From a i5 2500k to a i7 6700k: my impressions

Dave2150 · 11 Mar 2016 at 20:16

Scramz said:
This has been extracted from another forum. Hopefully this will help people understand the performance difference from a 2500k to a 6700k both clocked at 4.5Ghz. From this you will see the increased IPC performance gains.

Good find. Massive gains to minimum FPS - lets use this next time lunatics claim it's not worth upgrading from a sandybridge

Columbo · 11 Mar 2016 at 21:11

Ive recently gone from an i5 750 at 4.4ghz to a 6700k 4.5ghz and honestly feel like I've totally wasted my money, I couldnt be underwhelmed more. Games are certainly no better to my eye on 780 SLI at 1440p however that could be related to the games I play. Benchies certainly tell I'm getting more but I just dont 'feel' it.

Its great that the OP feels the boost of an upgrade but I cant help but think overclocking his 2500k would have yielded similar results, alternatively though its always lovely to have something new and shiny to play with and that feeling cannot be benchmarked!

Hedge · 11 Mar 2016 at 21:15

I went 2500k at 4.5 to 5820k at 4.5... Is it faster. Yes. Is it massively faster. Well.. Meh I guess money would be better on a 980ti

Guest2 · 11 Mar 2016 at 23:36

Install and run Cinebench R15 and see what score you get

http://www.maxon.net/products/cinebench/overview.html

~>Dg<~ · 12 Mar 2016 at 00:14

some people notice things some dont.

i guess its how receptive you are to fps changes and things like that.

CAT-THE-FIFTH · 12 Mar 2016 at 00:38

~>Dg<~ said:
some people notice things some dont.

i guess its how receptive you are to fps changes and things like that.

But also just randomly pushing graphs means nothing - what do they mean in actual gameplay?? Its why websites like The Tech Report look at percentile results too.

Plus how come all these comparisons compared a £150 Core i5 2500K with a £300 Core i7 6700K which has twice the threads??

Moreover,is the difference actually noticeable in a double blind study??

All the test sections are short in the post with the frametimes - some are like 15 seconds and some are only a minute.

However about something like 5 to 10 minutes of gameplay for each game with a video too??

You could easily find one area where one CPU fares worse than an other.

The Tech Report showed this years ago when they started their frametimes measurements,and showed even for graphics cards you could easily find one area where an AMD card fared worsed than a Nvidia card and vice versa.

I still remember this with Skyrim.

I tried the game with a GTX660 and an HD7850 and found one part of the game where the GTX660 would stutter and an HD7850 was smooth,and another place were the opposite was true. The same is true with CPUs.

This is the other thing - even if a CPU does improve performance over a Core i5 2500K,you will tend to find a bigger performance improvement going to a faster graphics card.

wam7 · 12 Mar 2016 at 02:00

CAT-THE-FIFTH said:
...This is the other thing - even if a CPU does improve performance over a Core i5 2500K,you will tend to find a bigger performance improvement going to a faster graphics card.

This sums it up perfectly.

Guest2 · 12 Mar 2016 at 08:53

On the performance of what though? A gpu intensive game of course will see bigger improvements with a better gpu. Try playing a cpu intensive game like starcraft 2 with 8 players or supreme commander. A gpu wont make any difference to calculations required when theres a lot going on.
A gpu wont do much for reducing encoding time either or general application use

Scramz · 12 Mar 2016 at 11:42

CAT-THE-FIFTH said:
Plus how come all these comparisons compared a £150 Core i5 2500K with a £300 Core i7 6700K which has twice the threads??

Because 2500K owners in this thread want to see if it is worth the money to performance to switch to a 6700K. If I owned a 2500K I would like to know this information before I purchased too.

It has been said for mean years that a I5 is all you need for gaming and moving to an I7 has kind of not worth it. Which to be honest you can clearly see from them results moving from 4 to 8 threads does not really increase gaming performance, but with the IPC improvements you do. Saying that, you will get the same gaming performance boost moving to the SL I5 as you would the SL I7.

Only reason to move up to a I7 is if you use multi threaded apps but then you would be looking at a Haswell-E over the SL i7 if that was the case.

lumik · 13 Mar 2016 at 04:48

ALL MMOs, RTS, Most RPG, Open world games like watchdogs.Those are games that are CPU bottleneck.
Also moded skyrim/Fallout4 is CPU bottleneck in scenes where are many NPC like cities.
Btw that IPC difference is pretty Big.
Skylake have 30-40% better IPC than sandy bridge in those games.
To match 4.5Ghz skylake you need 6-6.3ghz sandy ridge.Thats not small difference.

theMilitant · 13 Mar 2016 at 18:55

Hedge said:
I went 2500k at 4.5 to 5820k at 4.5... Is it faster. Yes. Is it massively faster. Well.. Meh I guess money would be better on a 980ti

I guess this quite succinctly encapsulates the point that people who keep saying a 2500k is fine for gaming are trying to make: To this day, you will still get more of a performance jump by upgrading your GPU rather than a 2500k.

Guest2 · 13 Mar 2016 at 19:31

theMilitant said:
I guess this quite succinctly encapsulates the point that people who keep saying a 2500k is fine for gaming are trying to make: To this day, you will still get more of a performance jump by upgrading your GPU rather than a 2500k.

This exactly. 6700k is apparently 20% faster than a 2500k clock for clock. I went from an i5 760 > i7 4790k for maybe 20-25% increase.

However, GPU wise went from an 8800 GT > GTX 670 and saw a huge gaming improvement. Both the GPU and CPU age gap was around 5 years.

If you play RTS games though, a GPU won't help much when you have huge maps with lots going on. My i5 760 used to grind to a halt with sort of thing but the i7 handles it fine

David Bisset · 13 Mar 2016 at 20:02

Can bring any CPU to it's knees if you go mad in SC2 unit-tester - just spawn a full ling&ultra army (ultras over lings) and do the same for other team, in close proximity. Only need 2 players worth of supply, think the auto-attacking logic is pretty heavy with so many targets to chose between, grinds to a messy halt :/

Edit: Not seen that vid in ages

love it for how oblivious Husky is...

CAT-THE-FIFTH · 13 Mar 2016 at 20:34

Guest2 said:
On the performance of what though? A gpu intensive game of course will see bigger improvements with a better gpu. Try playing a cpu intensive game like starcraft 2 with 8 players or supreme commander. A gpu wont make any difference to calculations required when theres a lot going on.
A gpu wont do much for reducing encoding time either or general application use

Dude I play games like SC2,SINs and SupCom and SupCom2 - if you actually played some of these games at LANs,unless ALL your mates have 10GHZ Core i7 6700K CPUs,games like SupCom are limited by the slowest CPU in the network.

Plus one of the BEST players I knew in SC2 was a decent Diamond League player and he was running the game on a Phenom II X4 965 and thrashing people having SB and IB Core i5 and Core i7 CPUs.

Also,some of the best players worldwide are running the games on laptops(not even the full sized ones),which most likely during extended games are not even Turboing fully due to thermal issues.

People who play these games competitively drop all the details down anyway - that is what I have seen.

Plus piddly little 20% to 25% increases mean bugger all if during the worst aspects of these RTS games - because 12.5 FPS is so much better than 10FPS.

Even with other CPU limited games like D3 during the Rifts,I do get better framerates than my mates with AMD chips with an IB Core i7 yet when the massive dips happen it happens to everybody and looking at the D3 thread people even with Haswell chips have seen the same.

Sometimes the limitations of online games are on the server side - EVE is a famous example.

Someone running SC2 or SINs on an HD6450 with a Core i7 6700K is probably going to have a much worse experience running it with an GTX970 and a Core i7 3770K or a Core i7 4770K.

Let's look at something like Total War:Attila.

The difference from going to a Core i5 2500K to a Core i7 4770K is much less than going from a GTX780 to a GTX980.

If you had a weaker card like a GTX680,it is even a bigger increase.

Its a total misnomer that people think many strategy games are not GPU heavy and are exclusively CPU heavy- more and more are using the compute capabilities of the graphics card to help the game run better or to increase complexity.

Games like Civ5,Total War:Attila and it appears Ashes of the Singularity are very GPU heavy strategy games as well.

This is why it makes far more sense if you have a reasonably CPU already to upgrade the card first if its not that new,for most games.

Compare the increases in graphics cards performance from generation to generation.

GTX680 to GTX780TI to GTX980TI.

A GTX980TI is twice as fast as a GTX680,and that happened in three years.

CPUs did not get anywhere near that especially for those overclocking over the last FIVE years.

David Bisset · 14 Mar 2016 at 07:41

That's a fair point cat, my work laptop sometimes is the limiting factor when playing SC2 but my desktop has not yet been (not a great graphics card but I play on fairly low settings anyway), held back by opponents machine occasionally though. I'd say over time this should change - except more and more people game on laptops which are aiming for power efficiency not speed so it's not really improving matters.

mdixson · 14 Mar 2016 at 07:53

Reading through I dont think this article has been mentioned - certainly worth a read http://www.eurogamer.net/articles/d...it-finally-time-to-upgrade-your-core-i5-2500k

dakaiez · 15 Mar 2016 at 19:51

Definitely noticed a difference going form a [email protected] to a [email protected]. That along with a 980ti upgrade, can throw everything at it.

In the Division my CPU was hitting 100% so i couldn't stream, now i can play and stream and still have 40%+ headroom.

Hedge · 15 Mar 2016 at 21:27

dakaiez said:
Definitely noticed a difference going form a [email protected] to a [email protected]. That along with a 980ti upgrade, can throw everything at it.

In the Division my CPU was hitting 100% so i couldn't stream, now i can play and stream and still have 40%+ headroom.

Are you sure the 980ti upgrade isn't the most noticeable factor here?

dakaiez · 15 Mar 2016 at 21:56

Hedge said:
Are you sure the 980ti upgrade isn't the most noticeable factor here?

Tried it with and without the 980ti just to check. Very noticeable difference in terms of what i could handle and lower end fps. I never go above 60% on my processor now even when streaming both with the 680 and 980ti.

Before without streaming in The Division was hitting 100%, big big difference. That and encoding videos is much much quicker and system as a whole feels well worth the upgrade.

Ayahuasca · 15 Mar 2016 at 22:36

Yeah for streaming an i7 is going to help massively.