Overclock.net - An Overclocking Community

Overclock.net - An Overclocking Community (https://www.overclock.net/forum/)
-   Intel CPUs (https://www.overclock.net/forum/5-intel-cpus/)
-   -   Skylake i7 6700K lose to Haswell i7 4790K in gaming? (https://www.overclock.net/forum/5-intel-cpus/1568288-skylake-i7-6700k-lose-haswell-i7-4790k-gaming.html)

trodas 08-06-2015 01:48 AM

According to the preliminary (NDA) and pulled down tests, Skylake is good for syntetic benches (109% over i7 4790K):
Skylake_wins_syntetich_benches.jpg

...but it sadly and surprisingly lose (4 tp 6%) in gaming tests to i7 4790K:
Skylake_gaming_failure.jpg

Sure, Anandtech is trying to cover this debacle, using an i7 4770K to unfair comparsion and ComputerBase tweaked the results to make Skylake look good (1% faster that i7 4790K) by throving syntetic benches in the mix, but gaming tests show that i7 4790K wins.

Only in Total War: Attila Skylake with GTX 980 Ti combination wins slightly faster result.

Unlocked_Skylake.jpg

At least the TDP get from 95W lovered to 91W (only marketing change?). OTOH the asking price will be 350$ for i7 6700K and i5 6600K will be for 243$.

http://wccftech.com/intel-unleashes-14nm-skylake-k-core-i7-6700k-core-i5-6600k-processors-91w-z170-platform-support/
http://wccftech.com/intel-skylake-core-i7-6700k-cpu-review-processor-gaming-performance/
http://www.computerbase.de/2015-08/intel-core-i5-6600k-i7-6700k-test-benchmark-skylake/

Sp33d Junki3 08-06-2015 01:51 AM

Who cares if it does not beat the 4790k.
It is better too see how it really does against Ivy and Sandy. Those are better to see if an upgrade is better.
Anything before Sandy would be a an upgrade.

flopper 08-06-2015 01:51 AM

5820k and x99 the way to go for upgrades.
wont do 5ghz with skylake, fail IMO.

Cyro999 08-06-2015 02:04 AM

rMqEcuB.png
kkpgKcE.png
1WU5J9Q.png
m5H3A0t.png
ke8wstf.png
JjFSUrt.png

Your site is testing incorrectly.

Finally, x264 encoding performance where you see 4.5ghz i5 6600k match a 9590 @5ghz
vM14zRW.png
^This is often tested wrongly, because reviewers use a 5 year old encoder version with a popular benchmark that is terrible and doesn't show performance improvements that newer CPU's have from avx2 etc. This test is accurate.

and IPC gains of 11.9% over Haswell, 37.7% over sandy bridge

can we stop pretending that skylake is a disappointment? It's exactly what i expected plus a little more (was erring on the cynical side). There's no need for this drama on every new CPU release when tock to tock it's not all -that- bad, especially considering that you'll probably get ~5% more from clocking 200mhz higher than Haswell (especially non-DC)

Sure, it should be 6 core. It could be faster - but it's here and it's NOT slower than a 4790k.

Imouto 08-06-2015 02:53 AM

Quote:
Originally Posted by Cyro999 View Post

Your site is testing incorrectly.

Yeah, because clocking a Sandy Bridge lower than anything else is doing things right.

I agree that Skylake is strictly a common tock, neither good or bad but that polish site is a joke.

Cyro999 08-06-2015 02:55 AM

Quote:
Yeah, because clocking a Sandy Bridge lower than anything else is doing things right.

They have several pages of tests with SB/HW/SL at the same clocks. In that case i'm not sure why no 4.7ghz sandy bridge on those gaming bench pics but i never even looked at that because 4.7ghz Haswell and Skylake are both there.

A joke why? It's the best skylake review out there. It's like 30 pages of stuff that's way more accurate than what most of the reviewers are posting. Actual non-GPU-bound tests, straight clock for clock IPC comparisons at 3ghz and 4.5ghz, etc.

Anateus 08-06-2015 03:08 AM

Those tests are strange. Stock skylake witcher 3 is 113fps, while ovetclocked to 4,7 is... 114fps.

Cyro999 08-06-2015 03:16 AM

Quote:
Those tests are strange. Stock skylake witcher 3 is 113fps, while ovetclocked to 4,7 is... 114fps.

Because it's GPU bound so useless as a test between a 4790k and 6700k

Kutalion 08-06-2015 03:26 AM

Those polish guys are famous for the crap inaccurate info they toss out.

Anateus 08-06-2015 03:27 AM

Quote:
Originally Posted by Cyro999 View Post

Because it's GPU bound so useless as a test between a 4790k and 6700k

Then why Crysis benches show the same? Welcome to the jungle is extremely CPU bound, and yet its just 1fps gain (which is in the margin of error).


All times are GMT -7. The time now is 03:25 AM.

Powered by vBulletin® Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.

User Alert System provided by Advanced User Tagging (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
vBulletin Security provided by vBSecurity (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.