.com/en/compare_cpu-intel_core_i7_8700k-763-vs-intel_core_i7_7700k-664
From this the result is:
Cinebench 11.5 ST
- Core i7-7700K: 2.36
- Core i7-8700K: 2.57 (+8.9%)
Cinebench 11.5 MT
- Core i7-7700K: 10.4
- Core i7-8700K 16.72 (+60.8%)
Cinebench R15 ST
- Core i7-7700K: 200
- Core i7-8700K: 218 (+9%)
Cinebench R15 MT
- Core i7-7700K: 995
- Core i7-8700K: 1523 (+53%)
Passmark CPU Mark
- Core i7-7700K: 13.098
- Core i7-8700K: 20.020 (+52.9%)
Geekbench 3 64-bit Single-Core
- Core i7-7700K: 5130
- Core i7-8700K: 5605 (+9.25%)
Geekbench 3 64-bit Multi-Core
- Core i7-7700K: 18.702
- Core i7-8700K: 28.581 (+52.8%)
So
Some strange old benchmarks being used there... I wonder why? (Intel friendly cough)
But we are looking at 9% single threaded performance increase, due to higher clocks, and a 3% possible increase in MT scores? Not much to bother about, especially as we also know that it will be hotter and may not clock so well using all threads, versus a 4 core design. Also new motherboard and chipset needed, again. Price will be interesting.
Great now lets do those comparisons with maximum overclock and see where they fall. This is a K chip after all.
*crickets*
The fact is you have 60.8, 53, 52.9, 52.8 with multi core. This averages to 54.8% in highly threaded task. Well you have 50% more cores So 0.81% IPC improvement isn't that impressive. That's assuming same clocks. Sorry to disappoint. (154.8%/6 - 100%/4= .81% improvement)
Intel has been hitting the thermal headroom top for years. Unless they somehow changed the physics of trigrate not to exponentially consume power with voltage increases.
3% in MT? You mean 50-60%? Will it be hotter? Its a bigger die and the same TDP. Price is 350-370.
3% in MT? You mean 50-60%? Will it be hotter? Its a bigger die and the same TDP. Price is 350-370.
Was KL manufactured on 14nm+ and CL will be 14nm++?
Zen+ should be an interesting test that’s coming up quite soon, as it should deliver higher clocks and slightly lower heat
That is a hard one. Since 32nm shrinking the node tends to reduce the clock. Well at lest until the process matures.
And that's the problem. Same TDP limits it's overclockability with all those cores. Intel has had the same issue for years and why everyone is stating "Stick with 4 core" for games.
And that's the problem. Same TDP limits it's overclockability with all those cores. Intel has had the same issue for years and why everyone is stating "Stick with 4 core" for games.
I dont follow you. At 95W you get 6 cores at 4.3Ghz.
Overclocking have to be seen. But if its behind 7700K I doubt it will be by much, thanks to 14nm++. You get a 4.7Ghz all core OC out of the box guarantee due to the turbo testing.So are you saying that this chip will overclock with all 6 cores as well or better than a 4 core chip, within the same power and thermal envelope?
No change in IPC and yet we need a new chipset.
How are these 2 linked? What if CFL uses IMVP9 VRM spec and SKL/KBL uses IMVP8?
Most of us feel that Intel changes Chipsets just a little too often, and for very little reason.
4.3 GHz is the turbo clock with all 6 cores out of the box.So are you saying that this chip will overclock with all 6 cores as well or better than a 4 core chip, within the same power and thermal envelope?
So are you saying that this chip will overclock with all 6 cores as well or better than a 4 core chip, within the same power and thermal envelope?
Intel has always been claiming something like 7-11% performance improvement in IPC performance per generation since Sandy Bridge's introduction. The reality of the situation is that we've seen increases that are more like 3%. That 11% boost usually coincides with the use of a new AVX type instruction set and encoding performance. Any claims Intel has ever made that come out to 20% or thereabouts have always been about performance per watt or coincides with much higher factory clocks compared to the last generation. Intel has never claimed a 20% IPC improvement since Sandy Bridge was released.
I really, really hope that this is the case...the crazy high power draw and heat are the only things that would hold me back from getting a TR platform, if I were to be upgrading.
Sandy Bridge (32nm)
i7-2600K = 3.4/3.8 GHz (4c/8t)
i7-2700K = 3.5/3.9 GHz (4c/8t)
Ivy Bridge (22nm)
i7-3770K = 3.5/3.9 GHz (4c/8t)
Haswell (22nm)
i7-4770K = 3.5/3.9 GHz (4c/8t)
i7-4790K = 4.0/4.4 GHz (4c/8t)
Skylake (14nm)
i7-6700K = 4.0/4.2 GHz (4c/8t)
Kaby Lake (14nm)
i7-7700K = 4.2/4.5 GHz (4c/8t)
Coffee Lake (14nm)
i7-8700K = 3.7/4.7 (6c/12t)
The only decrease in base speed since 32nm is the upcoming Coffee Lake, but that can be extrapolated to it having 50% more cores.
What are you talking about?I got banned for inferring that he worked for Intel a couple of months ago. But he constantly states how good Intel are, and how shit AMD are, and he has a mod here as backup. The mod never even warned me, he just banned me.
Intel has always been claiming something like 7-11% performance improvement in IPC performance per generation since Sandy Bridge's introduction. The reality of the situation is that we've seen increases that are more like 3%. That 11% boost usually coincides with the use of a new AVX type instruction set and encoding performance. Any claims Intel has ever made that come out to 20% or thereabouts have always been about performance per watt or coincides with much higher factory clocks compared to the last generation. Intel has never claimed a 20% IPC improvement since Sandy Bridge was released.