7800x vs 7700k, hint: the 7800x is slower(even when overclocked)

[kiska3]

Source: https://www.techspot.com/review/1445-core-i7-7800x-vs-7700k/

Quote

We tested both processors at stock settings as well as overclocked -- the 7700K was set at a relatively easy to achieve 4.9GHz, while the 7800X was at 4.7GHz.

Here are the results:

Spoiler

Power consumption

Well this is what you get for rushing a product release

For those who don't want to read the article Steve has made a video

 

 

[tom_w141]

We knew this anyway? HEDT isn't for gamers... Well not primarily. Games are still quad core optimised. If they were suited to more cores (which they may be in a few years) then you'd see the likes of Ryzen beating 7700k and the 7800X/7820X smashing the 7700k. At the moment due to the quad core focus of games, a highly clocked quad core will always win over more slower cores (assuming IPC is comparable).

 

EDIT: More cores is the better long term purchase and useful for more than gaming. Streaming/video editing for example. Intel knows quad cores won't live forever which is why coffeelake is hotly tipped to be bringing a 6 core SKU to the mainstream platform.

[tlink]

its clocked lower, of course its going to be slower in most games. that doesn't have anything to do with a rushed release.

[kiska3]

3 minutes ago, tlink said:

its clocked lower, of course its going to be slower in most games. that doesn't have anything to do with a rushed release.

If you look at the graphs, Steve also overclocked it to 4.7Ghz and the 7700k to 4.9Ghz, a 5% deficit in frequency

[tom_w141]

1 minute ago, kiska3 said:

Steve also overclocked it to 4.7Ghz and the 7700k to 4.9Ghz

*checks maths textbooks* Yep 4.7 is definitely lower than 4.9 chief. Seriously though I explained the exact reason for the results and why they aren't surprising in my above post.

[Simon771]

3 minutes ago, tlink said:

its clocked lower, of course its going to be slower in most games. that doesn't have anything to do with a rushed release.

 

Just now, tom_w141 said:

*checks maths textbooks* Yep 4.7 is definitely lower than 4.9 chief.

 

And you can clearly see that i7 7700k at stock is still better than 7800X at 4,7GHz. Now I might be wrong, but I think i7 7700k at stock doesn't get to 4,7GHz.

So much about your math

[kiska3]

Just now, Simon771 said:

And you can clearly see that i7 7700k at stock is still better than 7800X at 4,7GHz. Now I might be wrong, but I think i7 7700k at stock doesn't get to 4,7GHz.

So much about your math

Just about to say the same thing, you were just quicker.

Spoiler

And probably Westfield's free wifi, is not optimal

 

[Enderman]

2 minutes ago, Simon771 said:

 

 

And you can clearly see that i7 7700k at stock is still better than 7800X at 4,7GHz. Now I might be wrong, but I think i7 7700k at stock doesn't get to 4,7GHz.

So much about your math

7700k is kabby lake, 7800X is still skylake.

That's why it's slower.

[DarkBlade2117]

1 minute ago, Enderman said:

7700k is kabby lake, 7800X is still skylake.

That's why it's slower.

Idk about you... But I don't think Kaby has that big of an advantage over Skylake... Actually, last I checked the better performance of baby solely came down to the higher stock clocks.

[Simon771]

1 minute ago, Enderman said:

7700k is kabby lake, 7800X is still skylake.

That's why it's slower.

I do understand that. But in some cases there is more than 15% difference in FPS.

There isn't 15% difference between skylake and kaby lake IPC.

And that skylake-x is clocked at 4,7GHz while i7 7700k at around 4,4GHZ ... yeah it does boost to 4,5GHz but not on all cores.

 

At same clock speed, skylake-x should have 10% less FPS at most.

But we can see from graphs that you can clock skylake-x about 8% higher, and will still have 15% less fps in games.

 

But keep something in mind: it's 1080p teting ... 1440p and 4K would be probbably the same FPS.

And X299 is new platform. It still needs some polishing, just like Ryzen did.

 

in afew months, I'm sure Skylake-X will perform better in those games, and difference then will be about 10%, when comparing i7 7700k at 4,7GHz to i7 7800X to 4,7GHz.

[tom_w141]

8 minutes ago, Simon771 said:

 

 

And you can clearly see that i7 7700k at stock is still better than 7800X at 4,7GHz. Now I might be wrong, but I think i7 7700k at stock doesn't get to 4,7GHz.

So much about your math

Good point  I haven't got time to pour over the entire article at work but my guess would be the restructuring of the cache on skylake-X affecting the performance.

7 minutes ago, kiska3 said:

Just about to say the same thing, you were just quicker.

  Reveal hidden contents

And probably Westfield's free wifi, is not optimal

 

 

[tom_w141]

3 minutes ago, Simon771 said:

At same clock speed, skylake-x should have 10% less FPS at most

Erm IPC of Skylake and Kaby lake are the same... So at the same clock speed the difference should be 0%

 

4 minutes ago, Simon771 said:

15% less fps in games.

At stock its around 30% for GTA V for example.

 

5 minutes ago, Simon771 said:

1440p and 4K would be probbably the same FPS

Of course they would... You'll get the same FPS as a Pentium at 4k because you are GPU bound not CPU bound... This is why testing is at 1080p.

 

6 minutes ago, Simon771 said:

in afew months, I'm sure Skylake-X will perform better in those games

Why? You can't download more L3 cache.

[MageTank]

9 minutes ago, Enderman said:

7700k is kabby lake, 7800X is still skylake.

That's why it's slower.

That makes zero sense. Kaby and Skylake perform exactly the same clock for clock (aside from the speedshift tweak). Under a fixed clock workload, they have identical performance.

 

Now, with the L2 cache restructure that was brought about on the X299 SKU's, that is bound to change. For the larger 7900X, it showed a pretty single threaded performance boost over my 7700k (5ghz vs @done12many2's 4.9ghz) with him maintaining an all around lead on cinebench. However, it's yet to be seen whether or not other applications can take advantage of this larger cache pool. I certainly know AVX applications can, and do take extreme advantage of it, as seen by the massive gains in flops. 

[Enderman]

7 minutes ago, DarkBlade2117 said:

Idk about you... But I don't think Kaby has that big of an advantage over Skylake... Actually, last I checked the better performance of baby solely came down to the higher stock clocks.

It was a pretty significant improvement when talking about CPUs.

 

[Enderman]

Just now, MageTank said:

That makes zero sense. Kaby and Skylake perform exactly the same clock for clock (aside from the speedshift tweak). Under a fixed clock workload, they have identical performance. 

Then explain why the 7700k at 4.5GHz performs better than the 6700k at 4.5GHz, as seen in the benchmark above.

[Vode]

12 minutes ago, Enderman said:

7700k is kabby lake, 7800X is still skylake.

That's why it's slower.

This is statement is wrong on two levels.

 

1.) Kaby Lake is the same IPC wise as Skylake

 

2.) 7800X isn't Skylake, it's Skylake X which has a different cache structure and core topology.

[Enderman]

1 minute ago, Vode said:

This is statement is wrong on two levels.

 

1.) Kaby Lake is the same IPC wise as Skylake

 

2.) 7800X isn't Skylake, it's Skylake X which has a different cache structure and core topology.

1) Read above.

Clearly you missed all the benchmarks comparing the 6700k to the 7700k (at the same clock speed) when the 7700k launched.

 

2) The architecture is still skylake.

https://ark.intel.com/products/123589/Intel-Core-i7-7800X-X-series-Processor-8_25M-Cache-up-to-4_30-GHz

 

Just like Haswell-E was the same generation as Haswell.

[Simon771]

2 minutes ago, tom_w141 said:

Erm IPC of Skylake and Kaby lake are the same... So at the same clock speed the difference should be 0%

 

At stock its around 30% for GTA V for example.

 

Of course they would... You'll get the same FPS as a Pentium at 4k because you are GPU bound not CPU bound... This is why testing is at 1080p.

 

Why? You can't download more L3 cache.

Seems like you now took more time that me to check all the graphs lol

 

Why would it get better over time? Well why did Ryzen get better at 1080p? at release it was somehow horrible, but since release to today, it was improved by a lot. Drivers and optimisation.

It won't be any kind of miracle to gain 50% more FPS, but I can imagine 10-20% more with further updates and optimisations.

[MageTank]

6 minutes ago, Enderman said:

Then explain why the 7700k at 4.5GHz performs better than the 6700k at 4.5GHz, as seen in the benchmark above.

Margin of error. Go back and watch your own evidence. The 6700k is barely winning in The Division, and Tomb Raider, while trading blows in every other titles by 1-2% average at most. Where in that video is the 7700k performing better?

 

EDIT: @Enderman i'll save you the time buddy, DF provided the details in their article: 

 

The biggest swing in framerates was 2fps at most. Please, tell me you sincerely believe this isn't margin of error, and that Kaby holds some sort of secret IPC advantage. 

[Enderman]

3 minutes ago, MageTank said:

Margin of error. Go back and watch your own evidence. The 6700k is barely winning in The Division, and Tomb Raider, while trading blows in every other titles by 1-2% average at most. Where in that video is the 7700k performing better?

\

 

 

 

 

 

 

 

Are you kidding me?

Like half the video has the 7700k clearly above the 6700k...

 

There are many other benchmarks online that show the same, go look them up.

[Vode]

1 minute ago, Enderman said:

1) Read above.

Clearly you missed all the benchmarks comparing the 6700k to the 7700k (at the same clock speed) when the 7700k launched.

 

2) The architecture is still skylake.

https://ark.intel.com/products/123589/Intel-Core-i7-7800X-X-series-Processor-8_25M-Cache-up-to-4_30-GHz

 

Just like Haswell-E was the same generation as Haswell.

Where does the 7700k win? The first 8 minutes are at stock, the 7700k is clocked higher...

 

Haswell-E has the same cache hierachy as Haswell. Skylake-X does not have the same as Skylake.

[MageTank]

1 minute ago, Enderman said:

\

 

 

 

 

 

 

 

Are you kidding me?

Like half the video has the 7700k clearly above the 6700k...

 

There are many other benchmarks online that show the same, go look them up.

Oh hey, you used the beginning of the video, where they are running each CPU at stock clocks (4.5ghz all-core max vs 4.2ghz all-core max)? Good thing you told me to look at the 4.5ghz results instead, huh? Thanks for invalidating your own argument for me. You do this every single time to speak to me, and yet you never learn, lol. 

 

EDIT: Because I like my beatdowns to be thorough, here is your source, the one you chose to use against me: http://www.eurogamer.net/articles/digitalfoundry-2017-intel-kaby-lake-core-i7-7700k-review

 

And because you used an image to express your feelings, i'll use one too. 

 

[Enderman]

1 minute ago, Vode said:

Where does the 7700k win? The first 8 minutes are at stock, the 7700k is clocked higher...

 

Haswell-E has the same cache hierachy as Haswell. Skylake-X does not have the same as Skylake.

Watch the other half of the video.

You will see the blue line is almost always above the green line.

It's not that hard to see tbh, it's at the bottom left in case you didn't notice.

 

If they had the same IPC then you would only be able to see the blue line the whole time because green would be underneath.

[tom_w141]

7 minutes ago, Simon771 said:

Why would it get better over time? Well why did Ryzen get better at 1080p?

It will get a bit better by using Ryzen like optimisations such as better multithreading, using cores within 1 CCX (Ryzen only). What the dev can't do anything about is the lower L3 cache per core which is why I highlighted it as my best guess at why there is a performance dip.

[Enderman]

1 minute ago, MageTank said:

Oh hey, you used the beginning of the video, where they are running each CPU at stock clocks (4.5ghz all-core max vs 4.2ghz all-core max)? Good thing you told me to look at the 4.5ghz results instead, huh? Thanks for invalidating your own argument for me. You do this every single time to speak to me, and yet you never learn, lol. 

Look at the second half of the video...

Holy crap dude, it's not hard to look at two lines...

 

 

 

If you somehow can't tell that the two lines appart, you might want to get your eyesight checked, because it is obvious that the green and blue lines are not the same.