How does 4.8ghz 4790k compare to avg overclocked 6700k?

TechyGeek

Member
Feb 23, 2015
108
9
81
Hi Everyone,

I just want to make sure I don't have a buyer's remorse. I have 4.8ghz 4790k which I consider avg overclock. I want to know how it compares with newer avg overclocked skylakes?

Also,

How far are we from having $300-400 video cards that can do 4k gaming with 60fps? Will they be bottlenecked by my ASUS z97? I don't mind getting 2 if the cost is within my budget.

Thanks!
 

UsandThem

Elite Member
May 4, 2000
16,068
7,380
146
Keep the 4790k. You will not gain anything.

This bench shows a 4790k running at stock vs 6700k overclocked to 4.4 Ghz:

http://www.anandtech.com/bench/product/1260?vs=1548

As far as a card that can do 60 FPS on all games in 4k, your only choice right now is a GTX 1080. Even then, it can't do 60 FPS on all games at max settings.

You will not be bottlenecked by your CPU or motherboard.

Wait until tomorrow and read the reviews on the GTX 1070. Maybe in SLI, they can provide the FPS in 4k you are looking for. If not, spend the $600-$700 for a GTX 1080 and turn down a few settings in games that aren't getting 60 FPS.
 

UsandThem

Elite Member
May 4, 2000
16,068
7,380
146
Anand cpu(and memory) in games benchmarks are worthless since they are done in mostly gpu bottlenecked test places.


But yeah no point changing this cpu - the only real benefit of Skylake+z170 would be additional pci lances connected to m2 port

Pick whichever website benchmarks you believe, and they all show the same results.

I'm always amazed that people who don't like Anandtech reviews or data become members on its forums.

Of course most quick reference benchmarks for CPUs are GPU bottlenecked. People outside of enthusiasts want to see if they have X, and change it to Y, what Z will be.

Their full reviews show the full range of performance from IGP to high-end cards. However, most people running a 6700k aren't gaming on integrated graphics.

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation
 
Aug 11, 2008
10,451
642
126
When I joined the forums, Anand was still here and the reviews were much more timely, well designed, and complete. This site was my go to site for benchmarks. Now, however, I dont even go to the home page of the site, just the forums. A lot of good information can be obtained from the forums, taking everything with a grain of salt of course. Anand's gaming benchmarks, especially cpu, are very weirdly designed and seem to be all over the place.

In regards to the original post, Skylake with fast ram can be maybe 10% faster than Haswell overall, even more in a few corner case games.
 

Madpacket

Platinum Member
Nov 15, 2005
2,068
326
126
When I joined the forums, Anand was still here and the reviews were much more timely, well designed, and complete. This site was my go to site for benchmarks. Now, however, I dont even go to the home page of the site, just the forums. A lot of good information can be obtained from the forums, taking everything with a grain of salt of course. Anand's gaming benchmarks, especially cpu, are very weirdly designed and seem to be all over the place.

In regards to the original post, Skylake with fast ram can be maybe 10% faster than Haswell overall, even more in a few corner case games.

+1

Didn't Anandtech compare Skylake to Haswell using DDR4 2100Mhz memory? This memory speed can really limit Skylake which shows benefits all the way up to 3200Mhz, at least in many games.

I visit the homepage occasionally only to be generally disappointed by the quality of the articles. Techreport has also gone downhill since Scott left.


Oh and OP, if you're running a stable 4.8Ghz there's really no reason to switch to Skylake for improved gaming performance. That's an excellent overclock, I don't think Skylake
overclocks that well on average.
 
Last edited:

UsandThem

Elite Member
May 4, 2000
16,068
7,380
146
+1

Didn't Anandtech compare Skylake to Haswell using DDR4 2100Mhz memory? This memory speed can really limit Skylake which shows benefits all the way up to 3200Mhz, at least in many games.

I don't think it really would have mattered if they had tested it with faster ram. There are several users here who believe the review they did on DDR4 scaling on Haswell-E was inaccurate and/or misleading, and the review must use 'frame time variance' to truly show the difference. They would have used the same testing methods they always do, and the results wouldn't have been believed anyways.
 
Aug 11, 2008
10,451
642
126
I don't think it really would have mattered if they had tested it with faster ram. There are several users here who believe the review they did on DDR4 scaling on Haswell-E was inaccurate and/or misleading, and the review must use 'frame time variance' to truly show the difference. They would have used the same testing methods they always do, and the results wouldn't have been believed anyways.

What does haswell E have to do with skylake memory scaling. Several tests have shown significant performance increases with faster memory for skylake.
 

UsandThem

Elite Member
May 4, 2000
16,068
7,380
146
What does haswell E have to do with skylake memory scaling. Several tests have shown significant performance increases with faster memory for skylake.

I don't think it really would have mattered if they had tested it with faster ram. There are several users here who believe the review they did on DDR4 scaling on Haswell-E was inaccurate and/or misleading, and the review must use 'frame time variance' to truly show the difference. They would have used the same testing methods they always do, and the results wouldn't have been believed anyways.

Please re-read the post and the answer is there.

And it depends on what tests you are referencing, because there are lot of tests that show it doesn't. It has already been argued ad nauseam in the Memory and Storage sub-forum.

Anyways, I did not come in this post to argue DDR4 scaling. I was just responding to post #4 where the user said Anandtech CPU and memory benchmarks are worthless, and if they ran the same battery of tests on z170 motherboards, some users would not accept the results because they claim their testing methods are flawed.
 
Last edited:

bonehead123

Senior member
Nov 6, 2013
559
19
81
Well, in my case, I can say without a doubt that my skylake rig is considerably faster in every way than my last Z97/4790k@4.8ghz rig was.

Personally, I attribute this to

A) The switch to pcie/nvme m2 drives vs. sata-III
B) 3400 ram vs 2133
C) The change to PCIe 3.0/DMI 3, and the corresponding increase in lanes available to the cpu.

One caveat however, is that I do not game, so I can't speak to whatever improvements, if any, these changes make in that regard, but still, this is the fastest, most responsive system I have owned since I started using computers in 1988 :)
 

StrangerGuy

Diamond Member
May 9, 2004
8,443
124
106
Well, in my case, I can say without a doubt that my skylake rig is considerably faster in every way than my last Z97/4790k@4.8ghz rig was.

Personally, I attribute this to

A) The switch to pcie/nvme m2 drives vs. sata-III
B) 3400 ram vs 2133
C) The change to PCIe 3.0/DMI 3, and the corresponding increase in lanes available to the cpu.

One caveat however, is that I do not game, so I can't speak to whatever improvements, if any, these changes make in that regard, but still, this is the fastest, most responsive system I have owned since I started using computers in 1988 :)

Or D) Placebo. :rolleyes:
 

ClockHound

Golden Member
Nov 27, 2007
1,108
214
106
Well...if Intel can keep making faster placebos every generation, it will save on expensive IPC and process R&D.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,218
3,797
75
Well...if Intel can keep making faster placebos every generation, it will save on expensive IPC and process R&D.

That would be why they keep doing .1ghz "refreshes". And now Kaby Lake.
 

SteveGrabowski

Diamond Member
Oct 20, 2014
6,794
5,752
136
What does haswell E have to do with skylake memory scaling. Several tests have shown significant performance increases with faster memory for skylake.

I have actually seen increased performance from faster RAM in Haswell too. My system is

CPU: Xeon E3-1231v3
BOARD : MSI Z97S SLI Krait Edition
GPU: EVGA GTX 970 SC

and in GTA V I could force a drop to about 52 fps in this one area near Mt Chiliad with lots of shadows when using my DDR3-1600 RAM. After upgrading to DDR3-2400 on the exact same version of the game my framerate would only drop to 57-58 fps. This is at 1080p mostly ultra settings (grass moved down to very high). It was a 100% reproducible result and the only difference was the RAM. I also noticed a similar improvement in Fallout 4. On ultra settings except for god rays and shadow distance being both set to medium I would always get a drop to ~51fps or so with DDR3-1600 inside the speakeasy in Concord at this one spot where light comes in through a window and illuminates a bunch of particles and crap floating around in the air. With DDR3-2400 it didn't drop below 60 fps. This was back before the 1.3 patch came out that added a ton more graphical options like HBAO+ though.

I bring up these two games because I specifically remember exact places to trigger these minimum framerates and I know what the minimums were with the DDR3-1600 kit. The fast RAM was no help whatsoever in Crysis 3 though even though that's a game that hammers cpus when run at Very High settings. Overall I'd say it made no difference in most games I ran on both setups, since most were 100% gpu bound with a 970 at 1080p. But the higher minimum framerates in those two games were very noticeable to me. If I get an occasional drop to 57 fps or so I don't usually notice it unless I have a frame counter on, but a drop to 52 fps is immediately detectable to me.
 

Zodiark1593

Platinum Member
Oct 21, 2012
2,230
4
81
I have actually seen increased performance from faster RAM in Haswell too. My system is

CPU: Xeon E3-1231v3
BOARD : MSI Z97S SLI Krait Edition
GPU: EVGA GTX 970 SC

and in GTA V I could force a drop to about 52 fps in this one area near Mt Chiliad with lots of shadows when using my DDR3-1600 RAM. After upgrading to DDR3-2400 on the exact same version of the game my framerate would only drop to 57-58 fps. This is at 1080p mostly ultra settings (grass moved down to very high). It was a 100% reproducible result and the only difference was the RAM. I also noticed a similar improvement in Fallout 4. On ultra settings except for god rays and shadow distance being both set to medium I would always get a drop to ~51fps or so with DDR3-1600 inside the speakeasy in Concord at this one spot where light comes in through a window and illuminates a bunch of particles and crap floating around in the air. With DDR3-2400 it didn't drop below 60 fps. This was back before the 1.3 patch came out that added a ton more graphical options like HBAO+ though.

I bring up these two games because I specifically remember exact places to trigger these minimum framerates and I know what the minimums were with the DDR3-1600 kit. The fast RAM was no help whatsoever in Crysis 3 though even though that's a game that hammers cpus when run at Very High settings. Overall I'd say it made no difference in most games I ran on both setups, since most were 100% gpu bound with a 970 at 1080p. But the higher minimum framerates in those two games were very noticeable to me. If I get an occasional drop to 57 fps or so I don't usually notice it unless I have a frame counter on, but a drop to 52 fps is immediately detectable to me.
Could it be then that there was fairly little changes to Skylake aside from the new memory controller? Could explain Broadwell-C's quirks with that eDRAM too.
 

FlanK3r

Senior member
Sep 15, 2009
312
37
91
4790K 4.8 GHz will be similar as i7-6700K at 4.6 GHz. SOmetimes will be Devils Canyon better, soemtimes a little worse. In average the same.
 

guskline

Diamond Member
Apr 17, 2006
5,338
476
126
A 4790k at 4.8Ghz is a great overclocker. My 4790k is rock solid at 4.7Ghz and I'm happy.

I defer to FlanK3r on whether or not a 4.8Ghz 4790k is comparable to a 4.6Ghz 6700k since he has owned both cpus.
 

Dasa2

Senior member
Nov 22, 2014
245
29
91
no need to have buyers remorse
if your target was to run games at 144hz 1080p or one of a few cpu limited games then maybe skylake could make a noticeable difference probably not considering your current oc you would have to get very lucky

but at 4k in just about all games your cpu is more than powerfull enough and to run 60fps you really need two gtx1080 overclocked and that still wont be enough for some games you will be wanting the top end cards that are coming around the end of the year

with a single gpu you could probably return your cpu to stock clocks when gaming at 4k and get no change in fps
this does depend on what your playing though
 
Last edited:

Xellos2099

Platinum Member
Mar 8, 2005
2,277
13
81
Can anyone provide the setting they used for 4790k? I just got my processor pair with MSI Z97S SLI Krait Edition and i am a little lost on what to do. It just seems so different from oc 2500k. I am currently using A70 crosair cooler.
 

buklau

Member
May 4, 2012
135
0
76
I have the same 4790k delided o/c to 4.8ghz paired up with 16gb ddr3 2666 CL11 and I been itching to switch to a 6700k pair up with 4000+ memory later this year. Just sold my 980 and now waiting for 1070. :D
 

Jacky60

Golden Member
Jan 3, 2010
1,123
0
0
When I joined the forums, Anand was still here and the reviews were much more timely, well designed, and complete. This site was my go to site for benchmarks. Now, however, I dont even go to the home page of the site, just the forums. A lot of good information can be obtained from the forums, taking everything with a grain of salt of course. Anand's gaming benchmarks, especially cpu, are very weirdly designed and seem to be all over the place.

In regards to the original post, Skylake with fast ram can be maybe 10% faster than Haswell overall, even more in a few corner case games.

+2 I totally agree, I still use the forums but never visit
the main page now as reviews elsewhere are more informative.
The CPU gaming benchmarks on Anandtech are especially
worthless and GPU limited and their sole purpose appears
to be to boost the word count.
 

MajinCry

Platinum Member
Jul 28, 2015
2,495
571
136
The difference between Sandybridge & Skylake in gaming (non-emulation) was, what, 20% in favour of Skylake? An' the ol' margin of error is 5% in these performance comparisons.

Yer ain't going to notice squat difference. But if you do upgrade, feel free to send me the old CPU and motherboard.