Intel Targets AMD With Affordable Unlocked CPUs

EconolineCrush writes "For years, AMD has catered to gamers and enthusiasts with mid-range Black Edition processors whose unlocked multipliers make overclocking easy. Intel has traditionally reserved unlocked multipliers for its ultra-expensive Extreme CPUs, but it has now brought the feature to affordable models that compete directly with AMD's most popular processors. The Core i5-655K and Core i7-875K have two and four cores, respectively, and they're priced at just $216 and $342. It appears that both will easily hit speeds in excess of 4GHz with air cooling. Surprisingly, even at stock speeds, the i7-875K offers better performance and power efficiency per dollar than just about any other desktop CPU out there."

Comments

That's "frequency", not speed

[Anonymous Coward]

Really, you call yourself a news for nerds site.

Re:That's "frequency", not speed

[Anonymous Coward]

What do you expect? It's a kdawson submission. You know: the submissions that always plain incorrect or just total garbage

Re:That's "frequency", not speed

[poetmatt]

for once, quite accurate by the anon. Reviews about these have been inconsistent, some citing bad overclocking potential and generally being not for enthusiasts.

Meanwhile, others seem to state it's a full sweep and/or basically great .

I'm wondering if this is another scenario of handpicked engineering samples or not.

I'm not at all convinced that this is great, or horrible. Anyone care to weigh in with better comments than kdawson?

Re:That's "frequency", not speed

[fuzzyfuzzyfungus]

In any overclocking scenario results aren't assured(though, at certain historical moments, they very nearly have been for certain chips). It isn't a huge surprise that there is some variability being seen; but the small sample size(maybe a dozen review sites, with a chip or two each) doesn't let us say too much).

The only thing that would be really sleazy would be if the review processors "just happened" to perform atypically well compared to the ones that poor saps can actually buy. Since, though, the mixed reviews are coming from reviewers, that seems less likely, and that these chips are simply spotty overclockers more likely(unless, of course, some of the reviewers are reviewing "representative samples" kindly provided by Intel, and others are reviewing representative samples scored from somewhere in the distribution chain.

The fact that a chip only sometimes outperforms its sticker speed is irksome for the overclocker; but not a big deal. If Intel is feeding handpicks to the reviewers, that sucks.

Re:That's "frequency", not speed

[poetmatt]

Well, thats why I want to know. It's not at all unknown for them to handpick the samples. I remember some recent controversy about this with somebody, I forgot if it was Nvidia, AMD, or Intel, or all of them in general.

The sites with the great results seem to say "this kick ass", and the ones without are meh, so it doesn't seem to be indicative of whether this processor is even worth it or not, all things taken into consideration.

Re:That's "frequency", not speed

[noidentity]

And speed is simply the frequency at which something is performed. Who says this 4 GHz speed is clock rate, anyway? When I read it, I figured it was how many Libraries of Congress (LOC) it could process in a fortnight. Maybe it's just me.

Re:That's "frequency", not speed

[noidentity]

Errr, I mean LOC per second, as the latter part IS specified by Hz (cycles per second).

Re:That's "frequency", not speed

[geekoid]

Since the wavelength didn't change, that means the increase in frequency could change with a change in speed.

If it was do to a decrease in wavelength(lambda), then you would have had a point.

Re:That's "frequency", not speed

[Runaway1956]

Even with highly overclockable AMD's, you have to search for the production code, then determine which week it was produced. There may be one single week, or a run of two to four weeks, but you have to narrow it down to one of those good periods. One week, you get a near lemon that won't overclock 5% - the following week, you get an "OH MY GOD THAT'S FUCKING FAST! CHECK THE TEMPERATURE!" chip. Even among the known good overclockers, stability becomes an issue. Week 32 will run super fast, but will only stay up and stable for a week or so at a time, while week 47 will run the same speed forever.

Now, I'm pulling numbers out of my ass, but you get the idea here. If you're serious about overclocking, you have to hit the forums, and find out what other people are doing. Then, when you've identified a chip that does what you want, you start searching for it. But, be prepared to pay a premium.

Personally, I shop for a fairly good overclocker - I have two Opterons that are known to run at ~3GHZ and be stable. But, they are rated at 2 Ghz and 2.2 Ghz. I bumped them up a conservative 5% using the soft-menu or EZ-menu or whatever in the motherboard BIOS. Why, you ask? Well, I figure that I've got a good, super quality chip known to run for years under extreme abuse. They should last ME damned near forever!!

Well - maybe not that long, but at least until the next ice age. ;^)

Re:That's "frequency", not speed

[Skal Tura]

the review cpus "just happened" to work better than off-the-shelf cpus in the past too. This happened with last gen 40nm extreme cpus, or actually all quad 40nm. All review sites gave glorifying reviews running with nForce chipset and SLI, or tri-sli. We had 3 to test, 2 extremes, 1 regular (2nd fastest regular), none of them worked well with any nForce chipset if at all, and and SLI was pretty much no go without underclocking. 3 different motherboards, 3 different models were tried aswell.

CPUs were Q9550 and the relevant same series both extreme CPUs. This is widely known problem, admitted my mobo manufacturers to a degree, but never been to news.

Re:That's "frequency", not speed

[Hurricane78]

As they say: Those who don’t learn from history...

Intel has a history if anticompetitive behavior, tricks and fraud. If it were a person, it would be a large sea man with a tattoo and bad company in a biker bar, and have at least half a dozen jail tattoos.

I’m not saying AMD is an angel. Not even remoteld. But it’s always good to look back at history, imagine that company as a person, and think about, if you would even talk to that person...

Re:That's "frequency", not speed

[Anonymous Coward]

this is exactly the case.

my laptop with a "budget" c2d processor is good enough for anyone. Plays any games I want, it's great. Only improvement would be an SSD

my Destop is a top of the line i7 OC'd to 3.6ghz with 8gb of the fastest ram, SSD, and 2 of the fastest GPu's money can buy in the consumer realm.

Why? It's my hobby. I enjoy gaming. I rip my large bluray collection. with my i7, I can encode 3 movies a day to h264/mkv while sleeping/at work. my laptop takes 19 hours a movie.

Re:That's "frequency", not speed

[cfrankb1]

You got it wrong ;) The real question is: who would you rather have a beer with?

Re:That's "frequency", not speed

[Mashiki]

Considering overclocking at the best is always done by handpicking the CPU manufacture period it wouldn't surprise me. OC enthusiasts will always go for a chip that's got the highest stability, and highest OC rate; and there's no shortage of people who will blow the money trying to be first. I used to overclock my machines heavily because I could, this was when I was younger but the fun thing was back during the slocket days. Taking that little Cleron 266, 300 or 400, and because it was multiplier locked you simply upped the bus speed from 66 to 100 or 100 to 133, and then tinkered around a bit to get your PCI bus in sync. It was nice to be able to double your clock speed(300>~600mhz).

Re:That's "frequency", not speed

[Vigile]

PCPer review seems to like them a lot too: http://www.pcper.com/article.php?aid=924

Re:That's "frequency", not speed

[hairyfeet]

I just don't see their prices being that great. I mean we are talking $342 and $216 when I can get an AMD 6 Core for $199 and an OEM AMD Black Quad for just $118. Compare that to the Intel 6 core at $1000!

Sure if you have to have the absolute fastest silicon on the planet Intel is the one to get (or if you just HAVE to have the biggest ePeen) but unfortunately Intel knows that too and will make you pay big time. The flip side of course is that Intel has the absolute shittiest IGPs out there, and by cutting Nvidia off at the knees have made damned sure it's gonna stay that way.

So these new chips with their very uncertain OC really don't impress me. One really needs to look at the cost of the total solution when building a new machine, and right now I'd say the bang for the buck is firmly in the AMD camp for anything less than $1000.I have been building a lot of AMD duals and quads for customers and thanks to the excellent Radeon IGPs providing hardware acceleration of all the major formats even the low end duals and quads multitask beautifully while playing HD smooth as butter.

So I don't really know who these chips are for-OCers would be better off buying the high end Intel chips if they are going for max speed, while those with a budget would be a hell of a lot better off going AMD for the more affordable boards with better IGPs. And if all you care about is OCing on the cheap, you can get an OEM AMD Black Dual for $54 or an OEM AMD Black Triple for $89, and with the right board you can even unlock the missing cores. The dual would probably be the best for aggressive OCing thanks to the larger die allowing for more cooling.

So as a former lifelong Intel man I just don't get who these chips are aimed at. Plus with many parts of the country still reeling from a shitty economy you really need to look at the price of the total build, and in that case AMD wins pretty much any sub $1000 build IMHO. The much better IGPs and much more affordable high quality boards tip it over into the AMD camp. IMHO the Intel board situation is just still too high for quality boards.

Re:That's "frequency", not speed

[jvin248]

What both Intel and AMD have been lacking is an identifyable method of comparing "Speed" for any normal and performance buyer.

Look in most Sunday sales fliers and the retail outlets are forced to spend more space advertising storage size, screen inches, and wifi.

In the olden days you could easily tell that a Pentium2@450Mhz would be better than a Pentium1@100Mhz.

Now it's a tad confusing with all the sempron, celeron, i-this, and i-that. So people buy laptop screen size and hard drive space. They can have no lust for or bragging rights for "I'd spend another $100 to get 10% more Ghz!"

Sad days.

Re:That's "frequency", not speed

[kevinmenzel]

6-core with hyperthreading on each core. Which, as someone who does a lot of multi-tasking, audio work with software that has great HT support, and MP3 transcoding... is actually quite a performance boost. Or at least it was on my 840EE which was worth the approx $1000 I paid for it 5 years ago, considering I haven't run into any performance problems in 5 years, and have yet to feel the need to upgrade. Of course that's a specific use case... so... yeah.

Re:That's "frequency", not speed

[jvin248]

Does anyone know if either, both, or neither Intel and AMD produce chips and mating motherboards today that protect the CPU from overheating and burning out?

I'm sure the OC's will know!

A string of chips and systems through the late 1990's and early to mid 2000's I've had more direct experience with (I do repairs and upgrades) showed that the Intels could run until overheated safety circuits would shut them down. The AMDs would keep going until they audibly 'blinked' and maybe magic smoke comes out, never to do another calculation. The Intels when cooled down would restart.



o

Re:That's "frequency", not speed

[gumbi west]

it's really the reviewers that you should hold responsible for this. They need to be able to tell you how they got the chip. What else do they have but their reputation?

Re:That's "frequency", not speed

[lmnfrs]

I'm not at all convinced that this is great, or horrible. Anyone care to weigh in with better comments than kdawson?

It's not great. "Unlocking" is the removal of an arbitrary "feature"-type addition Intel came up with to disallow the price/performance leniency allowed by changing the multiplier. They came up with it when overclocking became too popular. Using a lower multiplier and higher bus than the defaults used to be great. Running the CPU at its max may be pointless if a higher bus allows the communication between CPU and graphics card to increase.

It is nice they're removing their removal of a powerful option.

Re:That's "frequency", not speed

[hairyfeet]

While I agree with your point completely, how many folks have a real need to pound the silicon THAT hard? I build new PCs for a living and I can say that even with the ultra low end AMD duals a good 70% of the time the things are running at 800Mhz, simply because average folks don't have enough heavy lifting to keep the cores fed. I like to game and edit A/V and even with that kind of hardcore lifting probably 70% of the time my AMD quad is idle.

But the fact that they are idle IS the reason why I've been upselling many of my customers to the new AMD quads. With the new quads having 4 cores idling at 800MHz means they can do all their basic tasks, surfing, email, listening to music, and have C&Q keep their machine whisper quiet while giving them a smooth experience. And of course by having a quad when they DO have heavy lifting it gets done MUCH quicker, while still leaving the machine responsive. And with the new Radeons having hardware transcoding built in tasks such as A/V editing are very smooth, much smoother than those shitty Intel IGPs. And at just $99 for a 2.9GHz quad, it really is an easy sell.

So while I have NO doubt there are folks out there where the $1000 for the new Intel quads would be money well spent, hell some folks might even need a dual socket 12 core, for most folks that additional $800 would be much better spent on bigger HDDs, an SSD boot drive, plenty of RAM, etc. And the fact that I built my AMD Quad, with a 925 CPU, 8Gb of DDR2 800, 2 500Gb HDDs, dual burners, an HD4650 1Gb, and W7 HP x64, all for less than the cost of a SINGLE Intel 6 core, is just frankly an insane deal IMHO. For all the jobs I can think to throw at it the AMD kicks butt and doesn't heat up my apt.

Re:That's "frequency", not speed

[Anonymous Coward]

Even with highly overclockable AMD's, you have to search for the production code, then determine which week it was produced. There may be one single week, or a run of two to four weeks, but you have to narrow it down to one of those good periods. One week, you get a near lemon that won't overclock 5% - the following week, you get an "OH MY GOD THAT'S FUCKING FAST! CHECK THE TEMPERATURE!" chip. Even among the known good overclockers, stability becomes an issue. Week 32 will run super fast, but will only stay up and stable for a week or so at a time, while week 47 will run the same speed forever.

Now, I'm pulling numbers out of my ass, but you get the idea here. If you're serious about overclocking, you have to hit the forums, and find out what other people are doing. Then, when you've identified a chip that does what you want, you start searching for it. But, be prepared to pay a premium.

Personally, I shop for a fairly good overclocker - I have two Opterons that are known to run at ~3GHZ and be stable. But, they are rated at 2 Ghz and 2.2 Ghz. I bumped them up a conservative 5% using the soft-menu or EZ-menu or whatever in the motherboard BIOS. Why, you ask? Well, I figure that I've got a good, super quality chip known to run for years under extreme abuse. They should last ME damned near forever!!

Well - maybe not that long, but at least until the next ice age. ;^)

So, you're saying the overclocking fanboy sites produce statistically valid results?

igive up

[Anonymous Coward]

whi has the world one icrazy? For goodness sake there are 26 other letters to choose from.

It seems to me that for too long naming conventions follow the turbo marketing rule and bear little significance to the product. icore... please.

But i suppose this is good news for people that brought a CPU that doesn't allow you to tweak the multipler. I understand that this makes it much easier to clock and keep a reliable system. But for most people this is not an issue. Maybe it shows Intel are realising how much they are loosing to the gamer / media center PC builder, that works on a budget (would have selected the AMD over the Intel for price vs speed).

I just use a dual core Intel that's not clocked, and it's fine for what I need. If anything it's the RAM that I need to upgrade (to 8Gb).

Re:igive up

[jaymz666]

26 other letters, or maybe only 25 other letters?

Re:igive up

[lyml]

a b c d e f g h j k l m n o p q r s t u v w x y z å ä ö

I count 28 of them on my keyboard.

Re:igive up

[eyrieowl]

GP speaks estonian?

Re:igive up

[Anonymous Coward]

Certainly doesn't speak English very well.

Re:igive up

[Anonymous Coward]

Don't forget ß and ü. And then there are all the funny french variations of about everyone of them.

Re:igive up

[aix tom]

The real funny thing is, it makes more sense for Intel to go i-crazy than it does for for Apple.

Re:igive up

[RightSaidFred99]

Umm, in their defense the word "Intel" does start with exactly one of the 26 characters in the alphabet, and it happens to be 'i'. Unlike some other companies...

Re:igive up

[confused one]

My keyboard only has 26... I feel robbed.

Re:igive up

[LinuxIsGarbage]

And they did it long before apple: http://www.weirdstuff.com/mas_assets/thumb/13034.jpg

4 GHz, eh?

[Pojut]

I thought we learned that, like sex and the Pentium 4, faster isn't always necessarily better...

Re:4 GHz, eh?

[Anonymous Coward]

I thought we learned that, like sex and the Pentium 4, faster isn't always necessarily better...

Well, that depends... have you seen some of the women around here?

Re:4 GHz, eh?

[fuzzyfuzzyfungus]

When making a direct comparison between chip X running at frequency Y and chip X running at frequency Z, speed does correlate pretty closely to frequency(unless you are being bottlenecked hard by some other aspect of the system). Assuming you don't care about puny things like "deafening fan noise" and "having to throw a baby seal into the boiler every 15 minutes just to keep the lights from flickering"(and, if you are a true overclocker, you care nothing for such trifles) faster is better.

Re:4 GHz, eh?

[ch-chuck]

sex and the Pentium 4

Is that the new TV show for geeks?

Re:4 GHz, eh?

[TheKidWho]

Yes, faster is always better.

What we learned is that high GHZ don't necessarily imply a chip is faster.

Re:4 GHz, eh?

[master811]

Have you seen ANY women around here?

Re:4 GHz, eh?

[Anonymous Coward]

It's going to be very hot!

Re:4 GHz, eh?

[Mister Whirly]

That is, until Apple also started using Intel chips.

Re:4 GHz, eh?

[TheKidWho]

Yes, because your 3.4ghz dual core is faster then my 3.2ghz quad core. This had nothing to do with the Athlons wooping the Pentiums or Intel releasing "slow" Core Duo processors that were "half the speed" of their Pentium 4s.

Indeed, let's just throw Apple into this for the +1 relevant.

Re:4 GHz, eh?

[Mister Whirly]

Well, the whole "GHz isn't everything" mantra was repeated over and over by Apple when they were using slower clock speed Motorolla PPC chips. I noticed how they liked to show how much better the new Macs with Intel chips performed over their older PPC models after they made the switch as well. Kind of like "ignore everything we said before about how inferior those Intel chips were over our Motorolla chips now that we are using them too".

So that is what I was referencing. I wasn't supporting or denying any actual claims about GHz comparative speeds or any of that nonsense. I am not really into the whole "comparing e-cock size" contests about hardware and which is better than which. If it works for me, I use it. When it doesn't anymore, I buy something that will.

Re:4 GHz, eh?

[Anonymous Coward]

[quote]When making a direct comparison between chip X running at frequency Y and chip X running at frequency Z, speed does correlate pretty closely to frequency(unless you are being bottlenecked hard by some other aspect of the system).[/quote]Not necessarily. For example, I have an app that runs 50% faster on a 2GHz Phenom II than it does on a 3GHz Pentium D, because the app has a working set size that fits in the Phenom II's L1 cache (which is four times as big as the Pentium D's). There isn't any bottleneck there other than the processor itself, as it still fits into the Pentium D's L2, but that is several times slower than the L1.

Re:4 GHz, eh?

[TheKidWho]

Right, AMD and Intel didn't participate in the "GHZ isn't everything" campaign either.

Re:4 GHz, eh?

[Chowderbags]

Not if it has sex in it (sex with oneself doesn't count).

Re:4 GHz, eh?

[Anonymous Coward]

He said processor X with processor X dumbass

Re:4 GHz, eh?

[Hurricane78]

Man, you two would not have this problem, if you would qualify your deliberately broad terms like “faster”.

- Faster processing per energy usage = better.
- Faster processing alone = Not better, if most of the energy it goes up into heat. (Pentium 4: 63%)
- Faster in terms on MHz = Not better, if it still does less instructions per clock cycle.

Re:4 GHz, eh?

[Mister Whirly]

Pointing out Apple did participate does not mean AMD and Intel didn't, and I never made any claims or inferred that anywhere.

Re:4 GHz, eh?

[mitgib]

sex and the Pentium 4

Is that the new TV show for geeks?

I head CBS ordered the pilot then never picked it up, but SyFy ordered 12 episodes to air after Eureka

Re:4 GHz, eh?

[gman003]

Except that, for the same chip model, increased frequency does increase speed. If you take two processors that are identical save for the frequency, the one with the higher frequency will be faster. Unless there's a massive bottleneck elsewhere, in which case they will perform at the same speed. Ignoring hardware damage from doing it wrong, you cannot slow your system down by upping the clock speed on your processor.

Yes, frequency is nearly useless when comparing vastly different designs, but that's not the issue at hand. If this were comparing the new i7 to the new Phenom, it would be relevant. But we're comparing i7s to i7s, where different frequencies are reliable indicators.

Re:4 GHz, eh?

[Anonymous Coward]

Yes becasue AMD and Intel also played the game means Apple didn't? Nice strawman.

Re:4 GHz, eh?

[TheKidWho]

Except that, for the same chip model, increased frequency does increase speed. If you take two processors that are identical save for the frequency, the one with the higher frequency will be faster.

Sure, so? I didn't say that wasn't true.

This comment is from a guy who runs an i7 920 at 4GHZ on air.

Re:4 GHz, eh?

[TheKidWho]

Yes, but you imply it was mainly Apples doing which is false.

Re:4 GHz, eh?

[Mister Whirly]

Where did I do that? I pointed out that Apple did it in reference to their switching chips, and said nothing else. Any implication was made by you, and not me. If you want to, how about I just give you my password and you can just speak for me since you seem to know what I am saying better than I do??

Re:4 GHz, eh?

[TheKidWho]

That would make things easier for the both of us.

Re:4 GHz, eh?

[Anonymous Coward]

Yes, that is correct.

What we learned is that just because TheKidWho is "right" doesn't necessarily imply that he is not a nitpicky idiot.

Re:4 GHz, eh?

[TheKidWho]

Hey hey, just because I'm a pedantic nitpicker doesn't mean I'm am idiot.

-sent from my iPad

Re:4 GHz, eh?

[Anonymous Coward]

... that weren't originally MEN?? :D

I mean hey, after the thousandth or so rejection, you start wondering if it wouldn't be better to just get a set of breasts of your own to play with... if you didn't already have them :D

Re:4 GHz, eh?

[jvin248]

I think AMD and Intel only participated in the "Mhz is everything" game, by the time they got to Ghz then they were convincing us of energy consumption (heat) per cycle and other mysterious measurments.

Re:4 GHz, eh?

[dave87656]

like sex and the Pentium 4, faster isn't always necessarily better...

Thanks, that made my day!

Re:4 GHz, eh?

[lmnfrs]

I always thought those sparkling action figures were hot.

Re:4 GHz, eh?

[petermgreen]

He said "chip X running at frequency Y vs chip X running at frequency Z" (emphasis mine). Generally that is true, if you take the same chip or even two chips from the same family and run at different clocks the higher clock will almost certainly bring better performance.

Sometimes cache comes into play but generally within a familily processors that are clocked faster will also have the same or more cache.

For example, I have an app that runs 50% faster on a 2GHz Phenom II than it does on a 3GHz Pentium D, because the app has a working set size that fits in the Phenom II's L1 cache (which is four times as big as the Pentium D's).
Assuming you really mean a pentium D and not a pentium dual-core are you sure the preformance difference is down to the smaller L1 cache and not just due to the pentium 4's general suckiness.

Yawn

[MBGMorden]

"just $216 and $342"?

The majority of regular users can get by with just about any modern processor on the market today. Just glancing at Newegg, single core CPU's are starting at $32. Dual cores at $50. Quad cores at $81. I personally haven't spent more than $100 on a processor in ages, and I'm more or less a power user (do heavy programming and video encoding as well as other such tasks on my systems).

Now, at work, for servers, and I'm sure other users who are doing things like heavy graphics editing and such, people do need faster processors, but the people doing such tasks are NOT going to give two shits whether or not the multiplier is unlocked (anybody using an overclocked processor in a professional environment is just asking for trouble).

So you're left with the absolute hardcore hardware enthusiast market. Even in this market though you're going to have the "I'm poor and don't want to spend much" people who are still going for the low cost ones and trying to push them, and the "I've got money to blow and want the fastest available" people who were likely buying the really, really expensive stuff already.

In short, I just don't see this feature, at the stated price points, as really having much of a market.

Re:Yawn

[thijsh]

Yeah, but the $32 CPU has an e-peen value of zero, while these babies raise your e-peen over 9000!!! It's 'the most e-peen for your buck' (I think that was in the small letters at the bottom of the Intel advertisement).

Re:Yawn

[drinkypoo]

In short, I just don't see this feature, at the stated price points, as really having much of a market.

Corvettes and Camaros sell Citations and Celebrities.

Or in other words, Intel will sell more CPUs than AMD if they can convince the world that they have a bigger penis. Nobody wants their CPU to come from small penis guy, or to imply that they are one.

It is rather crazy that gamers buy these, though. Far be it from me to tell anyone how to spend their money; whatever makes you happy that doesn't hurt anybody (indictments of the western lifestyle aside for now) is fine with me. But if you stay just behind the curve you can upgrade every year (to last year's kit) and still be able to play virtually every game at quite good settings. Buying the latest and greatest comes with a massive price:performance penalty. Once you get into the new generation of lower-power equipment it seems like there's little motivation to ride the upgrade train so far or long.

Re:Yawn

[Nadaka]

Exactly. You can get an unlocked quad core 3.2Ghz AMD 955 for $160 on newegg, that is more than enough performance for anything a "gamer" is going to be doing. The only reason to even want more than that is if you are chasing ego with benchmarks/folding@home or are doing bulk HD video encoding/graphics rendering.

Re:Yawn

[Lord Ender]

If you do computationally-intensive workstation tasks, like video editing, gaming, virtualization, or using java (sigh); you really will appreciate going from a $100 CPU to a $300 CPU. Using faster components also means having an overall less-frustrating experience with your computer.

At home, I have an i7, an SSD, a high-end NVIDIA GPU, and the fastest RAM my mobo can take. At work I have a computer made of the budget components you think are good enough. The difference is extremely evident. My computing tasks happen as fast as I can think at home. At work, I often have to wait for things to load, which can derail my train-of-though, lower my productivity, or just generally piss me off.

A few hundred more for good components is money very-well spent.

Re:Yawn

[rrhal]

Whats even crazier is that some of these chips will end up in the desktops of the pointy-haired and their chosen minions because they feel that "Big Penis" CPUs will reflect better business practice. Intel sells a huge number of chips at a premium price to corporations when cheaper AMD CPUs would do the same job.

Re:Yawn

[Anonymous Coward]

I agree with this. I read the summary and thought "how much?!?" One of my recent computer purchases landed me a good machine with new monitor for around $400 and they want to sell me just a CPU for $342? Granted, I'm not heavily into gaming or number crunching, but there's no way I'm spending that kind of money on any one part. I can go by a second machine for that amount.

Re:Yawn

[Fross]

There is a guy who posts this exact message, almost word for word, every time a new CPU or graphics card is announced. "This is useless - I'm a power user and I can get by with a can of tuna and a bit of string".

Well, I'm afraid I have to tell you, you're not a power user, if you don't need the power that is now available. 2005's power user, maybe. But if you want to do video editing (and I mean final cut/premiere, not reencoding your dvd rips), play the latest games etc, then you do need that hardware. That software is designed to run on that hardware. And if you manage your own machine, whether it's for gaming of photoshop or whatever, you're going to care that this thing gives you bang for your buck. for what it's worth, this new chip isn't the fastest available. It isn't even close. It's the best value high-end chip, with a view to become even better value if you're open to overclocking it.

If you're not in that target audience, then fine - why do you complain about it? Do you bitch that lamborghinis are too expensive? "$150,000? I have a ford cortina that I got for $500 and it gets me to the mall just fine!". You don't see there is a market for this, because you are looking at a sample size of 1. Intel and AMD have a multi-billion business riding on this, I for one trust they're going to have done their homework.

(And I'm interested in the chip too - I'm planning a system upgrade soon - first since my Q6600 - and I like high-end, value chips I can overclock!)

Re:Yawn

[Pharmboy]

I concur. I just set my watch back one year and save thousands of dollars on everything. New games $50? Nope, mine are $25. New processor $300? Nope, mine are $100 and runs my one year old games perfectly. My last "new" car was $30k new but I bought it with 8k miles and just under 1 year old for $20k with full warranty. I'm about to buy a pair of Motorola Droids, which I can get for $99-$199 for both (2y contract, yes). It doesn't always pay off, but on average it saves up tremendously without sacrificing anything but a little time.

The net results is that I actually can buy MORE toys for the same money. Delayed gratification can be a beautiful thing.

Re:Yawn

[amazeofdeath]

Having been part of PC building community for a long time with a lot of experience in everything from low-end as-cheap-as-possible builds to ridiculously expensive gaming builds, your comment is not based on reality. Unless you have a very short memory, about 4 years ago you had to spend at least $200 to get any dual core CPU (Pentium D or AMD Athlon X2), so I very much doubt your "ages" comment on a CPU suitable for your uses. And especially vid encoding benefits a lot from quad core CPUs with suitable architectures, so were you seriously informed about these things, you would have put the limit to that $200, so that Intel's i5-750 (good at gaming, excellent bang for the buck in other desktop task, like video encoding) would have fitted into the budget. Of course, if your time isn't worth anything, you might as well get the cheapest CPU out there, but that doesn't mean that for a lot of people the $100 extra would make their computing experience a lot less time-consuming.

Re:Yawn

[Plekto]

Or in other words, Intel will sell more CPUs than AMD if they can convince the world that they have a bigger penis. Nobody wants their CPU to come from small penis guy, or to imply that they are one.

Except that, as GM learned, when you're impotent, the customers figure out that you're a non-player pretty quick. The Corvette no longer sells anything except for itself. In simple terms, the Intel pricing structure completely misses the overclocker's market segment. The average person who wants this feature easily enabled like this isn't going to spend much more than $100 in the first place if they can possibly help it. Their "goal" is to make the cheap as dirt CPU work like the $300 one.

Re:Yawn

[twidarkling]

Well, I'm afraid I have to tell you, you're not a power user, if you don't need the power that is now available. 2005's power user, maybe. But if you want to do video editing (and I mean final cut/premiere, not reencoding your dvd rips), play the latest games etc, then you do need that hardware. That software is designed to run on that hardware. And if you manage your own machine, whether it's for gaming of photoshop or whatever, you're going to care that this thing gives you bang for your buck. for what it's worth, this new chip isn't the fastest available. It isn't even close. It's the best value high-end chip, with a view to become even better value if you're open to overclocking it.

Thank you! I mean, hell. I like to think I'm a power user (I'm really not, but I like to think it), and I've talked to people who *are* power users, and every single one of them is always wishing their machine was a little bit faster, had a little bit more space, something. They're never satisfied with their machines, they're simply limited either by resources or technology. Hell, the book designer I know just got a dual quad-core machine with a crap-ton of RAM, and he still has to wait ages on certain processes.

(Oddly, I'm running a Q6600 too, and thinking of upgrading soon.)

Re:Yawn

[drinkypoo]

While your point is well taken, will you appreciate as well the difference between a $200 and a $300 processor?

Is it possible to get two sextuple-core processors on a single motherboard yet?

Re:Yawn

[sjames]

So, how will the CPUs hypnotize our kids into forming an air force?

Re:Yawn

[Kjella]

Considering that the 875 is exactly the same as the 870 except for the unlocked multiplier, and the 870 has cost $562 then $342 is a quite significant drop and a better bargain no matter how you look at it but none of them are that great. Also, the title of Anandtech's conclusion is Final Words: Not for enthusiasts. It seems more like what's been happening in graphics cards, that you will be able to get some "pre-overclocked" systems. This is pretty much a competitior to AMDs $299 Phenom II X6 1090T, giving you 4 faster cores instead of 6 slower.

Though yeah, for desktop use I really feel the air is running out. New games are mostly built to be xbox/ps3 compatible and don't strain a modern PC much. Mostly it's server/workstation applications that really put them through their paces these days, but it's not that interesting for the home geek. And when games like Red Dead Redemption are xbox/ps3 with pc edition coming "later, maybe" you know the PC is dying as a game platform not matter how much you stick to your guns. I already have a Wii, I guess PC+Wii will get me through this generation but the next generation I fear it's almost certainly a console as my main gaming station. At least it'll make it easier to run Linux with what's left...

Re:Yawn

[KahabutDieDrake]

I'm with you 100%. If spending more means I don't wait as much, I'm sold. Partly because my hourly income is pretty high, and if I spend a few hours a week waiting, it pays for the new processor. And partly because it drives me batty to wait on computers.

I once had a boss that said "you tell me what you need, and we'll get it". What I needed was 3 high end desktops in a single workstation configuration using Synergy to share inputs. It was beautiful, for once, if I had to wait, I just switched machines. I kept them all running non-stop and my productivity was through the roof. It more than paid back the cost of 3 2500$ computers.

Re:Yawn

[geekoid]

Which part of overclocking makes you think this is for regular users?

Re:Yawn

[geekoid]

This is why it absolutely necessary for engineers to convinces upper management that their penis size correlate to how ell the engineering dept works.

Re:Yawn

[geekoid]

However it leads to being a social outcast.

Re:Yawn

[bored]

If you do computationally-intensive workstation tasks, like video editing...

I don't know about you, but my video editor, doesn't consume much CPU time. HD speed (or a sh*tload of RAM) makes a much larger diffrence. On the other hand, when I'm done with the editing and I tell it to create the video it can spin for a long time. Generally though, I batch those up and let them run when I'm not at the computer. In that case having them finish at midnight instead of 3 AM, really doesn't matter much.

But, compiling C++ template code on the other hand, that just EATS CPU like no ones business.

Re:Yawn

[Anonymous Coward]

There's an XKCD for that: http://xkcd.com/606/

Re:Yawn

[stim]

i think you are confusing the definition of power user. power user != tosses tons of money at shit because they are trying to fill some void in their heart. I can be a power user on a 7 year old solaris box. also, don't believe the lies, i play all the latest and greatest games and my computer is a couple years old.

Re:Yawn

[rrhal]

These mother boards will handle two 12 core Opterons: http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=2010200302%201071357038&name=Dual%20Socket%20G34

Re:Yawn

[Peach Rings]

It's a shame too, because the only reason Intel is making this offer at all is to compete with AMD's similar offerings. AMD should see more business for its influence on the market alone, but nobody has financial incentive to be altruistic toward the underdog. More "free" market forces at work unfortunately.

Re:Yawn

[Skuld-Chan]

Dell mobility has those motorola's for a dollar...

Re:Yawn

[drinkypoo]

The average person who wants this feature easily enabled like this isn't going to spend much more than $100 in the first place if they can possibly help it. Their "goal" is to make the cheap as dirt CPU work like the $300 one.

I'm not sure that's true. Clock modders seem to come in three groups, which I will now overgeneralize. There are underclockers who are trying to run quiet and cool, or fanless. There are overclockers who are trying to get the hotness on a budget, like the dual celery 300A owners of old. And then there are overclockers who want to have the fastest PC available, and are getting rates better than the fastest processor speeds you can actually buy. I could make an automotive analogy, but I think the list is fairly clear; there are plenty of people who will buy an expensive overclockable CPU if it lets them have the fastest computer in their neighborhood, which they can then brag about. Not that I'm immune; I like to brag about price:performance ratio, and performance:wattage.

Re:Yawn

[Skuld-Chan]

I built a machine similar to that at home - Intel Skulltrail - its kinda old now, but it has two quad core QX9775 cpu's inside (basically a 3.2 GHz Xeon that has an unlocked multiplier). I wouldn't really wish that process on most people - lots of really specialized components (at the time there was only one company that could provide me with DDR2 ECC dimms that fit - Kingston), 1000 watt power supply etc etc etc.

On the plus side its still a very fast PC :).

Re:Yawn

[Lord Ender]

Yes, because that can mean the difference between dual-core and quad/six-core. It also means a lot more cache. Modern operating systems tend to have many background tasks, and modern users tend to have many simultaneous interactive tasks and tabs. When your computer can dedicate a core to what you're working on, it will feel smother than when your computer is switching things in and out.

One other thing I should point out: more expensive CPUs give you more processing power PER WATT. That money buys you a reduction in your electrical bill, in the noise of your system fans, and in the amount of AC you need to keep your room cool ;-)

Re:Yawn

[MBGMorden]

Congratulations on missing the entire point of my post.

Nowhere in my post did I state that fast CPU's had no purpose. I explicitly stated the opposite. I for example, throw all the CPU I can afford when I setup a database server because there, I will see the ROI.

My statement was about the clock multiplier being unlocked on essentially a middle of the road (as far as cost - not performance or sales volume) chip. If the announcement had merely been about an ever-faster chip then I wouldn't have even cared.

MOST people are fine with chips costing significantly less. As I stated, you can get quad core chips now for under $100. Those are fine for 99% of what users do on a computer. If you think you need more than that to do do real work (or gaming) on a computer, you're just delusional. A 2+Ghz quad core chip is a lot more than a "can of tuna and some string".

For users with specific needs for high-cpu utilization tasks, they typically are going to be in a position where they can afford faster chips, and they're certainly not going to want to dick around with overclocking in a professional environment.

Re:Yawn

[Pharmboy]

However it leads to being a social outcast.

That is a valid point for someone under 25, but I'm in my 40s, albeit not typical, as I still game regularly and enjoy new technology on a regular basis. I also fish and garden a bit to make sure I get some exercise. I don't game socially (multi-player) very often simply because the demands of family and work make it difficult to schedule gaming, but I still play at least an hour a day. For most of us over 30, delaying gratification works out to save some money. This means when the rare something new comes out and I just MUST HAVE IT NOW, I don't feel so bad spending the money, as that is the exception rather than the rule.

Re:Yawn

[blahplusplus]

"Delayed gratification can be a beautiful thing."

But so is splurging, if you splurge big and ride that splurge for many years the cost can tend to average out to be the same or less as the 2nd hand buyer. For instance I bought at the high edge of the curve 4 ish years ago and I _still_ don't have to upgrade, I can still ride what I bought 4 years ago at a "higher price".

Re:Yawn

[Pharmboy]

I only buy primary systems every 4-5 years as well, and tend to buy high quality, although not the fastest processor or video. My one year old system runs a Q9550, which I bought just after the i7s came out. Video is a Radion HD4600, not top end but still quite good. While it wasn't highest end, a Q9550 is still a damn good processor by today's standards, as is the HD4600. The rest of the system is solid quality, dual DVD burners, 8mb ram, etc. I DO tend to upgrade the video card about 3 years in. Before discounts, I had about $1800 in the system with a 24" LCD, which is upper middle price range.

By delaying gratification, I am mainly talking about software, games, although it did apply to my hardware to a lesser degree. It only makes sense to buy quality hardware, although the fastest isn't usually the best value.

Re:Yawn

[Pharmboy]

Dell mobility has those motorola's for a dollar...

Thank you! I hit my 2 year mark in a month, and the wife and I both have our eyes on the Droid: me so I can ssh in and work from remote, the wife because, well, she just like shiny new toys. Now I can go buy a bunch of $25 games with the $200 I will save ;)

Re:Yawn

[MBGMorden]

Nothing. I discussed regular users in the first paragraph of a multi-paragraph response to define them as one portion of the overall user base that wouldn't be interested, so that as each portion discussed is deducted from that total, you're left with the group that this IS for - AND IMHO that left over group is a pretty small crowd.

Re:Yawn

[Plekto]

Yes, but consider that they upper-end group in your example is less than 1/3 of the potential market.

Fanless means low speeds or even *under-clocking* the FSB to keep temps down, so cheap and functional CPUs are the norm. The second group also is concerned with a budget. The third doesn't care and skips this middle-ground entirely.

Sure, this move by Intel helps some people, but AMD has a full line for the other two groups already in place. It really does strike me as a GM type move. Nice effort, but a bit too late to the ball game.

Re:Yawn

[evilviper]

My computing tasks happen as fast as I can think

I can see why they wouldn't want you to have a faster system, at work...

Re:Yawn

[drinkypoo]

Yes, but consider that they upper-end group in your example is less than 1/3 of the potential market.

You're making a false assumption. The market is not necessarily split into a given number of equally-sized groups. I suspect that the underclockers are the smaller group by far, for example, though I admit I haven't exactly conducted a survey.

Re:Yawn

[norminator]

You're so productive...

Which of the three did you use to post that on slasdot?

(I kid.. sort of)

Re:Yawn

[jvin248]

Also want to look at the natural stretch built into a system. Is the machine at it's upper end already - like getting a Pentium2@450Mhz or Pentium3@450Mhz. The Pentium 3 system could have been capable of reaching into 750Mhz range where the Pentium 2 never would. Or multiple ram banks open?

Re:Yawn

[jibjibjib]

I have a meaningful social life which doesn't depend at all on owning technology less than a year old. Is that unusual?

Re:Yawn

[MikeBabcock]

Not to mention you can get a black edition (overclockable) AMD Phenom II X4 (four cores) for $50 less than that 2 core Intel proc.

This isn't competitive.

Re:Yawn

[dave87656]

Agreed. If you have X$ to spend over time, you'll get more bank for your buck by not buying the latest and greatest. The upper-middle range will cost you half and you can update twice as frequently.

I am in a constant battle with those at work who believe that more expensive is always better. Problem is, one guy comes from the bankrupt company we just bought out. Their server cost, I kid you not, $100,000. We are running both businesses now on a $500 Dell desktop running as a server (plus another $300 box as a backup). The bankrupt company spent so much money on their server 8 years ago or so, that they blew their wad and couldn't afford to upgrade the server, which by today's standards, is incredibly slow.

Re:Yawn

[drinkypoo]

One other thing I should point out: more expensive CPUs give you more processing power PER WATT.

I haven't done the comparison here, but generally speaking, intel gets stupid expensive on the high end, whether you're talking about price:performance or ops:watt. It's a lot to pay for a minor improvement in power savings. Definitely looking forward to AMD's next process shift...

Re:Yawn

[petermgreen]

But so is splurging, if you splurge big and ride that splurge for many years the cost can tend to average out to be the same or less as the 2nd hand buyer.
But does "splurging big" on a computer really extend it's lifetime all that much. Is a $1000 computer really so much better than a $500 computer that it will have double the useful life.

I would get a decent (but not insane) CPU up front though, most other components (graphics, ram, storage) can be easilly upgraded later but CPUs can be a PITA due to frequently changing sockets (AMD have been better about this than intel recently but there is no gaurantee that will continue).

Re:Yawn

[Anonymous Coward]

And a GoatKCD for that too: http://goatkcd.com/606/ (warning: NSFW)

Re:Yawn

[binaryspiral]

It's a good price for CPUs compared to the H5550 Xeons we just ponyied up a grand a piece for. Rockstar performance comes at rockstar prices.

These are a steal.

Re:Yawn

[CAIMLAS]

New processor $300? Nope, mine are $100 and runs my one year old games perfectly.

Let me guess: you buy Intel processors?

New AMD processors can be gotten for $100, no problem. And yes, they'll play the new games just fine (the problem, of course, is usually going to be that $200 graphics card).

Nothing to lose

[Stan Vassilev]

Intel have nothing to lose anymore by keeping the multipliers locked: the bottleneck isn't with the CPU frequency anymore. The biggest differentiators in their higher end models are number of cores and cache size.

If they can get few more sales with a pointless gimmick some fall for, why not?

Re:Nothing to lose

[Anonymous Coward]

For games that don't use all cores, which is almost all games, CPU frequency is the only bottleneck left. And these CPUs cater to the gaming market, obviously.

Re:Nothing to lose

[petermgreen]

The biggest differentiators in their higher end models are number of cores and cache size.
I'd add the platform to that.

The low and midrange desktop chips (i3, i5 and i7 8xx) are on the LGA1156 platform. That means dual channel memory, a max memory of 16GB* and only 16 channels of fast PCIe**. What that means is that any addition storage controllers etc end up either stuck behind the PCH or stealing channels from the graphics card.

The high end desktop chips (i7 9xx) are on the x58 single socket LGA1366 platform, that means triple channel memory, a max memory of 24GB* and 36 channels of fast PCIe**.

The xeon 5500/5600 workstation/server chips are also on LGA1366 but it's a dual socket variant. That means six channels of memory (three per CPU), a max memory of 144GB* and 36 or 72 channels (depending on whether the motherboard vendor uses one or two IO hubs) of fast PCIe** (and yes there are boards that use all 72 e.g. http://www.supermicro.co.uk/products/motherboard/QPI/5500/X8DAH_.cfm ).

The higher end xeon chips support even more.

* Max memory figures for desktop chips assume 4GB modules and two modules per channel. I haven't seen a desktop board that claims to support higher configurations than this though it may technically be possible. Max memory for dual socket workstation chips assumes 8GB modules and three modules per channel, more is possible in theory if you can find 16GB modules which supposedly exist but i've never seen for sale.

** PCIe channels that run at the 2.0 speeds and are taken from the processor or IOH. Not the channels that are taken from the ICH/PCH that only run at 1.0 speeds and are potentially bottlenecked by the DMI/ESI connection.

Goes to 11

[wjousts]

It appears that both will easily hit speeds in excess of 4GHz with air cooling.

Nigel Tufnel: These go to eleven.

Why?

[Anonymous Coward]

"Enthusiasts" are people with more money than sense. Why would anyone pay more than $150 for a CPU these days? A quad-core 3.0Ghz chip is not going to be your bottleneck. Yeah, I guess if you spend all day ripping and encoding video then that extra 10%-20% might amount to a few minutes saved, but for most people spending the extra $150 on an SSD drive instead would give them a far more noticeable performance boost.

Or, if you've still got money to burn, buy a nicer monitor-- or a second one, for that matter. Or some quality case fans that don't make your case sound like a jet engine. Or a decent set of speakers.

The obsession with CPU "speed" is dumb.

LOL

[Jaysyn]

Well, I guess when anti-competitive practices fail.....

Re:LOL

[magus_melchior]

It's also amusing to see the new AMD-based notebooks lined up this year, after literally years of ceding the market to Intel. Maybe they run hotter, maybe they're not as energy-efficient, but I don't see those as reasons to let their mobile line languish.

Now if they could release XP drivers for some of their newer chipsets, that would be great-- unfortunately the vendors (*cough*Toshiba*cough*) have AMD by the nads on this one...

Depends on how you read the chart

[drinkypoo]

At the $200 price point, AMD is still killing it. Look at the scatter plot, and note what happens on the $200 line. Now, draw an imaginary $100 line, and check that out. It's all AMD. So while you may want to buy intel if you want today's fastest gaming machine, AMD is still the processor for those of us who want performance and money at the same time.

With that said, can anyone recommend a good AM3 air cooler that's not too spendy? I have a PhII X3 720 retail black edition that I'd like to overclock. The stock cooler won't cut it :) But I want to keep my budget very small, which is why I went AMD in the first place. So far, so good.

Re:Depends on how you read the chart

[Nadaka]

Not sure what you call "good" but I am using this one on my 140W 965, it is certainly better than stock but still a bit "budget". I have not tried overclocking it yet, no real need. But it should be able to handle the heat of your lower wattage processor while overclocked just fine.

http://www.newegg.com/Product/Product.aspx?Item=N82E16835103064

Re:Depends on how you read the chart

[drinkypoo]

What a coincidence, I got impatient and ordered one of those just now.

Re:Depends on how you read the chart

[amazeofdeath]

Try Coolermaster Hyper 212 Plus, or if you want something cheaper, Arctic Cooling Freezer series has been quite good.

You are wrong about the $200 price point: The chart has put the hexacore AMD above Intel's i5-750, which comes from some synthetic benchmark results and a few real-life apps, and it doesn't reflect the overall real-world results, where i5-750 is still the better choice (this comes again from higher clock speed and better architecture vs. more cores). However I agree that in the cheap end AMD is probably the better choice.

Re:Depends on how you read the chart

[magus_melchior]

My current build uses an Arctic Cooling Freezer. Nice big aluminum/copper heatsink with a big fan, quiet, and keeps my CPU at sub-50 even under load-- though your 720 BE might run hotter than my 810 if it's overclocked. At about $30 on NewEgg (and still being built & sold), I think it's a great cooler for the money.

My motherboard (Gigabyte) doesn't like the fan, however-- the MB's control software sometimes fires off a warning that it's not working correctly. I think it's because it's a 3-pin rather than 4, though I'll have to research that.

Re:Depends on how you read the chart

[Anonymous Coward]

there's a $20 rebate on this liquid cooler at tigerdirect.com

CoolIT Systems ECO-R ECO

i see it gets good reviews in general, and should work well with overclocking, i just ordered one for my phenom ii 720BE as well. i can only get 3.1 stable on air cooling

Re:Depends on how you read the chart

[Anonymous Coward]

Artic Cooling Freezer 64 FTW! no leds or stupid stuff pure 130W thermal glory. 20C less than stock HS in a OC Toliman Phenom 8650 with OCZ TIM and it's really quiet under 38C. Under $35, I bought it in Amazon and even with shipping to south America I got better bang/buck than the lame themaltake offers around here.

Not affiliated, only a happy customer, read the NewEgg reviews.

i0

4GHZ

[clinko]

The cards! they bounce TOO FAST!!

Choose wisely...

[KillShill]

http://clusty.com/search?query=intel+monopoly

http://clusty.com/search?query=intel+apartheid

Meh

[dlapine]

The article comparing values uses the highest price motherboard available for AMD for a "midrange" system, then claims that the Intel-based total system is a value. If you spend $350 on a 6-core processor, then spending $140 on a high-end motherboard is reasonable. If you're spending $99 for a low end AMD quad, you're probably in the market for more reasonably priced motherboard (~$100) to go with it. The comparison is valid for the high-end AMD cpus, but not their budget stuff, as a $40 drop in price is a big deal for a system with a $100 cpu.

That being said, being able to overclock this thing is directly aimed at the enthusiast market. "I got 6 cores, w00t!" "Yeah, well I'm at 4GHZ on a quad, so there!" It definitely improves the competition between the high end AMD hexa-cores and the midrange Intel quads, and makes the Intel option more appealing to the enthusisast.

Re:Meh

[Trev311]

The only problem with that is the AMD motherboards (especially in the price range they chose) come with USB 3 and SATA 6.0 Gb/s now. Want that on the Intel machine? Extra money... That tends to get rid of the whole Intel value thing since most of the time you're not going to be bottlenecked by the CPU and you get more for your money going with the AMD combination.

Re:Meh

[Rockoon]

Those hexa-core systems are reported to be very over-clockable. The 1090T is observed to be stable at 4.1ghz on stock cooling and the 1055T up to 3.9ghz.

In the end tho, its your memory over-clocking thats most important in gaming. Even the 4-core systems are often fighting for bandwidth at 3ghz+. Thats one of the reasons that Intel is so far ahead on the high end. Their smaller process allows for a much bigger cache. Once AMD moves to the next process size, its going to be a very interesting fight because Intel is nowhere near a drop in process size at this point in time.

If I was doing any sort of heavy calculation type stuff, there is absolutely no way that I wouldn't go with the 1090T. For gaming I would go with the 955 or an i5.

better performance per dollar ..

[Anonymous Coward]

only if you fudge the numbers.
TFA chooses a totally arbitrary system price, which is just high enough to make this cpu top the chart.
Any methodology that claims the $1000 i7-980X is better value than the $300 i7-920 is BS.

NO 1336? so you are stuck with 16 pci-e lanes so

[Joe The Dragon]

NO 1336? so you are stuck with 16 pci-e lanes so a good video card can uses that up and then when you add usb 3.0 and sata 6.0 you cut into the video pci-e lanes.

for $200 you can get a AMD board with 890fx that has more pci-e lanes so you can have 2 x16 video cards + room for sata 6.0 and usb 3.0 as well.

Summary conclusion and linked article is bull

[Rockoon]

Anyone who can build comparable systems for significantly less than those bozos, raise your hand.

Shift the second scatter graph to the left from anywhere between $200 to $400, and then draw a new conclusion.

Re:Summary conclusion and linked article is bull

[Anonymous Coward]

I don't get it. How are you saving $200-$400 and building anything like a comparable system?

Re:Summary conclusion and linked article is bull

[Rockoon]

You dont need a $150 motherboard for that setup. An $80 one would do just fine (still offering sata3.0, usb3.0, and ddr3/1333.) Save $70 here.

You don't need to spend $115 for a case and power supply, either. You can literally save 50% here without even trying.

For that video card you can get equal performance for $120, saving $40.

Whats the deal with spending $110 on a 1TB plater these days? $85 at most. 6GB sata doesnt help these drives, either. These drives cant even saturate 1.5GB sata.
I'm already at ~$200 savings, and thats using equal-feature parts. The system doesnt actually need equal feature parts tho, since hes gone overboard on some of the features: he doesn't need a 500W PSU, he doesn't need sata3.0, he doesn't need a pair of PCI x16 slots, he doesn't need to support DDR3-2000 since hes slotting DDR3-1333's, the motherboard doesnt need integrated graphics, ...

good amd MB are $150-200 good intel ones $200-250

[Joe The Dragon]

good amd MB are $150-200 good intel ones $200-250 and the amd boards have more pci-e lanes as well.

Nice to see AMD v Intel reversed...

[dtjohnson]

Intel Targets AMD...

Nice to see this. In the past 10 years it has always been 'AMD targets Intel.' AMD must be doing something right if Intel is taking notice of it and that means a little competition which is great for the future of the hardware.

Please explain overclocking.

[maillemaker]

Can someone please explain overclocking to me? Why are processors sold at a slower speed than they can actually perform at? Why don't they ship from the factory at their fastest speed?

Re:Please explain overclocking.

[xiando]

Simple. You run the CPU at a higher clock speed than what it was designed for. It works, and this will give you a higher clock speed, draw more power and produce more heat. The price different really isn't that big these days and nobody really cares about overclocking anymore, but it used to be a big thing back in the day.

Re:Please explain overclocking.

[adeft]

Car analogy: Every car comes from the factory tuned in a conservative manner regarding air/fuel mixture, spark advance.....even torque management and shift pressures/points (in the case of automatics) You are able to change these values for a greater increase in performance or fuel economy. This change (when pushed to the limit or performed incorrectly) can be risky and put extra strain on components of your vehicle.

Re:Please explain overclocking.

[Endo13]

Lots of reasons, but one of the biggest is because they're often not 100% stable at higher frequencies without increasing the voltage.

Re:Please explain overclocking.

[Fross]

Historically, they make a whole batch of processors together, then run some tests to see how fast each will run. Some go faster than others (or more to the point, some are reliable at higher speeds, some less reliable) so they get divided into different speed batches and sold at different prices.

The reasons behind the reliability are varied, however, and mostly down to heat dissipation. As the chip gets hotter (and they get v hot inside) it gets less reliable. The manufacturer tests against a standard heat diffusion system, but some people will spend more money on good cooling, either a big heatsink and fan, or even water cooling, or down to liquid nitrogen (!). The upshot is, any particular chip will be more reliable at a higher speed.

Sometimes, however, it's just about the market. It may cost $200 to manufacture your product, but some people want to buy a good product at around $300, some want to buy one at $500 and have the best. So you sell your product for $500, but nobble a few (say in firmware) to run a bit slower, and sell those for $300. Overall, you will sell many more that way than just the $500 ones, so you make more money.

This is EXACTLY what happened with the Geforce 6800GTS (If I remember correctly) - they used to manufacture dies with 4 cores on them. If it had 4 cores working, it was a high end model, if it had 3 cores working, it was a low end model. This allowed them to increase yields dramatically, as all the ones with just one fault still sold. However, the market demand for the low end one was far greater than the number of defective parts they had, so they ended up taking the 4-core model, locking one core in firmware and releasing it as a low end model - after all, selling a card at a cheaper price is better than not selling it at all, right?

The upshot was some people could programmatically unlock their 4th core, and get a high end card for low end price! I tried it myself, but had one with a defective 4th core so just got a bunch of video corruption until I locked it again :>

Re:Please explain overclocking.

[Anonymous Coward]

Where's the +1 Good Car Analogy mod?

Re:Please explain overclocking.

[ampathee]

Overclocking increases both frequency and heat output. Too much heat == many crashes and/or dead CPU.

CPUs are generally sold a speed that they are stable at, within the expected normal operating temperature range.

Overclockers usually replace the stock cooling system with something better before overclocking.

Re:Please explain overclocking.

[sjames]

It happens whenever production and market segmentation don't align.

When a new model comes out, especially if there was a die shrink, imperfections are significant. All chips are manufactured the same, but only a few will run reliably at the top speed. In a process known as bin sorting, all chips are tested at the top speed. Some pass and are so marked. They might run faster than that, they might not, but returns are costly and damage reputation, so a bit of margin is left at the top.

Everything that failed is tested at the next lower speed. The process continues down to the bottom of the line speed. The process may be more complicated in some cases. For example, at one time, a CPU with a single cache error would have a fuse blown to disable the half that had the problem and the chip became a Celeron.. These days, whole cores can be disabled. AMD sells those.

As production kinks get worked out and the process is adjusted, yields and quality go up to the point that they may produce more of the faster CPUs than the market demands. Rather than appearing to have production problems or turn down sales, they will mark the faster processors for the lower speed and sell them that way (at a lower but still acceptable profit).

That latter stage is where Intel starts to hate overclocking since lower marked cheaper processors likely WILL run faster. Perhaps their greatest concern there is shady gray market vendors who will buy cheaper processors in bulk and change the markings. Of course, with laser etched heat spreaders and such that is less of a real concern. Then it just becomes home users knowingly overclocking.

Re:Please explain overclocking.

[TheThiefMaster]

I even saw one with two cores (actually "quads" of four cores IIRC) locked, and we successfully unlocked one of them. The other was completely fucked. But you can't say no to adding 50% to the power of a card with just a couple of hours experimenting!

You've nicely summed up three of the kinds of overclocking (restoring the clock speed of a slow-binned-but-actually-really-good cpu, unlocking extra cores and actually overclocking a cpu using a decent cooling system). I remember in the Athlon XP era all the cool kids were taking mobile and other low-power cpus and overclocking the crap out of them. Most people had XP-M 2400s or 2600s, but one of my mates managed to find a Athlon XP-M 2800+ (2.1 GHz stock), and ran it at 2.8 GHz perfectly stable on just air cooling. Royally pissed off my brother with the water-cooled 2600+ that couldn't break 2.5 GHz. Considering the fastest chip of the era was the 2.3 GHz XP 3200+, getting 2.8 GHz out of an air-cooled setup was just insane.

Of course, not long after the Athlon 64s came out and eclipsed everything :)

Re:Please explain overclocking.

[drsmithy]

Can someone please explain overclocking to me? Why are processors sold at a slower speed than they can actually perform at? Why don't they ship from the factory at their fastest speed?

Binning and market segmentation.

Same reason you can buy basically the same car at multiple price points with different options.

Re:Please explain overclocking.

[drsmithy]

Historically, they make a whole batch of processors together, then run some tests to see how fast each will run.

Nitpick. They don't test each chip, they test a few samples from each batch and grade the whole batch based on them.

Re:Please explain overclocking.

[Anonymous Coward]

How bout this explination. Amd comes up with this new cool chip design and fabricates a large batch of them. They then run a quick stress test to see how far they can push the chips. The ones that are perfect get a bunch of memory strapped to them and are sold as the Phenom II X6 Super Black PRO edition. The ones that can't handle as much get the weak parts disabled and are sold off as the Athlon II X2 basic model.

Now fast forward 3 years of additional research and the engineers have weeded out most of the defects in the fabrication process. Now 90% of all the chips that come from the latest fabrications are fit to be sold as the Phenom II X6 Super Black Pro whatever model, however the majority of your market isn't interested in the super expensive top of the line parts, and more importantly won't pay for them. They just want their Athlon X2 basic models. So AMD just relabels the enthusiast class parts as budget parts, because they're all really the same part underneath.

Posted AC because I'm too lazy to login and find this again.

Re:Please explain overclocking.

[Anonymous Coward]

There's not always as large a market for higher end CPUs. Even if their factories are churning out CPU dies capable of running at the highest speeds, they may artificially reduce speeds to be able to sell more processors at lower price points. I'm not sure this happens as much as it used to, but it's just a means to grade the product. There are markets for processors between $50 and $1200, even if 80% of the CPUs perform like the $1000 model.

Re:NO 1336? so you are stuck with 16 pci-e lanes s

[Anonymous Coward]

No, here at Intel, we only support 1337.

Re:Goes to 11, Oblig

[NotOverHere]

Obligatory XKCD
http://xkcd.com/670/

Description is flawed

[iamhassi]

"Surprisingly, even at stock speeds, the i7-875K offers better performance and power efficiency per dollar than just about any other desktop CPU out there."

-1, Inaccurate

The 2.8ghz i7-930 is $199 vs $342 for a 2.93ghz i7-875K, so almost double the price for 0.13ghz more. How did the author see that and think "better performance per dollar"? The article he linked to even shows the better performance per dollar in a chart, and btw techreport that chart is pretty piss poor, shoving $200 processors on a chart that goes to $1200 just clumps 90% of the processors in the $50 to $400 range. Learn how to make a chart: you should have left off under $50 (no processors under $50) and anything past $1000 (no processors over $1000). Because of your crappy chart the i7-875 is right next to the i7-930 despite the $142 difference.

The i7-930 is locked but it does reach 4ghz on air rather easily.

I suppose all of this is a mute because the LGA 1156 platform and LGA 1366 platform are being discontinued next year, so if you don't already have a i7 compatible motherboard you'd be buying a board that won't be compatible with any cpus made 7 months from now. I wouldn't buy a i7 cpu unless intel started selling them for $50, while AM3 boards available now are compatible with future 16-core cpus

Re:Description is flawed

[geekoid]

Do you base all your decisions on rumor, or just ones that fit into your proAMD fanboiisms.

Re:Description is flawed

[travisco_nabisco]

I am glad someone else noticed this.

I also found that looking through all the charts on the linked page, if there is ever a tie between the AMD and Intel system, the Intel result gets a slightly longer bar on the bar graphs.

My final complaint with the linked article is that they are using a single synthetic benchmark for all of their results. To find a meaningful result I would expect them to run a full suite of benchmarks that would include: synthetic tests, rendering, gaming, productivity, etc. The average all the results in a meaningful way and use that average to determine the price / performance ratio.

Re:Description is flawed

[(H)elix1]

The 2.8ghz i7-930 is $199 vs $342 for a 2.93ghz i7-875K, so almost double the price for 0.13ghz more.

To be fair, the Microcenter price is a bit of a loss leader. Trying to get a 930 just about anywhere else, and it goes for ~$288. While the multiplier is locked, the FSB on both the 920 and 930 can be bumped up to some crazy amounts. My experience with the 920 let the x21 multiplier clock up to a stable 4.6ghz using a H50 cooler. Same sort of overclocks folks are seeing with the $1000 975 Extreme Edition version.

It will be nice when the stock multipliers start hovering around 4ghz by default, rather than everything dabbling around 3.

Re:Description is flawed

[TheTyrannyOfForcedRe]

I suppose all of this is a mute because...

-1, Poor Language Skills

The word you were looking for is moot not mute.

Also, the entire world does not live near a Microcenter.

Re:Description is flawed

[Anonymous Coward]

You're an Apple fanboy
You're an Intel fanboy
I'd bet big money that you're an nVidia fanboy too

*This guy must be pretty stupid* I have the bridge you're looking for!!!1

Re:Description is flawed

[InvisiBill]

I suppose all of this is a mute because the LGA 1156 platform and LGA 1366 platform are being discontinued next year, so if you don't already have a i7 compatible motherboard you'd be buying a board that won't be compatible with any cpus made 7 months from now. I wouldn't buy a i7 cpu unless intel started selling them for $50, while AM3 boards available now are compatible with future 16-core cpus

First off, 3Q11 isn't 7 months away.

And it's all relative. By that time, 1366 will be three years old. I'll have had mine for over 2.5 years. Note that I bought my i7 system because my high-end AMD board, which was 16 months old at the time, was only AM2 and therefore couldn't support the newer AMD CPUs. I had to buy a new board whether I went with AMD or Intel. Even the article you linked states that guaranteeing motherboard support this early in the game is difficult, but AMD is usually good about maintaining socket compatibility. You may be able to slip a Zambezi into your current day Socket-AM3 motherboards. You and your current AM3 board might end up being just as screwed as me and my 1366 board.

Plus you're forgetting that the Phenom II X4 was designed to compete with the C2Q, not the i7. In an old review (can't find the link right now), the i7-920 at stock 2.66GHz still beat out a Phenom II 955BE overclocked to 3.7GHz by 27% on some compression benchmark. http://www.bit-tech.net/hardware/cpus/2010/04/27/amd-phenom-ii-x6-1090t-black-edition/6 shows similar results - in the 7-Zip/Mplayer test, the stock 2.8GHz i7-930 beats the 3.87GHz Phenom II X6 1090T Black Edition by over 20%. Yes, the AMD has 50% more cores (not taking into account HyperThreading) and 38% higher clockspeed, but the Intel is still 20% faster.

AMD's CPUs can be a great value, especially for gaming where the GPU is usually the bottleneck. However, Intel's CPUs are simply more powerful for a large number of tasks. Depending on what hardware you currently have, there may or may not be a significant difference in upgrade costs between the two brands. It's pretty safe to say that the Intel will probably cost more, but you'll probably also be getting more performance out of it.

Add a TURBO button;, You'll be more than satisfied

[Anonymous Coward]

Turbo buttons always gave a decent boost in performance. I say, bring them back !!

Re:NO 1336? so you are stuck with 16 pci-e lanes s

[TeknoHog]

No, here at Intel, we only support 1337.

But over here at the 1336 0f 1337 63n700m3n, either one is fine.

Sex tip

[geekoid]

faster is better as long as the duration remains the same or longer.

Affordable?

[medelliadegray]

Intel is not affordable compared to AMD.

My system here has been frizzing out on me after 4+ years, and literally 2 hours ago I ordered an AMD 3ghz 3-core processor for 75 bucks. the board (with on board GPU) , 2gb memory, and cpu with shipping came to $230...

JUST 14 bucks more than the only Intel's dual-core processor listed in the article.

Re:good amd MB are $150-200 good intel ones $200-2

[amazeofdeath]

Based on what? There are plenty of good LGA 1156 mobos available in $100-150 range. I'd guess you have no idea of what you are talking about on the PCI-E lanes.

Re:NO 1336? so you are stuck with 16 pci-e lanes s

[amazeofdeath]

I call bullshit. USB 3.0 doesn't use PCI-E lanes.

Re:NO 1336? so you are stuck with 16 pci-e lanes s

[amazeofdeath]

To show how much full of shit you are, see this for example: http://www.newegg.com/Product/Product.aspx?Item=N82E16813131641

Triple SLI at X16/X8/X8, or dual SLI at full X16/X16. Seems to be somewhat more than your alleged 16 lanes.

To bad the cpus hear are 1156 that come with low p

[Joe The Dragon]

To bad the cpus hear are 1156 that come with low pci-e lanes.

Re:To bad the cpus hear are 1156 that come with lo

[amazeofdeath]

Too bad you don't know how much or little the PCI-E lanes have to do with the CPU.

so the usb chip use magic to talk the main chipset

[Joe The Dragon]

so the usb chip uses magic to talk the main chipset?

I haven't forgotten the pentium 90 math bug fiasco

[mrflash818]

I haven't forgotten the pentium 90 math bug fiasco.
http://en.wikipedia.org/wiki/Pentium_FDIV_bug

I have used AMD exclusively since then, and will continue to do so.

Re:To bad the cpus hear are 1156 that come with lo

[amazeofdeath]

And here's a tri-SLI LGA 1156 board for you: http://www.newegg.com/Product/Product.aspx?Item=N82E16813130258

Re:so the usb chip use magic to talk the main chip

[amazeofdeath]

Do you have any idea of what you are talking about here? If your theory of USB 3.0 eating up PCI-E lanes was correct, how would that not apply to AMD boards too?

The DIY modified yugo argument

[Cprossu]

I went through the numbers myself... they don't seem to add up very well as TFA would lead you to believe...

let's first blab about something totally unrelated---

Someone buys a Yugo and does a lot of modifications to it, another buys a stock low-mid range performance car-so something like this:
"How much money have you spent total on your car?"
Yugo boy: I spent $5,000 to buy my car, and $16,000 in modifications
Stock boy: I spent $21,000 to buy my car.

"What kind of gas mileage do you get? (in MPG)"
Yugo boy: I used to get 34MPG before I modified it, now I get 18, and I can only use high octane
Stock boy: I get 18MPG on regular 87 octane

"Compare your cars speed on a generic track with straight aways and turns, be courteous if you can"
Yugo boy: Stock boy's car is faster than mine in turns, but my car out-accelerates him in straightaways.
Stock boy: Yugo's car is faster than mine in straightaways when it's not broken down at the side of the road or if his engine doesn't hiccup, I have the edge in braking and turning.

"How do your cars do for everyday driving?"
Yugo boy: I've removed the power steering for wright and power, the ride is loud, bouncy, and stiff, but that just means I can feel and hear the road better, and if I am at a drive-through or a stop light for an excessive amount of time it might overheat.
Stock boy: I have a comfortable quiet ride with no issues, and my car is easy to drive.

"What do you do when things go wrong?"
Yugo boy: I have to fix it myself, buying any parts that broke.
Stock boy: My car is fully under warranty, and I can have almost any part that broke replaced for free, as long as it wasn't directly my fault.

"What options do you have to make your cars faster?"
Yugo boy: I am maxed out, all I could do is lessen the weight more and buy a different engine, both are extremely expensive to do
Stock boy: There are an array of bolt-ons, computer re-programmers, and other easy modifications I can do to this car to make it faster, quite a few of them don't even void the warranty.

so relating this to computer parts...
setting up 2 rigs almost the same (full atx name brand mobo with at least 2 pcie slots, radeon 5770 graphics, corsair 650watt psu, cooler master case, blu-ray reader, a dvd burner, a 1tb 7200rpm hdd), one with a core i7 920 and 3gb ddr3 1600 ram in tri channel ($1052 with a stock heatsink) and the other with a "new" core i7 875K and 4gb of ddr3 1600 in dual channel ($1062 with an aftermarket heatsink), the "modified yugo" ends up being more expensive than the stock mid range car... For me it's a no brainer, and this isn't even factoring AMD into the equation...

When we look at AMD it ends up being $670 for all the support products (same specs used, full atx brand name mobo with at least 2 pcie slots, radeon 5770 graphics, corsair 650watt psu, cooler master case, blu-ray reader, a dvd burner, a 1tb 7200rpm hdd), 4gb ddr3 1600 in dual channel, which leaves you quite a bit of money to spend on the CPU and heatsink you want, and there are a nice amount of options from $50-$310 depending on the amount of tinkering you want to do to get the performance you are looking for. (which correlates to a price range of $720-$980 without the heatsink.
So depending on the level of AMD tinkering you would want to do you could have $72-$82 differential for the high end AMD chip to spend on cooling/whatnot or if you look at the low end you could have up to $332-$342 differential between those intel options to spend on cooling/whatnot... (so you could do a nice heatsink fan/low level water setup with the high end cpu, or an all-out phase change setup with one of the bargain bin AMD chips)

So my opinion before this was announced and after it was is still the same... if you go intel, socket 1366 is the way to go, if you want a cheap capable rig, AMD has a lot to offer.
Some drastic price cutting could change my opinions though!

THANK YOU TO EVERYONE ABOVE.

[maillemaker]

Now I understand what is going on.

Basically, the chips are sold to reliably run at what they are set to run at, but they might run faster, especially if you put in robust cooling.

Interesting.

I just bought a new system with a i7-930. Supposedly it can be overclocked. I have not fooled with it. The system does have liquid cooling though.

Steve

Virtualization

[Anonymous Coward]

Am I the only one who cares about virtualization? AMD includes it in all their products - finding one from intel is a real PITA and not available in any reasonably priced cpus.

Re:Virtualization

[InvisiBill]

What? The cheapest LGA1156 CPU listed on Newegg, the Pentium G6950 for $98, supports VT. http://ark.intel.com/Product.aspx?id=43230 So does the cheapest i3 they offer, the $115 530. http://ark.intel.com/Product.aspx?id=46472

The older LGA775 chips did use VT as a selling point, but I believe that most current Intel CPUs support it. Since VT was required for Win7's VirtualPC XP Mode, it became a more standard option recently.

Mod parent down

[amazeofdeath]

*sigh* I thought this was a tech site. What does one have to do to counter this kind of junk?

Here's (an allegedly non-existent) LGA 1156 board with 3 x PCI-E X16 slots: http://www.newegg.com/Product/Product.aspx?Item=N82E16813130238

There's plenty of room for those USB 3.0 and SATA-600 cards with the extra 4 expansion slots, even if the parent seems to suggest that just having them integrated would eat up the PCI-E lanes.

amd boards have more pci-e lanes then the intel on

[Joe The Dragon]

amd boards have more pci-e lanes then the intel ones. But ones the listed intel cpus have only have 16 + DMI and DMI is full with Intel on board sata 3.0 / usb 2.0 / other on board stuff and the add on sata 6.0 / usb 3.0 have to use the 16 for video.

uses a switch the shears the 16 pci-e lanes

[Joe The Dragon]

uses a switch the shears the 16 pci-e lanes

Re:uses a switch the shears the 16 pci-e lanes

[amazeofdeath]

To do what? If you use a card for USB 3.0 or SATA-600 on the board, it definitely does nothing. You still have no clue of what you are talking about. If you are talking about asymmetric CrossFire / SLI (X16/X4) or scaling down the bus in those setups symmetrically (X8/X8), I'd suggest doing more research on real-world performance figures, and as the link I've given shows, it's true X16/X16 for a two-card setup. You might show a counter-example of your $200 890FX board doing X16/X16/X16 tri-card bandwidth then, or just admit (including all of the above text) that you still have no clue of what you are talking about.

Re:Add a TURBO button;, You'll be more than satisf

[Cprossu]

Actually the turbo button was for decreasing performance so your "old" applications could run as they intended....

Just marketing got a hold of it and just like the old fridge sized IBM tape drives back in the day that were going to be designed with a "Idle" light instead were granted with a "Ready" light (remember IBM's were never..IDLE!), the "make the machine go slower button so your old stuff runs" became "Push this button and when the light comes on you're running the computer faster than if it wasn't on!"

OT: If Intel is a salty biker, Microsoft must be..

[mollog]

If you are comparing Intel is a salty biker, Microsoft must be a Mafioso.

I would have compared Intel to a street lawyer.

Re:amd boards have more pci-e lanes then the intel

[amazeofdeath]

Any references? I didn't think so. Do you even know what DMI is? Would DMI be relevant, how would you think it would be relevant to PCI-E lanes?

The guy has kind of a point

[ashpool7]

Just because there's a x16 slot doesn't mean all the lanes get hooked up. I looked at the manual for the motherboard you listed. This is the relevant section:

3 PCI Express 2.0 x16 slots
* if you intend to install two expanson cards into both PCI_E2 & PCI_E4 slots, these two PCIE x16 lanes wll auto arrange from x16/ x0 to x8/ x8
* the PCI_E5 supports up to PCIE 2.0 x4 speed
* all PCIE x1 slots will unavalable when an expanson card was installed into PCI_E5 slot.

2 PCI Express 2.0 x1 slots
* all PCIE x1 slots will unavalable when an expanson card was installed into PCI_E5 slot.

The only thing that is not clear is that if the PCI_E5 slot (the lower one) is disabled when the upper two (E2, E4) have two x16 cards in them. Taking the most optimistic view of the manual specifications, at maximum, there are only 16 + 4 lanes available. As to if eSATA and USB 3 eat up the lanes... that's another story.

Re:The guy has kind of a point

[amazeofdeath]

And how is this relevant to the GP claim of only 16 PCI(-E) lanes available to Intel CPUs? I don't mean to be disrespectful*, but your post is completely irrelevant: I can show LGA 1156 boards with a total of at least 34 PCI-E lanes in use, which just decimates the original argument. The "step-down" of PCI-E lanes in dual/triple card usage bears no meaning, as there are boards with more than 16 lanes in use, contrary to the original bogus argument.

*16+4 is still more than the original argument claims, even if you aren't paying attention to the other PCI-E slots.

Re:The guy has kind of a point

[ashpool7]

http://en.wikipedia.org/wiki/LGA_1156

Looks like it's 16+4. The +4 is channeled out of the PCH (http://en.wikipedia.org/wiki/Intel_P55). Therefore when you're using anything on the PCH (like USB and SATA), you're stealing bandwidth from the non-video PCIe slots, because it shares the same connection to the CPU. So he was a little off on his point about other interfaces stealing bandwidth from the video card. Turns out it only steals bandwidth from whatever you have in the third x16 slot or the x1 slot.

The Asus M4N98TD EVO can do 16+16+2 with an AMD chip. I think the point was that for the segment that this chip is targeting, it is severely hobbled in terms of PCIe bandwidth. These chips can only do x8/x8 SLI and AMD chips can do x16/x16 SLI.

I'd like to see a board that disputes that the 1156 socket can only handle 16+4 lanes.

Re:The guy has kind of a point

[amazeofdeath]

http://en.wikipedia.org/wiki/LGA_1156

Looks like it's 16+4. The +4 is channeled out of the PCH (http://en.wikipedia.org/wiki/Intel_P55). Therefore when you're using anything on the PCH (like USB and SATA), you're stealing bandwidth from the non-video PCIe slots, because it shares the same connection to the CPU. So he was a little off on his point about other interfaces stealing bandwidth from the video card. Turns out it only steals bandwidth from whatever you have in the third x16 slot or the x1 slot.

The Asus M4N98TD EVO can do 16+16+2 with an AMD chip. I think the point was that for the segment that this chip is targeting, it is severely hobbled in terms of PCIe bandwidth. These chips can only do x8/x8 SLI and AMD chips can do x16/x16 SLI.

I'd like to see a board that disputes that the 1156 socket can only handle 16+4 lanes.

Yes, looks like 16+4 or whatever, it's still more than the 16 claimed by original poster. As you did not bother reading the whole discussion, here's X16/X16 SLI: http://www.newegg.com/Product/Product.aspx?Item=N82E16813130258

And there's the board that disputes what you are asking for. Not to mention that if you know what you are talking about, true X16/X16 does not really make a difference in most common situations.

Re:The guy has kind of a point

[ashpool7]

Oh, I was reading it. Trying to figure out how they could do that. Turns out it's the nForce 200 chipset, which has a x16 connection to the boards, but the outgoing connection to the chip is still stuck at the speed of the DMI. So, really, it's just faking the available bandwidth and multiplexing the uneven split out to the cards. The available bandwidth to the cards is still 16+4 in terms of PCIe lanes.

A 2.0 PCIe slot is what, 8 gig/sec? Seems like AMD's HT link is about the same. I guess the deal is that without more HT/PCIe links on AMD/Intel chips (which would be FX/opteron chips with three links and socket 1366 chips) you're right, you won't see squat.

Re:The guy has kind of a point

[Hydian]

As has already been said, they work that magic by bridging an Nvidia NF200 that provides the additional lanes and that communicates over the same shared 2GB/s DMI as everything else. The P55 still only provides 16 PCI-E 2.0 lanes via the CPU and 8 PCI-E 2.0 lanes (that are signal rate limited) via the chipset. This limit is what potentially hurts USB 3.0 and SATA 6.0 cards.

In contrast, the 890FX has 42 PCI-E 2.0 lanes. 32 dedicated to graphics slots, 4 to a PCI-E x4 slot and the rest are for single x1 slots.

Block diagram of the P55 is found here: http://techreport.com/articles.x/17513
Block diagram of the 890FX is found here: http://techreport.com/articles.x/18825

As far as it making a difference...crossfire slot speed only makes a difference when you are using it, but USB 3.0 and SATA 6 can make a difference all of the time.

It is an unfair comparison. The 890FX is a top end board while the P55 is midrange.

Great, now I feel ripped off

[GameboyRMH]

- i7 940 owner

Re:Great, now I feel ripped off

[Anonymous Coward]

Don't confuse your feelings you're not a normal human, you're an intel customer, what you describe as "feeling ripped off" is the utmost nirvana available for experiencing in this mundane and AMD ridden unholy land. There, feel better?

Correctamundo/Agreed, 110% (with actual proof)

[Anonymous Coward]

"If you do computationally-intensive workstation tasks, like video editing, gaming, virtualization, or using java (sigh); you really will appreciate going from a $100 CPU to a $300 CPU. Using faster components also means having an overall less-frustrating experience with your computer." - by Lord Ender (156273) on Friday May 28, @11:30AM (#32376720) Homepage

Agreed, 110%, with evidence/proof thereof here too:

1.) My last system was an AMD X2 64 4800+, 2400mhz stock speed.

2.) My latest system utilizes an Intel I7 920 QuadCore, 2600mhz stock speed (iirc, & that's what I left it at too, summer time here & all).

VERY similar to yours too (I quote your listing of that below for a "compare & contrast" in fact)

To "justify my purchase", via an actual test?

Well...

I wrote up an application that deals HEAVILY in string processing (very, Very, VERY CPU-INTENSIVE work as I am sure you all realize), & wrote it in the fastest language for that I know of, in Borland Delphi 7.x with inline asm routines + heavy compiler optimizations. Purpose is to:

A.) Remove HOSTS file entries duplicates

B.) Alphabetize entries (for faster loads into sorted lists OR the local DNS cache (when I leave it on that is, can't with larger HOSTS files)

C.) Change the larger & slower 127.0.0.1, or even 0.0.0.0 blocking IP address entries into faster 0 based ones (works on Windows 2000/XP/Server 2003 & produces a 35% smaller file on average, which = faster loading file of course)

Best my AMD 4800+ could do the job in, with current dataset size (I still have that system)? 4 hours (over 1 million HOSTS files entries)...

My new Intel I7 920 Core system does the job in less than 2 hours time...

(And, it WILL/CAN do the job faster too (that is, if I don't let the app do it "all @ once" (meaning leaving a file in 1 million++ lines form, & instead, busting it into 8 smaller equally sized parts - takes only 24 minutes that way is why, & took me QUITE A WHILE to quit "beating my head" on profiling & doing inline asm routines that only gained me TINY %'s of gain in the end, vs. working on the dataset itself which gained me HUGELY as you can see above))).

So, do INTEL's I7 newest CPU's live up to their billing on PROCESSOR HEAVY/CPU INTENSIVE work-related tasks?

Absolutely!

(It's the "why" of why I shelled out for the I7 920 in fact, & it did what it said it would do, and it performs the string processing/cpu intensive work 4x as fast as my AMD did!)

"At home, I have an i7, an SSD, a high-end NVIDIA GPU, and the fastest RAM my mobo can take. At work I have a computer made of the budget components you think are good enough. The difference is extremely evident." - by Lord Ender (156273) on Friday May 28, @11:30AM (#32376720) Homepage

Amen/Agreed, 110%... see the above as my proof thereof in agreement with you, by ALL means!

APK

P.S.=> My system makeup, for your reference in comparison to how you designed yours:

1.) Intel Core I7 920 CPU (2600mhz stock speed)

2.) WD Velociraptor 10,000 rpm SATA II 300gb Disk #1

3.) WD Viking 10,000 rpm SATA II 150gb Disk #2

4.) Gigabyte IRAM 4gb Solid-State Drive (SSD) with "REAL RAM" on it (DDR2), not slower on writes FLASH RAM

5.) NVidia 8800 GTX OC/SuperClocked (by EVGA)

apk

Re:Correctamundo/Agreed, 110% (with actual proof)

[Bill, Shooter of Bul]

Interesting post.
Few questions unanswered by it:

What size (MB and #lines of host files) of a dataset did you actually have?

Why did you use Delphi 7, exactly? Its just an odd choice (from 2001), not universally proclaimed as the fastest in the world. Is it because its the fastest that you are familiar with? Its also obviously not 64 bit.

Why are you doing server work on a desktop? Typically that kind of a workload is performed on servers. It doesn't really require a gui, just brunt horsepower. It isn't something that most desktop users would be doing, not even power users like graphic designers or cad users. There is no argument that you could benefit from faster processors in a server. However, that is not what we are talking about here.

Why are you comparing the x2 4800 from 2005 to a top of the line processor from today? That's not a $100 processor of today, its a discontinued not even available in most web stores, processor. $100 gets you a amd phenom II 3.2 ghz processor today. See how much of a difference there is between that and your i7.

Confused...

[DarthVain]

The Core i5-655K and Core i7-875K have two and four cores, respectively, and they're priced at just $216 and $342. It appears that both will easily hit speeds in excess of 4GHz with air cooling.

I can buy a regular i5-750 quad core that will air OC to 3.6-3.8Ghz for about 215$ CND. So I can get a slightly faster dual core for the same price? yhippie.

So what is so special about these chips and why would I buy them? Anyway they don't seem all that special a deal. Unlocked multiplyer is neat, but the pricing VS performance I don't think makes any sense.

Same goes for the folks that buy a 200$ CPU and a 100$ cooler, then they could have just bought a 300$ CPU to begin with. Particulary when they mess with socket specs all the time making your previous cooling setup worthless.

Anyway if your doing it for fun whatever, it is. Just it doesn't make sense all the time. Getting as much out of your CPU just makes sense, but buying assuming OC is a bad idea, even with an unlocked multiplyer, there is nothing to say it will hit the 4+Ghz mark, and if you don't nothing you can do about it.

Re:I haven't forgotten the pentium 90 math bug fia

[cyberjock1980]

AMD never had the TLB bug...

http://en.wikipedia.org/wiki/AMD_Phenom

I have used Intel exclusively since I could build my own machineat 13 years old, and will continue to do so.

Re:Whats the point?

[Anonymous Coward]

Note to mods: Flamebait != I disagree...

oops

[ashpool7]

I mean 2.0 x16 PCIe slot.

Re:Splurging

[jvin248]

I splurged 12 years ago on a $2500 pc and used it until I upgraded five years ago to a then two year old desktop. The new machine was being scrapped out because the heatsink mounts broke and it would overheat. Easy repair. Used this machine until a month ago. The replacement is a free Xeon Quad-core with a failed HDD - easily repaired and has provided great uptime.

Of course, I squeeze a lot more useful work out of these machines because they run Linux, without cycle-sapping anti-virus programs.

o

It reeks of desperation

[mrmeval]

AMD must really be worrying them.

Why Delphi 7? SPEED IN STRINGS & MATH! apk

[Anonymous Coward]

"What size (MB and #lines of host files) of a dataset did you actually have?" - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

A custom HOSTS file I've been building since 1997, and when I "merge in" new entries from other sources, it usually 'shoots up' to around 1 million lines to go thru, weighing in @ approximately 24mb in size (using 0.0.0.0 as the initial blocking address format).

----

"Why did you use Delphi 7, exactly?" - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

First, because I like and am very familiar with Object Pascal (& lately, I've been "checking out" LAZARUS (IDE for "Free Pascal"), because it does 64-bit binaries, & I am 'slowly but surely' making the move to that here.

I actually program fluently in 12 computer programming languages, but for the PC, in 32-bit? I like Delphi the most, even to this very day, because the reasons I noted in my last post, & again here now.

(Additionally, as I mentioned in my last post to which you replied, Borland Delphi 2.0 beat the snot out of my former favorites in VB5/6 and even MSVC++ 6.x in 7/10 tests performed on various operations, but most of all, because Delphi's Object Pascal 7.x code was shown DOUBLING (and doing far more than that to VB5, even though it had a 'watered down' MSVC++ 5.x compiler engine) MSVC++ in math & strings work (which every program does work in mind you - but the one I am working on here & noted in my last post does EXTENSIVE work in strings comparisons in)!

----

"Its just an odd choice (from 2001), not universally proclaimed as the fastest in the world. Is it because its the fastest that you are familiar with? Its also obviously not 64 bit." - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

Nope, again, for 64-bit coding in PASCAL, I use LAZARUS &/or FPK 2.2 or 2.4... or, of course, Microsoft Visual Studio... but, per what I wrote above, & from as far back as 1997? Delphi does an especially good job of strings and math work, & is a really nice 32-bit IDE and programming language!

(Again - Delphi quite literally doubles the speed of code it produces, over that from MSVC++ typically in those areas)

----

"Why are you doing server work on a desktop? Typically that kind of a workload is performed on servers." - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

I am not sure what gave you the impression that I was doing "server work" on a desktop, but... I am not, so you know (just my personal workstation @ home is all this system is...).

----

"It doesn't really require a gui, just brunt horsepower." - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

I have contemplated re-writing this app in pure console/tty terminal/dos prompt (character mode), and I actually have done parts of it this way, & I saw a 10-fold increase in its speed (which is typically, less "weight" & message-passing overheads are present when you do a console-mode app)... but, I like GUI. I built it for myself, and a few of my pals use it also, so GUI was "the way to go" is all (easier for them to use).

----

"It isn't something that most desktop users would be doing, not even power users like graphic designers or cad users. There is no argument that you could benefit from faster processors in a server. However, that is not what we are talking about here." - by Bill, Shooter of Bul (629286) on Friday May 28, @05:29PM (#32382404)

I generally do MOSTLY coding work here (programming), & the "more power the better"... in CPU, & disk especially (a lot of what I have been doing lately is "file I/O based" in nature, so the disks matter here too, & of course, string processing is MOSTLY ALL CPU!)

Re:NO 1336? so you are stuck with 16 pci-e lanes s

[yanyan]

You said it man. It simply isn't 1337 enough.

I7-875K places 17th in price/performance

[dave87656]

the i7-875K offers better performance and power efficiency per dollar than just about any other desktop CPU out there."

Yep, coming in near the top at .... place 17. http://techreport.com/articles.x/18988/2

The author must be an Intel fan. Almost all the best price/performance numbers are in AMD's offerings.

Does he get a kick back?

[kenp2002]

Seriously is kdawson getting kickbacks for posting this crap. Since the "ignore kdawson" filter has been broken I find the need to keep score on bullshit he keeps posting and he's in the red for my bullshit-o-dometer.

It's like some smarmy marketing prick sends a submission to kdawson and says "there's 200 bucks in it for ya if you post it."

I mean seriously Taco, you need to look into this shit...