Why the 4690K is a underperforming bottleneck!

[Overkilled]

Ok so this here is my Firestrike score

 

http://www.3dmark.com/3dm/9017242

 

And this is a i7

 

Theres no real performance benefit unless youre doing rendering they said. The 4690k performs just as well in benchmarks and games they said.

Edited October 24, 2015 by Godlygamer23

[OAcesync]

Ok so this here is my Firestrike score

 

http://www.3dmark.com/3dm/9017242

 

And this is a i7

 [spoiler]  [/spoiler]

 

Theres no real performance benefit unless youre doing rendering they said. The 4690k performs just as well in benchmarks and games they said.

It's called synthetic benchmarking for a reason

[Interfectorem]

There is a Physics test in 3DMark, which the 5820K will perform better in and can affect the overall score in the end. Not to mention that the card in the 5820K test could have been overclocked, if you ever look at benchmarks using 3DMark, GPU overclocking really boosts the score.

[Overkilled]

 

There is a Physics test in 3DMark, which the 5820K will perform better in and can affect the overall score in the end. Not to mention that the card in the 5820K test could have been overclocked, if you ever look at benchmarks using 3DMark, overclocking really boosts the score.

*plus 620mhz on the mem. My graphics score isnt the problem its the physics and combined
Sure its got more cores but the difference shouldnt be that big. Even the 4790k can easily break 10k on the physics score

[Interfectorem]

 

*plus 620mhz on the mem. My graphics score isnt the problem its the physics and combined

Of course physics score is much better on the 5820K because it's purely a CPU intensive task. The 5820K is a more powerful processor in these regards, thus resulting in a higher score. I don't see the surprise here.

 

 

Sure its got more cores but the difference shouldnt be that big. Even the 4790k can easily break 10k on the physics score

We also don't know what (and if) the overclock is on the 5820K. Let me download 3DMark real quick and run the test on my system, which is a 5820K (4.4 GHz) + GTX 970 (1565MHz)

[Overkilled]

r

Of course physics score is much better on the 5820K because it's purely a CPU intensive task. The 5820K is a more powerful processor in these regards, thus resulting in a higher score. I don't see the surprise here.

Obviously the difference is there but I never expected near 10,000 more on the physics score

[Interfectorem]

r

Obviously the difference is there but I never expected near 10,000 more on the physics score

Four 3.5 GHz threads is quite aways from twelve (potentially) 4.0+ GHz threads. We'll see when I download 3DMark, it'll be about 20 minutes and I'll report back.

[Overkilled]

Four 3.5 GHz threads is quite aways from twelve (potentially) 4.0+ GHz threads. We'll see when I download 3DMark, it'll be about 20 minutes and I'll report back.

No my cards at 4.3 if it was 3.5 it would barely muster up 7k

[App4that]

I've been saying it, but no one listens.

[spwath]

Of cource the i7 does better in benchmarks....

 

But games it wont matter.

[AlTech]

Of cource the i7 does better in benchmarks....

 

But games it wont matter.

Unless your running triple SLI or Crossfire it won't matter and by that point you should be looking at a six core i7 or an eight core for more PCI-E lanes rather than pure performance.

[Godlygamer23]

If you actually look at the individual benchmarks, you'll find something interesting. The 4690K gets over 14,000 points for Graphics, 8440 points for Physics, and 5288 points for the Combined Score. On the other hand, the 5820K gets 13,672 points for Graphics, 17,701 points for Physics, and 5058 points for the Combined Score. Your outlier is the Physics score which I believe is able to use more than 4 threads. The other two tests, the 4690K actually ends up doing better with the 970 versus the 5820K.

 

The title of this post is sensationalist.

Edited October 24, 2015 by Godlygamer23

[harrynowl]

Did you even read your own results...?

 

It's quite clearly the difference in the physics score that makes up for the higher mean score on the i7 system. Which makes sense because the i7 is the faster CPU.

 

What it doesn't translate to is a bottleneck in a real-world scenario.

[Interfectorem]

No my cards at 4.3 if it was 3.5 it would barely muster up 7k

Here are the results on my system. 

 

http://www.3dmark.com/3dm/9017872?

[Overkilled]

Did you even read your own results...?

 

It's quite clearly the difference in the physics score that makes up for the higher mean score on the i7 system. Which makes sense because the i7 is the faster CPU.

 

What it doesn't translate to is a bottleneck in a real-world scenario.

sigh I know its the physics score the i7 is obviously faster but surely the difference shouldnt be that big?

[Godlygamer23]

sigh I know its the physics score the i7 is obviously faster but surely the difference shouldnt be that big?

Why can't it be? That's more than double the physics score of the 4690K. The 5820K has 6 cores with hyper-threading. If the test can use all 12 threads, that would explain why there's a big difference.

[NumLock21]

sigh I know its the physics score the i7 is obviously faster but surely the difference shouldnt be that big?

Well looks like it is. Time to upgrade.

[Glenwing]

Ok so this here is my Firestrike score

 

http://www.3dmark.com/3dm/9017242

 

And this is a i7

 

Theres no real performance benefit unless youre doing rendering they said. The 4690k performs just as well in benchmarks and games they said.

 

If you want to get real-world performance numbers, you have to actually test real-world programs. So which games did you test?

 

You do realize that the entire point of a CPU-only benchmark (which is what a physics test is) is to show the maximum computational power of the CPU, right? The difference you see in those scores is precisely why the i7 destroys in rendering, but it's not and was never intended to be taken as an indicator of game performance.

[App4that]

Witcher 3 shows improved fps in cities when using a 4790k verse a 4690k. A better CPU WILL give you higher fps. Why this forum is so unwilling to admit this I will never know.

[Megahurt]

Of course if your only aim is to get the best 3D mark score you can, an i7 is better than an i5, because the physics test isolates the cpu, and the i7 is obvioiusly a better cpu, but that doesn't mean the i5 is causing any bottlenecks in real world situations.

[Monarch]

The difference you see in those scores is precisely why the i7 destroys in rendering, but it's not and was never intended to be taken as an indicator of game performance.

 

Some games do have complex physics.Take Crysis 3 for example. Grass physics destroys i5's. The difference between 4690k and 4790k with HT is up to 30%.

 

 

Another example is Witcher 3, which is CPU-bound in cities and villages and you get microstuttering even with a 4790k. 

 

 

So imo the 3dMark's physics test is an indicator of how different CPUs will perform in CPU-heavy scenarios in games. 

 

Witcher 3 shows improved fps in cities when using a 4790k verse a 4690k. A better CPU WILL give you higher fps. Why this forum is so unwilling to admit this I will never know.

 

Thank you.

[CalebTheEternal]

r

Obviously the difference is there but I never expected near 10,000 more on the physics score

You also need to look at other i5 CPUs you're looking at all i7 which have hyper threading. 

[Sauron]

Firestrike is meaningless

[Overkilled]

If you want to get real-world performance numbers, you have to actually test real-world programs. So which games did you test?

 

You do realize that the entire point of a CPU-only benchmark (which is what a physics test is) is to show the maximum computational power of the CPU, right? The difference you see in those scores is precisely why the i7 destroys in rendering, but it's not and was never intended to be taken as an indicator of game performance.

Knowing my setup can run any game at a decent framerate at high settings on 1080p is enough. Having a good firestrike score is way more importajt as long as my games play at 60 or more fps

[Monarch]

Firestrike is meaningless

 

No, it's not. i5's perform significantly worse than i7's in CPU-heavy scenarios, which is exactly what Firestrike shows.