i5 2500K vs i7 2600K, does cache size matter?

[Rangoon]

I am trying to choose between these two processors. Is the main difference between them the cache size between 6MB and 8MB? Will they both overclock to relatively the same degree? I would like to get a water cooling setup this time around. Will both of these chips likely OC to 4.6GHz easily on water? Higher?

I mainly play games, but also record those games. I do some video and audio work as well. I play most genres of games. Some are CPU limited and others are GPU limited on my current system (GTX 285, e8400 @ 3.8GHz). I would like to be able to play today's titles and into the next two years with as little trouble as possible (to record on max settings at 1920x1200).

What exactly does the cache do for me? Is it mainly beneficial when 4 threads are being used? Or will it help prevent slowing down in CPU-intensive games even with just 1 or 2 threads? Is it relevant?

Thanks!

 

[Achilles97]

Hmm. I don't have any useful information to share about the cache, etc, but I can say that my 2500k easily overclocked to 4.5 on a Mugen3 and a $115 mobo. NCIX pricematched Microcenter so i got my 2500k for $179 shipped and no tax. I just use thr computer for gaming. I am extremely pleased with the 2500k given what I paid and the usage I get out of it.

I am sure there will be other posts with more specific performance information to help you decide.

Good luck.

 

[Puppies04]

The really aren't going to be able to notice the difference between the i5 and i7 in 99% of what you do. The 2500k is a very capable CPU, you will hit 4.6ghz or higher, it all depends if you get a really top notch example or not as to how much voltage you have to put through it, also it comes down to how far you are willing to push temps which shouldn't be a problem on water.

I don't recall actually seing a benchmark done with HT disabled on the i7 to see what difference the extra 2mb of cache makes but I doubt it would make any discernable difference in the tasks you are talking about doing.

 

[nanaki333]

the HT in the 2600 actually takes a performance hit in some games.

 

[Puppies04]

the HT win the 2600 actually takes a performance hit in some games.

Also HT increases heat output so if you OC to a max temp and not a max vcore the I5 should end up with a slightly faster OC although it all depends if you manage to get a chip from a golden batch. I see people claiming 5ghz on 1.38v and I have had to set my 2500k to 1.41 to get stable at 4.7 although I'm still trying to tweak some other settings to lower that by a notch or 2

 

[OILFIELDTRASH]

Look at the anandtech review. The 2600k regularly gets a few more fps in the gaming tests even though it's basically running at the same speed. That has to be because of cache size difference.

 

[dma0991]

If you have the cash to but a Core i7 2600K, it is definitely the best choice but I'd be just as happy with a Core i5 2500K for the majority of the things that I do. If HT takes a hit during your gaming sessions, disable it in your BIOS. Either chips will OC to 4.6GHz easily on air, even more so with a decent WC setup. A Core i7 2600K is nothing more than a higher binned Core i5 2500K.

 

[TakeNoPrisoners]

Cache size does matter. If you disable the HT and run it at the same speeds as the i5 2500k, you will get slightly better performance, it won't be much better though.

The i7 2600k also tends to overclock better.

 

[BD231]

12mb nehalem had *zero* performance increase over the 6mb versions, so I highly doubt you'd notice the extra 2mbs. Right now it's only there for the added demands of HT.

 

[TakeNoPrisoners]

12mb nehalem had *zero* performance increase over the 6mb versions, so I highly doubt you'd notice the extra 2mbs. Right now it's only there for the added demands of HT.

SB-E shows a performance boost from extra cache.

http://www.anandtech.com/show/5091/...-bridge-e-review-keeping-the-high-end-alive/6

"The larger cache helps give the 3960X a 9% advantage over the 2600K in Dawn of War II."

 

[Rangoon]

Look at the anandtech review. The 2600k regularly gets a few more fps in the gaming tests even though it's basically running at the same speed. That has to be because of cache size difference.

Couldn't that also simply be the 0.1 GHz difference in clock speed? If they were both clocked to 4.6GHz, would there be a difference? I will look for that; surely someone has a direct comparison somewhere...?

 

[OILFIELDTRASH]

Couldn't that also simply be the 0.1 GHz difference in clock speed? If they were both clocked to 4.6GHz, would there be a difference? I will look for that; surely someone has a direct comparison somewhere...?

2500k clocked at 3.33, 2600k 3.4ghz

No way you can get that many more fps just from that difference. That would mean at 4ghz you'd be getting 15 more fps.

 

[Puppies04]

A Core i7 2600K is nothing more than a higher binned Core i5 2500K.

I would love to see your proof for this when I have seen 2600k's needing 1.45v to hit 4.8 and 2500k's needing less than 1.4v.

Pretty weird binning process if you ask me, as far as I am concerned a 2500k is just a 2600k with the HT and 2mb of cache disabled so intel can make a bit more profit on the I7

 

[dma0991]

I would love to see your proof for this when I have seen 2600k's needing 1.45v to hit 4.8 and 2500k's needing less than 1.4v.

Pretty weird binning process if you ask me, as far as I am concerned a 2500k is just a 2600k with the HT and 2mb of cache disabled so intel can make a bit more profit on the I7

I said higher binning, not higher quality. You would also have to consider the fact that the Core i7 2600K may need the extra volts to go with HT enabled. You could probably see a smaller disparity in voltage if HT is disabled on the Core i7 2600K.

 

[Rangoon]

...I have seen 2600k's needing 1.45v to hit 4.8 and 2500k's needing less than 1.4v.

Hm, so the 2500K may be better/easier overclocking? So it could wind up being a better performer based on that. Is it just that much leaner with the HT and larger cache disabled? Therefore less voltage required and more potential for speed?

However it also seems more likely that over the next couple of years games will use multiple cores more routinely, and therefore the more robust architecture of the i7 may preform better in those conditions?

 

[Puppies04]

SB-E shows a performance boost from extra cache.

http://www.anandtech.com/show/5091/...-bridge-e-review-keeping-the-high-end-alive/6

"The larger cache helps give the 3960X a 9% advantage over the 2600K in Dawn of War II."

I would love to know how he can say that the performance boost is from the extra cache and not from the quad channel memory/more optimised mobo socket design/extra cores etc etc etc.

Check out the 2500k @ 3.3ghz ---41fps
2600k @ 3.4ghz ---41.1fps

... yeah extra cache really helps a lot I would expect more than 0.1fps difference from the 100mhz difference in favour of the I7.

The differences you see between the I5 and I7 in those tests are easily explained by different clock speeds/margin of error and even if we are talking about a 2-4 fps increase with an I7 can you really pin that on the cache size and completly ignore HT?

 

[OILFIELDTRASH]

Hm, so the 2500K may be better/easier overclocking? So it could wind up being a better performer based on that. Is it just that much leaner with the HT and larger cache disabled? Therefore less voltage required and more potential for speed?

However it also seems more likely that over the next couple of years games will use multiple cores more routinely, and therefore the more robust architecture of the i7 may preform better in those conditions?

It's just a luck of the draw type deal IMO. I don't believe the 2600k is any less capable or anymore capable of overclocking than the 2500k. Really just depends on if you get a good chip.

Neither one is going to bottleneck any gpu you get. Maybe 3 gtx 580's.

 

[TakeNoPrisoners]

I would love to know how he can say that the performance boost is from the extra cache and not from the quad channel memory/more optimised mobo socket design/extra cores etc etc etc.

Check out the 2500k @ 3.3ghz ---41fps
2600k @ 3.4ghz ---41.1fps

... yeah extra cache really helps a lot I would expect more than 0.1fps difference from the 100mhz difference in favour of the I7.

The differences you see between the I5 and I7 in those tests are easily explained by different clock speeds/margin of error and even if we are talking about a 2-4 fps increase with an I7 can you really pin that on the cache size and completly ignore HT?

Games really arn't memory limited. Quad channel memory is overkill and I do not think dual channel memory is a bottleneck. The extra cores make a difference because they can spread out windows processes on those cores.

I do think cache makes a difference depending on the game, WoW really likes extra cache.

 

[Puppies04]

Games really arn't memory limited. Quad channel memory is overkill and I do not think dual channel memory is a bottleneck. The extra cores make a difference because they can spread out windows processes on those cores.

I do think cache makes a difference depending on the game, WoW really likes extra cache.

Think what you like but until i see the 2600k getting 20-30fps better I don't believe the extra cache makes 1 jot of difference in gaming. the I7 has 33% more cache compared to the I5. Where are the increased framerates to show that and im not talking 1-5fps

 

[OILFIELDTRASH]

Think what you like but until i see the 2600k getting 20-30fps better I don't believe the extra cache makes 1 jot of difference in gaming. the I7 has 33% more cache compared to the I5. Where are the increased framerates to show that and im not talking 1-5fps

The question in OP was whether cache makes a difference. Yes it does. Is it worth the extra $100 or so? That's up to the OP to decide.

 

[LiuKangBakinPie]

SB-E shows a performance boost from extra cache.

http://www.anandtech.com/show/5091/...-bridge-e-review-keeping-the-high-end-alive/6

"The larger cache helps give the 3960X a 9% advantage over the 2600K in Dawn of War II."

Yet each core are only restricted to 2.5mb cache on the E versions. Less cores active means 2.5mb less cache per core

In gaming the cache wont make a difference like a better gpu would

 

[BladeVenom]

For gaming get the 2500k.

 

[IntelUser2000]

Yet each core are only restricted to 2.5mb cache on the E versions. Less cores active means 2.5mb less cache per core

In gaming the cache wont make a difference like a better gpu would

There's nothing else that suggests that's true and each core has to be bound by only 2.5MB cache.

When proving a theory that's widely different from general beliefs, proof must be provided.

 

[sm625]

I started typing a post about how I could never find ANY benchmark comparing a 2500k and a 2600k with HT off and running the same speeds. Then I tried yet another google search. After combing through 19 results , my stupid piece of iCrap safari joke crashed and my post was gone! I was dealing with this crap on windows 10 years ago. But you know what? It don't happen anymore on windows. Anyway, amazingly enough there doesn't seem to be any results but I gave up looking because it's a Saturday night and this pos ain't no computer.

 

[Rangoon]

For gaming get the 2500k.

Is that because of no HT, it's cheaper, and can overclock just as well as the 2600K?

What about a 2600K, with HT turned off, and overclocked to 4.6? Great idea just $100 more?

If it really is that simple, then the 2500 does sound like the better option. And this is the same for recording games? And the same for basic video/audio editing? And having a small basic array of other tasks running in the background like Ventrillo/TeamSpeak?