Upgrade from 2500k to 6700k

S

Starview

n00b
Joined
Jul 9, 2011
Messages
15
So after doing lots of research with most sources claiming that you wouldn't get much in the way of perceivable differences in a clock for clock comparison I went ahead and did it anyways.

I wish I had ran more benchmarks before I did the swap but at least I can post these two for anyone who might find this thread via google or something. I am seeing higher average frame rate in Witcher 3 also (wish it had a built in benchmark). Probably not worth the upgrade cost for most people but I kept reading that I would get almost zero improvement from the i5 @ 4.5GHz.

Unigine Heaven ran @ 1080, Tomb Raider (2013) ran @ 1440

i5 2500k @ 4.5 GHz
16GB DDR3 1600
2x 780 GTX SLI

vs

i7 6700k @ 4.5GHz
32GB DDR4 3200
2x 780 GTX SLI

In both captures the i5 is the 1st picture.

1080
TtrUQZl.jpg


1440
P6uwNd8.jpg


1440
714A256.jpg
 
Try doing it with a single 780, you should not have a difference at all imo. Looks like previous CPU bottleneck or PCIe bandwidth bottleneck?
 
It's possible I suppose, Z68 chipset is only PCIe 2.0 vs the Z170 at 3.0. Both would be running at 8x in SLI mode though.
 
You should compare it against 2600k. Clock for clock 6700k is faster but once you overclock 2600k higher it won't.

Also, single card, please.
 
Last edited:
With more powerful GPUs you'd see an even bigger difference.
 
a3venom said:
Try doing it with a single 780, you should not have a difference at all imo. Looks like previous CPU bottleneck or PCIe bandwidth bottleneck?
Click to expand...

no.. Skylake it's at least 30% faster than sandy bridge. and he's not only going from sandy bridge to skylake he also jumped from i5 to i7. so in several games he will see a massive jump in FPS there are actually a lot of games that benefit from the extra threads and architectural improvements (tri-gate transistor from Ivy bridge, 14 phase pipeline, AVX2 from haswell, and more important skylake features with improved front-end and out of order buffer executions[Improved and optimized hyper-threading], but also VALU[third vector integer ALU] which also have improvements beyond only IPC)

Iching said:
You should compare it against 2600k. Clock for clock 6700k is faster but once you overclock 2600k higher it won't.

Also, single card, please.
Click to expand...

you couldn't be any more wrong ;) even at stock clocks the i7 6700K it's simply faster than my old 3770K with my max 4.8ghz overclock in every task I could throw at it. and now OC'd to the same 4.8ghz its performing a good 30%(and more in some games) faster than my old chip.
 
Araxie said:
no.. Skylake it's at least 30% faster than sandy bridge. and he's not only going from sandy bridge to skylake he also jumped from i5 to i7. so in several games he will see a massive jump in FPS there are actually a lot of games that benefit from the extra threads and architectural improvements (tri-gate transistor from Ivy bridge, 14 phase pipeline, AVX2 from haswell, and more important skylake features with improved front-end and out of order buffer executions[Improved and optimized hyper-threading], but also VALU[third vector integer ALU] which also have improvements beyond only IPC)
.
Click to expand...

You are absolutely wrong, if i run this with a 780 on a 4th gen i3 vs a 6700k oced the results should be in 1 - 2 fps of each other because this is an extremely GPU limited benchmark. Thats why i was shocked to see such an improvement which has to be because he has 2 x 780s, which could result in CPU being the bottleneck in his case since he has 2x the GPU power or it could be PCIe thing ( probably not)

Quix said:
With more powerful GPUs you'd see an even bigger difference.
Click to expand...

this is right, and with less powerful your difference would cease to exist between those 2 CPUs (imagine if you do this with a 750ti)
 
a3venom said:
You are absolutely wrong, if i run this with a 780 on a 4th gen i3 vs a 6700k oced the results should be in 1 - 2 fps of each other because this is an extremely GPU limited benchmark. Thats why i was shocked to see such an improvement which has to be because he has 2 x 780s, which could result in CPU being the bottleneck in his case since he has 2x the GPU power or it could be PCIe thing ( probably not)



this is right, and with less powerful your difference would cease to exist between those 2 CPUs (imagine if you do this with a 750ti)
Click to expand...

This guy is hilarious. 30% faster. :p
 
a3venom said:
You are absolutely wrong, if i run this with a 780 on a 4th gen i3 vs a 6700k oced the results should be in 1 - 2 fps of each other because this is an extremely GPU limited benchmark. Thats why i was shocked to see such an improvement which has to be because he has 2 x 780s, which could result in CPU being the bottleneck in his case since he has 2x the GPU power or it could be PCIe thing ( probably not)



this is right, and with less powerful your difference would cease to exist between those 2 CPUs (imagine if you do this with a 750ti)
Click to expand...

"If I run"? so you can't test and see results yourself right? then you can't speak for your own experience and first hand how the CPUs perform, I really do not care about these benchmark posted, I talk about my own experience with single and couple of 980TI.. want to see your i3 destroyed? try Crysis 3, Watch Dogs, Hitman Absolution, GTA V, BF4 (multi), FO4, Dragon Age Inquisition, Shadow of Mordor, Project Cars, as few examples if you need more just tell me.. and you will see how you i3 are easily a massive bottleneck beyond 50% of what an i7 can offer. You know I most of the time and generally agree with you about the performance an tiny i3 can offer but being honest and having couple of those tiny chips I know they lack of enough performance in the mentioned games, I know there's an issue with tomb raider 2013 that game run equal with 2 or 8 threads so it shouldn't be a huge difference as shown in those pics, however in the games i've mentioned you can find differences even greater than what are posted here.
 
Araxie said:
"If I run"? so you can't test and see results yourself right? then you can't speak for your own experience and first hand how the CPUs perform, I really do not care about these benchmark posted, I talk about my own experience with single and couple of 980TI.. want to see your i3 destroyed? try Crysis 3, Watch Dogs, Hitman Absolution, GTA V, BF4 (multi), FO4, Dragon Age Inquisition, Shadow of Mordor, Project Cars, as few examples if you need more just tell me.. and you will see how you i3 are easily a massive bottleneck beyond 50% of what an i7 can offer. You know I most of the time and generally agree with you about the performance an tiny i3 can offer but being honest and having couple of those tiny chips I know they lack of enough performance in the mentioned games, I know there's an issue with tomb raider 2013 that game run equal with 2 or 8 threads so it shouldn't be a huge difference as shown in those pics, however in the games i've mentioned you can find differences even greater than what are posted here.
Click to expand...

Nobody cares, people are talking about those benches posted above, and they are extremely GPU limited, you can get a dual 32 core xeon or i3 it will be the same with a single 780. Admit you are wrong when you are wrong, dont act like a kid and change topics.
 
Araxie said:
if you have the money, buy one and test yourself =) you will then understand what im speaking.
Click to expand...

Why would I waste money on Skylake? Money = weak argument. I prefer spending on GTX 980 TIs than CPU that will do nothing for me.
 
When it comes to gaming GPU > CPU, and there have been a ton of professionals benchmarking older CPUs versus newer ones and they all had lackluster results. Most non-professional people that get large boosts in performance can probably be attributed to a clean/fresh OS installation.
 
I'd also like to see single card benchmarks just in case. Sandy Bridge still packs a lot of punch but PCI-E 2.0 is getting severely bottlenecked by the latest videocards especially when you SLI them and the problem is only going to get worse as time passes. That port was relevant when when GTX 480 was the hot shit and that was ages ago! (edit: no wait, I think it was the 280 that was the hot shit?) A single 980Ti could very well be bottlenecked by it slightly.
 
I'm actually pretty shocked at the performance increase. Was expecting 10% difference at most honestly, not 15%-20%.
 
These "I love my 2500K I will stick with it my whole life" cheap skids are everywhere on tech forums, even on reddit. Their collective denial IMO is not based on facts and logic, but on emotion. They can't afford to upgrade so they use myths like skylake has no improvement in games to make themselves feel better, to justify their behavior.
 
Iching said:
This guy is hilarious. 30% faster. :p
Click to expand...

Generation wise he isn't wrong. Actually he was being a little more conservative I think. I think without oc'ing even Intel says it averages 10% increase per gen give or take. And not just on one benchmark but a suite of benchmarks testing the whole platform.

The only thing I have here to test it was my spare i5-2500 Vs 3570k Vs 4590. In games the 4590 didn't all of a sudden get 20% more fps but it did consistently get better fps with the same card at 1080 and 1440p gaming.
 
elfdemon said:
These "I love my 2500K I will stick with it my whole life" cheap skids are everywhere on tech forums, even on reddit. Their collective denial IMO is not based on facts and logic, but on emotion. They can't afford to upgrade so they use myths like skylake has no improvement in games to make themselves feel better, to justify their behavior.
Click to expand...

I'm not one of those guys but I did give my buddy my old 2500k system with a 750ti as a gift and he thinks it's the second coming of Jesus. He wanted to pay for it and I told him the fact that I stopped him from getting an alienware r51 was payment enough. I estimate for general game play and usage at 1080p the 2500k will be pretty damn handy for a while.
 
OP, could you disable HT on the 6700K and redo the test? This would make a more interesting almost 'apples-to-apples' comparison.
 
Last edited:
exlink said:
I'm actually pretty shocked at the performance increase. Was expecting 10% difference at most honestly, not 15%-20%.
Click to expand...

http://www.overclock.net/t/1578480/i5-2500k-4-5ghz-vs-6700k-4-5ghz-in-games

This is about the most extensive testing I have seen of Sandy vs Skylake, HT on and off and single GTX 970. Max FPS did not rise much mostly but holy those minimum and average FPS! That is a pretty big deal on how smooth your gaming experience is. You know, maybe it really is time to retire my SB system. :(
 
Last edited:
elfdemon said:
These "I love my 2500K I will stick with it my whole life" cheap skids are everywhere on tech forums, even on reddit. Their collective denial IMO is not based on facts and logic, but on emotion. They can't afford to upgrade so they use myths like skylake has no improvement in games to make themselves feel better, to justify their behavior.
Click to expand...

For most people there is no need to upgrade.
The higher the framerate, the higher the need for CPU.
If you dont care about hugging 60fps and max quality settings, most will cope fine.

When you dont like cheap, its better not to take cheap shots at people.
Being correct helps as well :p
 
elfdemon said:
These "I love my 2500K I will stick with it my whole life" cheap skids are everywhere on tech forums, even on reddit. Their collective denial IMO is not based on facts and logic, but on emotion. They can't afford to upgrade so they use myths like skylake has no improvement in games to make themselves feel better, to justify their behavior.
Click to expand...

Go, go, go spend your money. Are you actually trying to say something?
 
Iching said:
Go, go, go spend your money. Are you actually trying to say something?
Click to expand...

Well since amd people aren't worth fighting anymore since they haven't been relevant in 5 years. It's time to feast among ourselves as Intel users. :)

Truth be told I have noticed a small pack of i-2500k is still just as good people. They find it hard to accept that it's just as good but it's still great.

I think 90% of people who still use over clocked 2500k are reasonable people who know they don't have the best computer but still happy with how it performs. The rare few who thinks that there's no improvement in Skylake though do need to stop kidding themselves.
 
Nehalem -> sb -> ib -> hw -> bw -> skylake, the bumps are adding up, plus ddr4 bandwidth. It just took forever. The old 2500k still got 97 avg fps in the one game tested ...
 
xorbe said:
Nehalem -> sb -> ib -> hw -> bw -> skylake, the bumps are adding up, plus ddr4 bandwidth. It just took forever. The old 2500k still got 97 avg fps in the one game tested ...
Click to expand...

This is exactly why I'm sticking with my 2500k: I'm happy with the minimum frame rate performance this year. Maybe next year I won't be, and I'll consider an i5 or i7 upgrade path. It won't last forever, but it will last a stupidly long time!

One thing I do have my eye on: mainstream platform is supposed to get an upgrade with Cannonlake!

http://hexus.net/tech/news/cpu/86879-intel-go-beyond-quad-core-cannonlake-processor-family/

And this makes sense since Broadwell-E is getting the bump to 10 core. Gotta make the mainstream platform a little more competitive!
 
So shouldn't the 2500k be OC to 4 so its the same speed as the 6700k?
 
There is a pretty big gaming difference with Skylake vs. Sandy Bridge (even Ivy Bridge). Saying other wise you are just fooling yourself.

OP thanks for the benchies, hope you like your upgrade! You'll see a pretty big difference in other well threaded games like BF.
 
Trimlock said:
There is a pretty big gaming difference with Skylake vs. Sandy Bridge (even Ivy Bridge). Saying other wise you are just fooling yourself.

OP thanks for the benchies, hope you like your upgrade! You'll see a pretty big difference in other well threaded games like BF.
Click to expand...

Oh, please stop. BF?o_O One can't win with internet warriors.
 
Last edited:
First off, I want to say thank you for writing such a comprehensive post comparing the two platforms. It is very much appreciated.

However, I am a bit perplexed as to your findings. Anandtech did a simililar review of Skylake vs Sandybridge (The Intel 6th Gen Skylake Review: Core i7-6700K and i5-6600K Tested), and did not notice any appreciable difference in games between the two platforms.

Anecdotally, comparing the two rigs in my sig (Westmere 1st generation i7 with Haswell, and there is very little difference between the two when gaming at 1440p - 3D Mark Firestrike GPU Score, Dragon Age Inquisition and Call of Duty Modern Warfare 3 all report the same performance in game. It was actually what motivated me to make the x5660 my primary gaming rig, as it was simply more fun to work with.

Interesting results none the less. Gaming improvements aside, I think that Skylake brings enough to the table to warrant an upgrade should your budget allow for it. However, all things being equal, if I had to spend my money between upgrading a first generation i7 to Skylake, or upgrading the GPU and sticking with the older platform, I would choose the latter.

Edit: I noticed Bittech did a similar evaluation but both platforms overclocked to 4.8ghz - with the exception of Alien Isolation (and in all fairness, they were exceeding 150fps anyways), Skylake and Sandybridge were the same
 
Last edited:
caltech31 said:
First off, I want to say thank you for writing such a comprehensive post comparing the two platforms. It is very much appreciated.

However, I am a bit perplexed as to your findings. Anandtech did a simililar review of Skylake vs Sandybridge (The Intel 6th Gen Skylake Review: Core i7-6700K and i5-6600K Tested), and did not notice any appreciable difference in games between the two platforms.

Anecdotally, comparing the two rigs in my sig (Westmere 1st generation i7 with Haswell, and there is very little difference between the two when gaming at 1440p - 3D Mark Firestrike GPU Score, Dragon Age Inquisition and Call of Duty Modern Warfare 3 all report the same performance in game. It was actually what motivated me to make the x5660 my primary gaming rig, as it was simply more fun to work with.

Interesting results none the less. Gaming improvements aside, I think that Skylake brings enough to the table to warrant an upgrade should your budget allow for it. However, all things being equal, if I had to spend my money between upgrading a first generation i7 to Skylake, or upgrading the GPU and sticking with the older platform, I would choose the latter.

Edit: I noticed Bittech did a similar evaluation but both platforms overclocked to 4.8ghz - with the exception of Alien Isolation (and in all fairness, they were exceeding 150fps anyways), Skylake and Sandybridge were the same
Click to expand...
Even his findings don't seem like a lot of difference. If you want a newer system it works out good.(selling your old system to get some back makes it a bit better)
There could even be memory difference between the speeds that will change it.
 
Thanks for the responses guys. As I said I re-purposed a lot of the 2500k build into the 6700k so I would need some parts to test it again. Once I get some free time I'll put the 2500k on a new psu and try with a single GPU.

And wow this guy linked above did a great job. He might already have all the benchmarks needed.

i5 2500k 4.5Ghz vs 6700k 4.5ghz in games
 
MaZa said:
This is about the most extensive testing I have seen of Sandy vs Skylake, HT on and off and single GTX 970. Max FPS did not rise much mostly but holy those minimum and average FPS! That is a pretty big deal on how smooth your gaming experience is. You know, maybe it really is time to retire my SB system. :(
Click to expand...

Sorry for double post but I just wanted to say that the frame minimums is where I was hoping to see improvement. All I am looking for is a smoother experience as you say. I know 2x 980ti would have given more performance than the CPU\MB\RAM upgrade but I'm hoping to hold out until pascal for GPUs.

For the guys looking for more comparisons these two videos linked in another thread were one of the reasons I made the jump. Test multiple games and you can see the frame drops with nice real time graphs. Might be interesting to some of you.

He does: 6700k vs 4790k vs 3770k vs 2600k

Overclocked:

WyPBIbr.jpg
 
Clock for clock i7-6700k seems to be not so impressive. That video should settle it.

 
Here is my Haswell 4790k Benchmark with two 980 ti in SLI at 4.8ghz.. Looks like skylake doesnt have much on me.
mIiAWZe.png

pCCH25z.png

YtvE0or.jpg
 
looks like to me the OP has a 20% jump in speed over IVY to Skylake. My benchmarks would show less than a 10% difference in cpu processing speed. At 1080p my min was 37.7FPS vs 38.4 on skylake at 1080p at 4.8ghz . IF the OP overclocked another 300mhz his min FPS would be around 38fps in heaven also.

gaming might be around the same percentage from sandy to skylake ... but overall system performance i would think going from sandy to skylake should net a 30% increase in other tasks , such as encoding, video editing etc.

If you own a 2600k or a ivy bridge i noticed a huge difference in dolphin emulation. I could barely run any game until i went to haswell. Skylake i7 has almost a 10 percent increase in performance with dolphin from haswell i7 .
 
Last edited:
You must log in or register to reply here.
Back
Top