Intel Core i7-875K And Core i5-655K Battle Beyond 4 GHz

[SpadeM]

[citation][nom]cangelini[/nom]Throw some suggestions out there! I'm always watching the comments for feedback iTunes is but one element of the tests, and to be fair, it's a nice change of pace from all of the threaded transcoding apps that demonstrate the same thing, for the most part ;-)[/citation]
Glad someone is listening
Well my suggestions are as followed:
1. For the Audio department i'd go with something that can convert cd's to flac or flac to mp3 aif aac and the works, maybe run it as a batch process. (I personally am using goldwave for that, plus it adds filters and distorsions)
2. On the video side I'd go with avidemux or virtualdub and convert into x264 with mkv/mp4 package since it's becoming the norm these days (take anime and compressed HD video)
3. I like the 3dsmax since I use it, but spicing it up with some CAD, maybe some solidworks or why not even the good ol Blender since it's free. And if it's possible some other photo suite, be it from Corel or whoever.
4. Definitely some office/open office/flash tests, content creation as a batch process and time it from start to finish.
5. Games .. definitely no console ports or games that were made with consoles in mind. Crysis Warhead, Alien vs Predator, Metro, some racing simulator, a strategy game, basically games that have either a wide install base OR are using new engines and technologies.
6. Anti-virus programs, you can always use the free editions of avira, avg, mse, and so on and so forth.

I know that changing a benchmark suite takes time and $ but, I believe it could be more fair and close to a real life usage scenario.

 

[masterofevil22]

[citation][nom]youssef 2010[/nom]I wonder if they disabling 2 cores of the 6 on the P II x6 will help improve gaming performance.[/citation]


Honestly, I think the Phenom II X6 scores lower than the Phenom II X4 because of the more "advanced" power management.

I've noticed on mine (X6) that it will throttle down to very low Mhz ~1150 even when running at 4Ghz. I imagine disabling this aggressive power management would allow for speedier game scores since most are only running on a fraction of it's cores and it would not have a lag while "throttling up" in Mhz.

 

[Tonkyboy]

I'm still running my old C2D E6300 (stock 1.86 Ghz) at 3.33 Ghz.

That's a 1.47 Ghz overclock on a 1.86 Ghz chip. On air. With a £15 cooler. On an Asus P5K mobo. For well over 2 years.

My next rig will be AM3, Phenom II X4 965 BE. But no cash for the project yet

 

[gkay09]

Well atleast this is a good start on Intel's side IMO...Like they said in the article, this should expand to the rest of the chips also...

 

[zodiacfml]

same old price structure by Intel, quite perfect though so as not to affect the value other chips.
anyways, new parts won't sway me from buying the AMD black editions. there's not much performance difference.
the savings will get me a reasonable a heatsink.

 

[liquidsnake718]

Intel needs to have a k series for the 1366 boards!



Not to say Id get it since I already am building around a i7930, which I just got this past weekend.

 

[silverblue]

[citation][nom]masterofevil22[/nom]Honestly, I think the Phenom II X6 scores lower than the Phenom II X4 because of the more "advanced" power management.I've noticed on mine (X6) that it will throttle down to very low Mhz ~1150 even when running at 4Ghz. I imagine disabling this aggressive power management would allow for speedier game scores since most are only running on a fraction of it's cores and it would not have a lag while "throttling up" in Mhz.[/citation]

I think the X6 returns to the original Phenom's independent core clocking system, albeit with much quicker ramp and slam delays (I think those are the terms?). Thus, it's very elegant and results in a far smaller performance drop than we saw back with Barcelona. If you throw something relatively heavy at the Phenom II X6, you wouldn't ever notice a difference between having all cores locked at the same frequency and all cores working independently anyway.

 

[juliom]

[citation][nom]silverblue[/nom]I think the X6 returns to the original Phenom's independent core clocking system, albeit with much quicker ramp and slam delays (I think those are the terms?). Thus, it's very elegant and results in a far smaller performance drop than we saw back with Barcelona. If you throw something relatively heavy at the Phenom II X6, you wouldn't ever notice a difference between having all cores locked at the same frequency and all cores working independently anyway.[/citation]

The independent clocks for each core on Barcelona was a great idea. The only problem was the crappy thread scheduler that Vista had. It made the threads bounce cores all the time, thus the cores were constantly slowing down and ramping up. Win 7 is a bit better on that regard, but I imagine not by much. Anyway, the new X6 only uses it for Turbo and not idle clocks, I think.

 

[wh3resmycar]

oh the pain of having your epeen cut in half.

 

[silverblue]

In MW2, the Thuban is the only CPU that seems to increase speed as you go up in resolution... and, with AA on, is the best performer at 2560x1600.

As for the strange result in Crysis, I doubt it's due to non-power-of-2, otherwise we'd have heard about it by now with the X3s. Was this repeatable? Could it be down to the aforementioned change in power management, and if so, why now and not in other resolutions/with other settings?

 

[JeanLuc]

[citation][nom]SpadeM[/nom]2. On the video side I'd go with avidemux or virtualdub and convert into x264 with mkv/mp4 package since it's becoming the norm these days (take anime and compressed HD video)[/citation]

I don't know about Virtualdub becoming the norm, it's been around for years now and most people I speak to who do a lot of video encoding use Handbrake.

Chris you should really lose the DVIX & XVID benchmark, these avi containers are limited to two threads at the most and it's been that way for years now there not going to change which makes them pointless for testing these highly threaded/CPU's.

 

[2shea]

it's obvious that intel doesn't unlock the i5 750 and 930 for precise the reason stated: they are way to popular to upgrade to unlocked versions that will effectively cancel all orders for the normal ones prolly. Although very good oc, the i5 655 is a dual core and in these times I want a triple core minimum for multitasking and rarring, videoeditting and such. Pure gamers could use a fast dual core though.
good article and very nice oc's on all cpu's. The six core still disapoints me a bit in performance, I would have expected more power from the two extra cores. Now it only matters in the specific multicore/thread optimized programs which are not that abundant.

 

[SpadeM]

[citation][nom]JeanLuc[/nom]I don't know about Virtualdub becoming the norm, it's been around for years now and most people I speak to who do a lot of video encoding use Handbrake. Chris you should really lose the DVIX & XVID benchmark, these avi containers are limited to two threads at the most and it's been that way for years now there not going to change which makes them pointless for testing these highly threaded/CPU's.[/citation]

I was refering to x264/mkv/mp4 becoming the norm not the software you encode with. So many free encoders out there.. take your pick, but anything would be better then what's now available in the benchmark section

 

[dertechie]

Eh. You can overclock the older models fine without it, this just makes it easier. Still probably looking at an i5-750, taken to 3.2 with Turbo still on to hit 3.84. Not really interested in what you yourself described as suicide run OCs.

 

[saint19]

Nice review man, now I have wait to see the price of a full build with the 785k and maybe I build another rig with an Intel CPU to add to my X4 955 @ 4.0GHz.

 

[jimishtar]

glad to see intel playing ball with us like I often put it: the empire always strikes back. I cant wait to see how AMD will respond to this new chips.

 

[Reynod]

Really enjoyed this.

Very well written Chris.

 

[cknobman]

Great article. Sad to see how much AMD current processors really lag behind Intel. But even though AMD processors get owned easily when comparing core to core all my computers are still AMD. Intel really shoots itself in the foot with overall platform costs. Knowing how much cheaper and more full featured the AMD platform keeps me coming back every time. Being able to buy an affordable motherboard with plenty of Sata ports, USB 3.0, integrated high def sound, ext. Sata, non gimped pci express lanes, etc..., etc... is really a blessing for the builder on a budget.

 

[sublifer]

Great article Chris! Keep up the good work.

Totally agree with you about their choice of models for the K series. Best guess is that is probably the sku they had the most of and felt comfortable in potentially wasting should something have gone wrong.

 

[Guest]

I was personally looking forward to the new anti-virus benchmark. There was mention of doing away with AVG and moving onto Kaspersky, but I haven't seen the change. I was interested due to me having to run a number of virus sweeps for family and friends.

 

[JohnnyLucky]

Thanks for including a few mainstream benchmarks, especially Adobe photo processes. Much appreciated.

 

[triculious]

interesting pieces of hardware Intel brought up

+1 on the choice of models for this k series but it makes me wonder why core i7 920 and core i5 750 weren't used to compare against on the Intel front

 

[shin0bi272]

[citation][nom]cangelini[/nom]Throw some suggestions out there! I'm always watching the comments for feedback iTunes is but one element of the tests, and to be fair, it's a nice change of pace from all of the threaded transcoding apps that demonstrate the same thing, for the most part ;-)[/citation]

dBpoweramp batch converter can run several different audio file type conversions (flac and mp3 to name a couple) and it will convert a different song per core (does NOT work for hyperthreaded cpus with multiple cores though).

 

[falchard]

Good review. My only disagreement is on the game benchmark suite. You choose a bunch of video card taxing games. You should have chosen CPU taxing games like RTS (Total War), or MMORPG.

My next 3D app processor is most likely going to be the successor to Magni-cours using a G34 platform.

 

[dragonmathics]

Chris,

I loved the article. Solid. And, since you asked for them, here are my suggestions. I'm an avid follower, and I love the hardware, but I don't get to build or game as much as I'd like, so here's what someone like me would like to see.

1. Maybe you have this on-site somewhere, but I couldn't find it. How about a permanent link on your main page that explains in detail your standard benchmark suite, what each benchmark is used for, and what it is meant to test for (For example, why run the iTunes bench if you know the cpu with the fastest clock speed will win? Are you looking to confirm the speed of the cpus, or are you looking for core advancements that might give you unexpected results?)? This way you could explain specific findings in the comparison article without having to re-explain what the benchmark specifically tests.

2. I agree that MS Office/Open Office benchmarks would be a great idea for home PC builders.

3. Please use a greater variety of games in the benchmark suite. Yes, you obviously want the most crushing, system-defeating games installed to test how potent the hardware is, but how about an MMO (incredibly popular, a useful test) or cpu-intensive RTS game thrown in to see how the cpu load affects system performance?

4. I realize you avoid this kind of thing to keep the charts less confusing, but this article was missing a comparison to a "normal" chip. Yes, these scores and times look great, but how do they stack up against a stock i7-920? Or one that's been overclocked? We have no basis for comparison, since such a cpu might never have seen this particular test rig before.

Love the articles. Please, keep 'em coming.