GeForce Titan: 2500K vs 4790K (66 games tested)

[BrightCandle]

All depends what you want to test. If you want to see the difference between the CPUs then a lower resolution and setting might be an appropriate way to do it. If however the point is to show the difference as many people would play the game then real settings are the right way to do it. Yes many games are GPU limited, but we can see a few showing gains with more cores, we can see some showing quite significant advantage to HT and if you look at an i3 v i5 it becomes clear that a lot of the big games today want at least 4 cores.

From these tests you can either conclude there isn't much value in an upgrade because most games don't benefit, or you can conclude that some key AAA PC games show moderate to significant gains from haswell even at high settings.

 

[YBS1]

That's the point - to analyze real world performance of a CPU not reduce IQ to show differences between CPUs in order to justify upgrades. We do not care about 1280x1024 CPU benches on a Titan since who will play at such resolution with such an expensive GPU. Of course for someone running 2 overclocked 780s/290s at 1080/1440, there will be a larger difference between an overclocked i5 and 4790k @ 4.8Ghz. However for the majority of single-GPU owners there is still little legitimate reason to upgrade the CPU platform from 2nd gen OC i5/7s. I would bet i5 2500k @ 4.5Ghz paired with dual 880s will destroy an i7 4790k @ 4.8Ghz or even i7 5930 @ 5ghz with 780Ti SLI in 98% of games.

I understand your point, but my reasoning is for CPU testing I want to see what the CPU is capable of. Reason being in the current PC climate you're likely to go through multiple GPU upgrades before a CPU upgrade. While my X58/W3520 was my main gaming pc it spanned all the way from the 4870/280 generation to the 7970/680 generation. While the real world results between it and a lesser CPU might not have been striking back in the 4870 days, it certainly would have been with the 680. Further, I always keep in mind my CPU is important for more activities than gaming in which it won't be hamstrung by a lack of GPU power.

 

[toyota]

Which is what I have been saying for some time now however someone will always come along and state that you need a top of the line cpu to play computer games and try to justify it. Those charts show little to no difference across most games when comparing an older i5 to a new i7 that has significantly higher clock rates. When it comes to gaming your always better off spending the bulk of your money on the video card setup or even an ssd. Also when you get into games getting over 100fps I doubt many people here can say that they can see a serious difference between 140fps and 170fps. If you can than kudos to you because I cant and could care less as long as the game is running smooth around 60. Interesting to see games like bioshock or crysis that I play somewhat often have no fps difference or 3-4 tops.

he did not use HT so it was not like using an i7. not that HT would help in any of the games he tested anyway though. and a 2500k is hardly old and even at stock can maintain over 60 fps in all but a few games so of course a 4790k is not going to be faster in 98% of cases at 2560 with settings he is using.

that said he really missed the boat on the Crysis runs as that game will drop into 40s with a stock 2500k in some areas. its basically impossible to stay above 60 the whole time in that game no matter how much hardware you throw at it.

and if someone had a 120/144 hz screen they will find themselves very disappointed as even an oced 4790k cant hold those framerates in many games no matter how much gpu power that you have.

what I like about my 4770k is the ability to play games without having to worry about my cpu getting close to maxed out. before there were some games hitting over 90% cpu usage and Crysis 3 was pegging it at times. Crysis 3 runs smoothly now and I can use vsync finally. its still sad that its such a miniscule overall improvement though.

 

[poohbear]

no BF4 benchmarks? a lot of future games will be based on the FB3 engine, so would've been nice to see.

 

[Attic]

Crysis 3 with HT on/off on the 4770k between stock and 4.2ghz.

http://www.youtube.com/watch?v=ZJ7eOUA-5VY



Wolfenstein The New Order needs quite a bit of CPU horsepower to maintain 60fps mins IME.

WatchDogs appears impossible to get 60fps mins.


What stuns me is how little improvement has been made in the $200 area for intel since the 2500k. The new haswells are better but to get a real boost from OC'd 2500k in CPU limited games you need intel 6c/12t cpu's.

 

[Cerb]

You da man BFG10K.

I just upgraded to a 4690k yesterday from i3-4330. My main concern was to go i7 for HT or not. Given it's roughly $100 I think with your look here we can say for a majority of folks buying a CPU for gaming that the i7 is not worth it, but definately get an i5 over any i3.

It can help with some games, but some it won't. Some it still hurts min times for. Given the higher cost, IMO, it's not yet worth it for gaming, unless you're choosing between diminishing returns on GPU v. diminishing returns on CPU, with a sufficient budget (like i5+780 v. i7+770 at 1080P, or higher). Games are generally going to basic work queue models, where what needs doing goes into a queue and is then processed by a thread that needs something to do. So in the long run, HT will work out well across the board, as long as Intel keeps beefing up the ALUs and caches.

I haven't encountered any problems with my Xeon E3 and HT, thus far, with older games, either, but I also don't know what specific titles BFG10K had issues with, or those issues.

 

[escrow4]

Crysis 3 with HT on/off on the 4770k between stock and 4.2ghz.

http://www.youtube.com/watch?v=ZJ7eOUA-5VY



Wolfenstein The New Order needs quite a bit of CPU horsepower to maintain 60fps mins IME.

WatchDogs appears impossible to get 60fps mins.


What stuns me is how little improvement has been made in the $200 area for intel since the 2500k. The new haswells are better but to get a real boost from OC'd 2500k in CPU limited games you need intel 6c/12t cpu's.

I was running Wolfenstein on my stock 4770 and it maintained a solid 60FPS 95% of the time. There were next to no dips. At all. Crysis 3 physics are all over the place, hence the massive dips.

 

[Carfax83]

Crysis 3 with HT on/off on the 4770k between stock and 4.2ghz.

http://www.youtube.com/watch?v=ZJ7eOUA-5VY

A solid 10 FPS boost, which is very significant. Hyperthreading will only continue to get better and give greater returns as 3D engines become more multithreaded and Intel refines the technology..

All the next gen engines/games will be capable of using at least 6 threads, which makes having an HT enabled CPU worth the extra cost.

 

[Attic]

It's hard to imagine highly clocked i5's not being suitable for 3-4 years. Crysis 3 stands out in that area as a point of interest. Sometimes I think it comes down to poor coding. i7 4770k@4.2ghz and it's 50fps in that section? I guess we're not sure if that's gpu limit, but given HT off was 40, it looks CPU limited.

Watchdogs is also insane trade off for massive CPU grunt vs frame ceilings. In watchdogs the gameplay deliver for that grunt is lackluster, the section of Crysis in that test was more fun IMO than all of Watchdogs for me so far. Hell, even watching that crysis 3 bench may have done more for me than Watchdogs up to this point.

Ok watchdogs rant /off

I think 6c/12t will be the next place to get excited for, for folks who were rousted by the 2500k and 2600k delivery of their time. That being said, the 4790k is an exciting chip and i'm somewhat rethinking my purchase of not going with it. Still at this time, that extra chunk of money for pure gaming is not quite worth it IMO.

Hard because most games are not going to even tap the newer i5's at close to 4ghz, but then some games an overclocked 4770k isn't up to the task of 60fps mins at high details.

 

[toyota]

It's hard to imagine highly clocked i5's not being suitable for 3-4 years. Crysis 3 stands out in that area as a point of interest. Sometimes I think it comes down to poor coding. i7 4770k@4.2ghz and it's 50fps in that section?

Watchdogs is also insane trade off for massive CPU grunt vs frame ceilings.

I think 6c/12t will be the next place to get excited for, for folks who were rousted by the 2500k and 2600k delivery of their time. That being said, the 4790k is an exciting chip and i'm somewhat rethinking my purchase of not going with it. Still at this time, that extra chunk of money for pure gaming is not quite worth it IMO.

Hard because most games are not going to even tap the newer i5's at close to 4ghz, but then some games an overclocked 4770k isn't up to the task of 60fps mins at high details.

perhaps look closer as its his gpu that becomes the limitation at 4.2. with a faster gpu he would be over 60 fps in that section.

there are always going to be games that need more cpu or more gpu or more vram. if you want to always get the best out of every game then you have to be willing to spend more money and or upgrade more often.

 

[Deders]

That grass section is CPU limited, there was a test I saw recently (and I mentioned it in an earlier post) where they were comparing an i5-4690 and a i7-4790, both at 4.6 GHz and the i7 managed to keep the framerate above 60fps for 98% of the time, the i5 would dip below much more often.

It's a section of the game that is notorious for CPU usage, it uses all 4 cores at 100% on my rig, and Brightcandle said the same for his 6 core.

 

[toyota]

That grass section is CPU limited, there was a test I saw recently (and I mentioned it in an earlier post) where they were comparing an i5-4690 and a i7-4790, both at 4.6 GHz and the i7 managed to keep the framerate above 60fps for 98% of the time, the i5 would dip below much more often.

It's a section of the game that is notorious for CPU usage, it uses all 4 cores at 100% on my rig, and Brightcandle said the same for his 6 core.

again you can see the gpu usage go to 98-99% when he oced the cpu to 4.2. with faster gpu the fps would have gone up there. there are other parts of the game that are more cpu limited than that.

 

[BFG10K]

Further, amongst users running higher than the current de facto standard of 1920x1080, I would expect multi-GPU usage to climb disproportionally as compared to those running lesser resolutions. Thus running multiple GPUs in your testing would have shifted the burden in many of those games back to the CPU.

Not necessarily, not if one increases the GPU settings, as multi-GPU users should be doing. Increasing graphics settings (e.g. 4K resolutions) is the whole point of multi-GPU.

 

[BFG10K]

that said he really missed the boat on the Crysis runs as that game will drop into 40s with a stock 2500k in some areas. its basically impossible to stay above 60 the whole time in that game no matter how much hardware you throw at it.

Like I said, at the settings I use I'm GPU bound in the areas that perform the worst. Oubadah posted 41/55 FPS in this area:

I notice a distinct lack of AA in your screenshot, which means you too choose to lower certain settings to get better performance.

 

[toyota]

Like I said, at the settings I use I'm GPU bound in the areas that perform the worst. Oubadah posted 41/55 FPS in this area:

http://s29.postimg.org/g32f562lz/Crysis.jpg

I notice a distinct lack of AA in your screenshot, which means you too choose to lower certain settings to get better performance.

what does AA have to do with what I was saying? I said it drops below 60 fps with a 2500k and that is FACT. so your logic as usual is that I should turn up the AA until I am gpu limited instead of cpu limited? who cares in this particular case as the point is that a stock 2500k itself does not have the ability to keep the framerate from low 50s or even high 40s. most people buy a high end gpu to stay above 60 fps and you cant do that in the old Crysis which again was the point here.

 

[tidy]

.

 

[Black Octagon]

Gogo Sandy! Still haven't replaced mine. Thanx BFG

My only very small criticisms would be:

1) should have benched with both CPUs OC'd to at least 4.2-4.5Ghz. The 2500k is SUUUCH an easy OC-er that running it at stock makes no sense.

3) Should have enabled HT on the 4790k. No point holding back a feature like that in my view. The 2500k will hold its own regardless, and it's good to know exactly how much better the newer CPU is

 

[RussianSensation]

While the real world results between it and a lesser CPU might not have been striking back in the 4870 days, it certainly would have been with the 680. Further, I always keep in mind my CPU is important for more activities than gaming in which it won't be hamstrung by a lack of GPU power.

Totally understand your point; if a gamer wants to keep the CPU for 2-3 generations of GPUs, then it is better to spend a little extra for the i7 over i5 no doubt. $100 is not a lot over 3-4 years of CPU ownership to cover the instances where HT will give a noticeable boost in CPU limited cases. But given the current PC climate (Skylake out in 12-15 months), and X99 out in September, maybe i5 2500/2600k users maybe should consider the 5820 OC not 4790K or just wait it out to a new socket, PCIe 4, DDR4 and as a bonus 14nm overclocking should allow 5Ghz OC hopefully.

 

[Pandora's Box]

Can't believe we have gone 3 pages without anyone mentioning the lack of minimum fps numbers. I would think those numbers would show the difference (if any) between processors.

 

[Attic]

That grass section is CPU limited, there was a test I saw recently (and I mentioned it in an earlier post) where they were comparing an i5-4690 and a i7-4790, both at 4.6 GHz and the i7 managed to keep the framerate above 60fps for 98% of the time, the i5 would dip below much more often.

It's a section of the game that is notorious for CPU usage, it uses all 4 cores at 100% on my rig, and Brightcandle said the same for his 6 core.

For those interested, that vid is about half way down in this article

 

[Deders]

For those interested, that vid is about half way down in this article

Nice 1

 

[BFG10K]

Can't believe we have gone 3 pages without anyone mentioning the lack of minimum fps numbers.

A minimum by itself is worthless since any random event can affect it given it's a single point in time (by definition).

You either need sustained minimums or some kind of frame over time analysis graphs. Both were outside the scope of this exercise, which is primarily to test games nobody else does.

 

[BD2003]

It doesn't seem like older CPUs have any issue keeping up with 60fps....but how does the story change when the target is 120fps, which isnt all that uncommon nowadays? I suspect the differences will become much more evident.

 

[Ausm]

Nice post but no test on WOW??

 

[tidy]

.