• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Real tangible difference between i7 and i5 at same fps?

opethdisciple

opethdisciple

opethdisciple

Soldato
Joined
18 May 2010
Posts
23,335
Location
London
Lets just forget about DX12 for a second, which supposedly will favour cores over clockspeed.

What I'm trying to work out is, is there any real world difference between two cpus one with more threads/cores and one with less threads/cores but both deliver the same fps?

Looking at benchmarks doesn't give me a real feel for if there is difference. The problem is you look at benchmarks and they show that tow different cpu architectures both get the same fps, so is there a real tangible difference?

Take this as an example:

5820k with a 980 = 45fps in a particular game
3570k with a 980 = 45fps in a particular game

Would you notice any difference between the feel of the game?

I'm trying to determine if I should upgrade my 3570k to a 6770k or a 5820k (with a 970).
 
muon said:
A great video here. Their channel is very good, I recommend subscribing.

https://youtu.be/WZ_5p9wd2dk

https://youtu.be/JWxncqbe1H8

The answer is, even with trying to achieve a CPU bound scenario, the difference isn't much. At more normal settings, there is even less of a difference.
Click to expand...

I'll check the videos out on a sec.

As an example, I've been watching videos on you tube of other people 5820k's. I saw one video which showed that the Witcher 3 scaled across 6 core of the 5820k. Meaning my 3570k is holding the 970 back as it has 2 less physical cores.

However if both GPU's are the same and the fps is comparable, is there any real world difference?
 
Witcher 3 is capable of maxing out the 6600K at stock. They mention it in one of their videos, but the framerate still is high and doesn't appear to be holding the Titan X OC they are using back.

They will release a video soon with the 6700K and they might also include a comparison with the 5820K.
 
Last edited:
it all comes down to the game itself and whether it can utilise more cores/threads

if it can then the chip with the most cores/threads will be much better

if it cant then there will be no difference between a multithreaded/core chip and one without
 
from kitguru
http://www.kitguru.net/desktop-pc/zardon/ocuk-titan-riptide-x99-system-review/21/
Luke Hill’s opinion:
With such a small pricing gap between a system based around the Haswell-E i7-5820K, an entry-level X99 motherboard, and 16GB of quad-channel DDR4 and one that utilises the Skylake i7-6700K, a mid-range Z170 motherboard, and 16GB of dual-channel DDR4, the decision process is a tough one. As we highlighted in the Skylake CPU review, those who prioritise cores, cache, and plentiful PCIe connectivity should opt for the 5820K system. Those who value a feature-filled platform with many of the latest motherboard goodies, such as USB 3.1, will be best served by a Skylake system.
Focusing on CPUs, the 5820K has extra cores, more cache, and plenty of PCIe lanes in its favour, however the 6700K can typically be overclocked higher, has better IPC from newer architecture, and is equipped with an iGPU which may (and this remains to be seen or validated) bring benefits with DX12’s multi-GPU options. There’s unlikely to be a noticeable performance difference between the 5820K or 6700K CPUs in games, although GTA V has shown that a well-optimised game engine can indeed take advantage of several threads (we saw 10 threads being utilised in our testing shown here).
I think that OCUK’s decision to opt for the 5820K is a smart one. Coupling it with the incredibly popular GTX 970 GPU leaves plenty of room for an SLI upgrade in the future. The extra PCIe lane flexibility for an X99 system also means that a PCIe SSD is a viable future upgrade without significant change being made to the hardware. And for those times when you aren’t gaming, six well-overclocked Haswell-E cores will provide significantly higher performance in multi-threaded workloads than four fast Skylake cores would.
The sheer quantity of gamers and enthusiasts who consistently reference how happy they are with their 3- or 4-year old Sandy Bridge or Ivy Bridge systems makes me think that CPU performance is more valuable to them than a feature-healthy chipset and overall platform. And it is for that reason that I think opting for the higher-performance, six-core Haswell-E 5820K is the correct decision by OverclockersUK./QUOTE]

if that helps
Click to expand...
 
Very VERY VERY little difference unless you are running multiple GPU's. Then, its only a very VERY little difference :)
 
muon said:
A great video here. Their channel is very good, I recommend subscribing.

https://youtu.be/WZ_5p9wd2dk

https://youtu.be/JWxncqbe1H8

The answer is, even with trying to achieve a CPU bound scenario, the difference isn't much. At more normal settings, there is even less of a difference.
Click to expand...

These were the first clips I thought of as well.

OP - look at the frame time graphs (second link), the i7 is slightly less "dippy" than the i5. That's your real-world difference.

Edit: actually that one is across generations - have a look for an i5 vs i7 one from both Ivybridge/DC for a fairer comparison.
 
Last edited:
opethdisciple said:
I'll check the videos out on a sec.

As an example, I've been watching videos on you tube of other people 5820k's. I saw one video which showed that the Witcher 3 scaled across 6 core of the 5820k. Meaning my 3570k is holding the 970 back as it has 2 less physical cores.

However if both GPU's are the same and the fps is comparable, is there any real world difference?
Click to expand...

Using more cores and needing more cores are not the same thing so your 3570k isn't necessarily holding your 970 back.

On a single GPU there is seldom any real improvement in frame rates or frame times from having an i7 over an i5 at the same clock speed. On multi-GPU setups it's more common to see gains, especially in frame times (smoothness) though framerates can improve too in some games (once you're on e.g. quad graphics the difference will likely be substantial though you'll want to be on an x99 i7 in that case to have enough lanes to run all the cards well, never mind the extra cores... tho these help a bit too :p)

One rare downside of the i7 is that to save power it parks cores/threads and can stutter when then trying to use them - this only happens in some games and is pretty rare but maybe worth mentioning.
 
For "60Hz" gaming there is very little difference between an i5 and i7 in many games (there are a few exceptions) especially if you use vsync - infact sometimes the extra overclockability of an i5 can be a benefit. However I think pure gaming benchmarks can sometimes be a bit misleading - if you have a lot of stuff running as well as the game especially companion programs like voice comms, streaming/capturing, etc. then the extra threading units of the i7 can be of benefit to keep things a little more responsive and smooth. (Even seen cases where the 8 core AMD CPUs can actually give a smoother experience in those kind of scenarios over an i5 even with a little less FPS though that is generally an exception).

There can be a more noticeable difference if like me you like to run 120fps/hz or multiple GPUs sometimes in terms of FPS (i.e. high end multi GPU) or just in frame time consistency which doesn't show in raw FPS numbers - when you have a situation with 4 cores fully loaded versus 8 threads with moderate load spread over them - even though you still have 4 real cores it can sometimes work out smoother even at the same FPS but to really notice that you generally have to go above 60fps/hz gaming.
 
SKYMTL@Hardware Canucks | Posted: 6th August 2015 said:
Since the PCH on X99 doesn't have Gen3 lanes, any storage solutions that require high bandwidth interconnects will grab lanes from the CPU, thus limiting available graphics lanes.

For example, on most X99 motherboards installing two GPUs and an M.2 SSD will result in the second GPU running at x4 speeds.
Click to expand...
Source
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom