HEXUS.reviews :: Intel Core 2 refreshed - QX6850 and E6750

[Tarinder]

Today, Intel is launching a revised range of mid-market-to-high-end Core 2 processors, still 65nm but with a faster front-side bus. Are they fast enough to get CPU lovers everywhere reaching for their credit cards? And how do they stack in terms of value after AMD's latest round of price-cuts? We look at the QX6850 and E6750 to find out.

Check it out.

[nugation]

There are a couple of things I don't agree with in this article.

First is the pricing. It's a little tough trying to figure out what the price will be in the UK compared to the US but not impossible. Your "guess" at the retail price seems to be WAY off to me. I searched computerprices.pricegrabber.co.uk and could only find one seller and their price wasn't reasonable so I think we should ignore it as price gouging.

The lowest US price on pricegrabber for the 6000+ is 169.00 and 106.42 in the UK. That's a conversion of .6297. If we compare a more readily available part like the e6700 we find a US price of 314.00 and a UK price of 192.99 for a conversion factor of .6146.

The $183 price you listed for the e6750 is for OEM parts which you can't buy (legitimately) from retail. So if we take the conversion factor from the e6700 and apply it to the US price of the e6750 I think we get a more realistic picture of what to expect the price to be.

Another pricing issue I have is the assumption you make about where the e6750 will land relative to the 6000+. Currently, in the US the e6750 is available for 26% more than the 6000+. You also fail to mention the possibility that AMD may further reduce the price of the 6000+ in response. You also claim that, because of the price drop, the most direct competitor to the 6000+ will be the e6600. After reviewing your review of the 6000+ which (in some cases) compared it with the e6600 this appears to be true from a performance standpoint.

However, the e6600 currently costs 33% more than the 6000+ - yes- even more than its big brother the e6750. This is going to take some fancy footwork from Intel to get the pricing stabilized given that it is already upside down. Given all that it seems to me that your “assumption” is unfounded at best.

The other main issue I have is the award you give it. It seems that either you don’t take your awards very seriously or they don’t really mean much. Take a look at your own benchmarks. All of them, both individually and as a whole. The gaming benchmarks show both CPUs to be producing more-than-perceivable frames rates. That is a tie. There is no benefit to producing more frames per second than a human being can see. The WAV and DivX encoding benchmarks… The 6000+ took 28 seconds longer than the e6750. 28 seconds. Those people that compute pi for a living out to 10M places day after day after day will certainly be excited to find out they can save 6.84 seconds every time they do it. I mean, seriously, could you have found a less relevant gauge of performance?

If you look at the big picture the e6750 really begins to show its true colors. If you compute the percentage of difference between the e6750 and the 6000+ for EACH of your benchmarks – including the power consumption – and then average those out something interesting shows up.

On average, across all the benchmarks you ran, the e6750 outperforms the 6000+ by 12.17%. Now, understanding that you used, as a whole, one of the worst compilations of tests I’ve seen in quite some time I think it reasonable to include someone else’s results for comparison. That’s only fair isn’t it, since you are comparing the products of two competitors why not have a taste of your own medicine? Anandtech’s average across all their tests works out to be 9.37%. That’s pretty close so no harm no foul.

But what does that 2.8% margin of error mean to a consumer, you remember them don’t you? They are the people that look at the “recommended” award at the end of your review and think- ah, must be ok, Hexus recommends it. If they ignore you and buy the 6000+ it would cost them US 4.73 or US 6.38 if they went for the e6750. If I based my purchasing decision on your “recommended” award I would ask for a refund. But I digress; my purpose here is not to point out how your ridiculous test suite and award system is.

The bottom line is that you are correct in that the e6750 outperforms (sometimes) the 6000+ in various benchmarks. But at what cost? Does that not count for anything? Don’t get me wrong, the performance crown is fun to wear and I certainly enjoy it when AMD wears it but that is not the whole picture. Life is NOT solely about performance or we would all be driving cars made in Italy.

In order to get the 12.17% more performance you claim the e6750 will get you (on average) you have to pay 26% more for it. That is a difference of US 58.99. So let’s turn that $60 into something real. If you’re a gamer that’s a free badass heat sink, the next step up in power supplies or about 5 cases of beer. If you’re a business that’s buying three new computers you would save $176.97 or if you’re a larger business buying 1000 systems you would save $58,990. This doesn’t take into consideration lower infrastructure (mobo and other supporting items) costs typical of an AMD system.

Speaking of power… What if you try to justify the additional cost by the amount of money you will save in power? I went to the energy information administration’s website to get some power costs; here’s what I found. The national average cost for power is US .1017 per kilowatt hour for residential use. Commercial power is cheaper (go figure). If you run your PC at full load 24/7 for a whole year you would save US 75.73 using the e6750. If you simply idled for a year you would save US 13.36. A 50/50 split between max and idle would save you US 44.54. The most realistic figure is more like 25% full load and 75% idle in my opinion. That would render a total yearly savings of US 28.95. This ultimately means that after, on average in the US and using your computer at full load 25% of the time and idle the rest, you would have to keep your e6750 two years and thirteen days to break even.

Ultimately my point is that this review doesn't provide a very clear picture of what's going on. It doesn't provide much relevant data one could use to make a purchasing decision. Assumptions are made with no explanation of the basis. Price is discussed but very one-sided. You make claims and assumptions about Intel's forward pricing but ignore AMD. Repeatedly, throughout the article, you "omit" data that creates a "feel" that the e6750 is better than the 6000+ but better is a far different term than out performs.

The summary is a train wreck. You spend more time talking about what might be possible and what may happen in the future than actually summarizing the article. You claim that you, if you were looking for a quad core, wait until after the price drop but a little further down you claim this is a golden time to buy a CPU. Well which is it? Now or wait?

The second to that last paragraph really buries you. "Right now, though, there's exceptional value to be had..." It is this sentence that got me started on this rant in the first place. All of my complaints here have revolved around this statement. The operative word in it is value. Value is defined as the ratio of performance to cost when the term is applied to a product. So I agree that both the e6750 and the 6000+ both have value but, in the last paragraph the product with less value is "probably" chosen. The award goes right along with this line of non-thinking.

Strong performance at a competitive price is what the award says. But I say nay-nay. It's such a simple thing. Go to the grocery store and by some food. Some of the items you get will be popular name brands some will not. You make your decision based on how bad the food will taste vs. how little the food costs. The performance of these two parts is very similar, on average a difference of 12.17% or 9.37% depending on who you ask. The performance of both parts doesn't render any software unusable or unplayable and the performance differences are measured in seconds. The issue you completely ignore is the extra 26% you have to pay to get that 9% or 12% of extra performance.

In its simplest form you reduce the performance of the e6750 by 12% making it equal to the 6000+. Reduce the price by the same amount you did the performance and you will find yourself paying an additional 14% in cash. My question is for what? You would "probably" recommend people spend (no matter how you measure it) more money for the e6750 than the 6000+ and fail to provide a single reason that is related to the reason they got the award. In a word, value.

Please explain to me where I can find the missing 14% of my money in an e6750.

[PD HEXUS]

Originally Posted by nugation

...me... ...rant...

Webster's Revised Unabridged Dictionary - Cite This Source
Nugation

\Nu*ga"tion\, n. [Cf. OF. nugation.] The act or practice of trifling. [R.] --Bacon.

Webster's Revised Unabridged Dictionary, © 1996, 1998 MICRA, Inc.

[nugation]

Yup, you caught me. I'm a trifler. LOL.

[James Thorburn]

A lot to cover here so excuse me if you think I've missed anything out.

Originally Posted by nugation

The lowest US price on pricegrabber for the 6000+ is 169.00 and 106.42 in the UK. That's a conversion of .6297. If we compare a more readily available part like the e6700 we find a US price of 314.00 and a UK price of 192.99 for a conversion factor of .6146.

The $183 price you listed for the e6750 is for OEM parts which you can't buy (legitimately) from retail. So if we take the conversion factor from the e6700 and apply it to the US price of the e6750 I think we get a more realistic picture of what to expect the price to be.

Pricing for the E6750 was taken from pre-order prices we found on a UK retailer.
The E6700 should drop to match the E6750's price on the 23rd of this month.

Originally Posted by nugation

Another pricing issue I have is the assumption you make about where the e6750 will land relative to the 6000+. Currently, in the US the e6750 is available for 26% more than the 6000+. You also fail to mention the possibility that AMD may further reduce the price of the 6000+ in response. You also claim that, because of the price drop, the most direct competitor to the 6000+ will be the e6600. After reviewing your review of the 6000+ which (in some cases) compared it with the e6600 this appears to be true from a performance standpoint.

We use UK retailer pricing, which at the time of writing were as stated in the article.
AMD dropped prices on the 9th and have already denied they are going to be dropping prices further.

Originally Posted by nugation

However, the e6600 currently costs 33% more than the 6000+ - yes- even more than its big brother the e6750. This is going to take some fancy footwork from Intel to get the pricing stabilized given that it is already upside down. Given all that it seems to me that your “assumption” is unfounded at best.

Intel's price cuts come on the 23rd, and details of these have been circulating on the rumor mill for months. Things should fall into line .

Originally Posted by nugation

The gaming benchmarks show both CPUs to be producing more-than-perceivable frames rates. That is a tie. There is no benefit to producing more frames per second than a human being can see.

We are using these games to test the raw performance of the CPU, and therefore run at just 1024x768 with reduced detail level. This allows us to stress the CPU, to highlight the differences in performance. Even in these circumstances we are able to hit a GPU limit on the Core 2 CPUs.
While it is true that rendering above 60fps will be outside the realms of what an LCD will display, the frame rate is not a constant. It will fluctuate with the amount of action on screen, and while a higher average does not guarentee a higher minimum frame rate it can give some kind of indication.
Unfortunately our testing suite doesn't feature games which give their minimum and maximum frame rates at present.

Originally Posted by nugation

The WAV and DivX encoding benchmarks… The 6000+ took 28 seconds longer than the e6750. 28 seconds.

Given that both these tests take roughly 2 minutes to complete on the E6700, 28 seconds is a fairly substancial increase, roughly 25%.

When I encode DVDs I've bought for my Media Centre I run a two-pass, highest quality encode, with the whole process taking in the region of 4 hours (for a longer film).
A 25% increase in time would see the whole process take an extra hour to complete. This isn't a small, intangible, difference.

While we would love to test CPUs by encoding an entire movie this simply isn't practical in any kind of realistic time-frame.

Originally Posted by nugation

Those people that compute pi for a living out to 10M places day after day after day will certainly be excited to find out they can save 6.84 seconds every time they do it. I mean, seriously, could you have found a less relevant gauge of performance?

While it may not be directly relevant to how people use their PCs, PiFast is a quick and easy way of comparing two CPUs.
We would not recommend a CPU solely off the back of a strong PiFast result anymore than we'd recommend one for scoring well in ScienceMark, but they are popular with our readers for showing the differences between architectures.

Originally Posted by nugation

If you look at the big picture the e6750 really begins to show its true colors. If you compute the percentage of difference between the e6750 and the 6000+ for EACH of your benchmarks – including the power consumption – and then average those out something interesting shows up.

On average, across all the benchmarks you ran, the e6750 outperforms the 6000+ by 12.17%. Now, understanding that you used, as a whole, one of the worst compilations of tests I’ve seen in quite some time I think it reasonable to include someone else’s results for comparison. That’s only fair isn’t it, since you are comparing the products of two competitors why not have a taste of your own medicine? Anandtech’s average across all their tests works out to be 9.37%. That’s pretty close so no harm no foul.

Having taken a quick look at Anandtechs, and our own, benchmarks, those percentage differences are very similar if you exclude the game tests.
Anandtechs overall percentage difference appears to be reduced by virtue of them having a greater number, and greater proportion, of game tests in their benchmarking suite. These, as with ours, show little variance between CPUs due to GPU limitations.

Originally Posted by nugation

But what does that 2.8% margin of error mean to a consumer, you remember them don’t you? They are the people that look at the “recommended” award at the end of your review and think- ah, must be ok, Hexus recommends it.

We do recommend the Core 2 Duo E6750, it does, in our opinion, represent the best value for money of the CPUs we looked at in this review. This is based off of a combination of benchmark results, price, power consumption, and overclocking potential.
We do not make this recommendation unconditionally however. If you already have a Socket AM2 motherboard then the AMD Athlon 64 X2 6000+ makes a great upgrade at a strong price, but given a clean sheet for a system build we would spend the extra £15 on an E6750.

Originally Posted by nugation

If they ignore you and buy the 6000+ it would cost them US 4.73 or US 6.38 if they went for the e6750.

Errrrrm. I must be missing something here but I don't get what you mean here.

Originally Posted by nugation

If I based my purchasing decision on your “recommended” award I would ask for a refund. But I digress; my purpose here is not to point out how your ridiculous test suite and award system is.

The bottom line is that you are correct in that the e6750 outperforms (sometimes) the 6000+ in various benchmarks. But at what cost? Does that not count for anything? Don’t get me wrong, the performance crown is fun to wear and I certainly enjoy it when AMD wears it but that is not the whole picture. Life is NOT solely about performance or we would all be driving cars made in Italy.

In order to get the 12.17% more performance you claim the e6750 will get you (on average) you have to pay 26% more for it.

Given the price differential (roughly £12 at the time of writing) we have no qualms recommending the Core 2 Duo E6750 over the 6000+.
The E6750 works out 11.11111 (re-occuring) more expensive than the 6000+. Given that in some benchmarks it is over 30% faster than the 6000+, and does not fall behind in any of our benchmarks other than theoretical memory tests.

As previously stated pricing is a bit of a quagmire at the moment, AMD are fresh off the back of a price cut while Intels is less than a week away. Give it a week (co-inciding with the E6750's retail availability) and things will make more sense.

At this point we'd hold off purchasing any Core 2 parts until the 23rd, unfortunately Intel pulled in the NDA for these parts and weren't clear on the date of launch or price drops.

Originally Posted by nugation

Speaking of power… What if you try to justify the additional cost by the amount of money you will save in power? I went to the energy information administration’s website to get some power costs; here’s what I found. The national average cost for power is US .1017 per kilowatt hour for residential use. Commercial power is cheaper (go figure). If you run your PC at full load 24/7 for a whole year you would save US 75.73 using the e6750. If you simply idled for a year you would save US 13.36. A 50/50 split between max and idle would save you US 44.54. The most realistic figure is more like 25% full load and 75% idle in my opinion. That would render a total yearly savings of US 28.95. This ultimately means that after, on average in the US and using your computer at full load 25% of the time and idle the rest, you would have to keep your e6750 two years and thirteen days to break even.

In the UK electricity is significantly more expensive, roughly double the price compared to the US.
Of course even ignoring this there are the environmental implications. I'm not going to get preachy on this as I'm hardly the most environmentally friendly person but cost is not the only reason to advocate an energy efficient approach.

Originally Posted by nugation

Ultimately my point is that this review doesn't provide a very clear picture of what's going on. It doesn't provide much relevant data one could use to make a purchasing decision. Assumptions are made with no explanation of the basis. Price is discussed but very one-sided. You make claims and assumptions about Intel's forward pricing but ignore AMD. Repeatedly, throughout the article, you "omit" data that creates a "feel" that the e6750 is better than the 6000+ but better is a far different term than out performs.

Intel's pricing data for the 23rd has been available for some time, AMD have denied they are going to drop prices further, especially since they dropped them less than 2 weeks ago.

Originally Posted by nugation

The summary is a train wreck. You spend more time talking about what might be possible and what may happen in the future than actually summarizing the article. You claim that you, if you were looking for a quad core, wait until after the price drop but a little further down you claim this is a golden time to buy a CPU. Well which is it? Now or wait?

Our comments were made from the perspective of after the price cuts, which given the previous comment should have been implied. Intels lack of communication over whether this had been pulled in with the NDA led to some confusion here, but things could have been re-jigged to make things clearer.

Originally Posted by nugation

In its simplest form you reduce the performance of the e6750 by 12% making it equal to the 6000+. Reduce the price by the same amount you did the performance and you will find yourself paying an additional 14% in cash. My question is for what? You would "probably" recommend people spend (no matter how you measure it) more money for the e6750 than the 6000+ and fail to provide a single reason that is related to the reason they got the award. In a word, value.

Please explain to me where I can find the missing 14% of my money in an e6750.

Being a UK based site we worked from UK pricing, which shows an 11.1111% pricing differential.
While the mean performance differential is in the region of 9-12% this hides the large performance advantage the E6750 has shows in some tests.
If you focus on media encoding tests the E6750 is roughly 23% faster than the 6000+ and these are the kind of real-world, time consuming tasks where the Core 2 Duo does show its true value.

Add in the 6000+'s poor performance in the multi-threaded tests, the Core 2 Duo completes the DivX encoding while playing Quake 4 in almost the same time as the Athlon 64 X2 6000+ takes running DivX alone, and we can see the E6750 does offer obvious performance benefits. Enough, in our opinion, to justify the extra price.