Maximum PC Core 2 RSS Feed

Intel's Core i5 Analyzed and Tested (150+ Benchmarks)

Gordon Mah Ung — Tue, 08 Sep 2009 10:00:00 -0500

Intel’s latest troika of new CPUs brings Nehalem goodness to the masses

Nehalem for everyone! That simple sentence best explains Intel’s brand-new series of CPUs, which is sure to please budget users everywhere while confounding power users.

Why would a new CPU that gives you the best bang for the buck in town be greeted nervously? Because Intel’s new CPU brings with it a new socket as well as a new infrastructure. This new infrastructure is essentially a fork in the road that forces users to make a difficult choice: Save money today but get locked out of the high-end, or splurge today knowing that the budget CPU is damn near as good as the top-end part.

For the details on Intel’s new budget monster, savor our full report, consume the specs, and then digest the benchmarks to see just which path your next PC should take.

Meet Lynnfield

We normally don’t use a CPU’s code-name once its real name is known, but to help keep your head from popping off over Intel’s confusing naming scheme we’re going to rely on some code-names here for clarity.

Intel’s newest CPU family is code-named Lynnfield. The lineup includes the 2.93GHz Core i7-870, the 2.83GHz Core i7-860, and the 2.66GHz Core i5-750. Lynnfield chips use essentially the same microarchitecture as Intel’s original Core i7 CPUs, which were code-named Bloomfield, but the new CPUs are incompatible with existing Core i7 motherboards. That’s right, you could walk into a store and buy a Core i7 CPU that will not work with the Core i7 motherboard you just bought. Likewise, the Core i7 heatsink cooler you bought may not work with a new Core i7, either.

Despite its smaller size, Intel’s new Lynnfield Core i7/Core i5 CPU (right) actually features an increased transistor count of 774 million and a larger die size of 296 square millimeters, compared with the 731-million-tranny Bloomfield Core i7 (left) and its 263 square-millimeter die.

Socket Switcheroo

The most notable difference in this new crop of Core i7s is the socket. For Lynnfield, Intel is introducing the LGA1156 socket. This socket is, as stated, incompatible with the current LGA1366 motherboards and CPUs. To irk you even more, even the heatsink cooler mounting holes are incompatible, so you probably couldn’t use an LGA1366 cooler, even if you happened to have one. And even more annoying to enthusiasts, LGA775 coolers are also incompatible. Earlier this year, Intel execs told Maximum PC that LGA775 cooler compatibility was being considered for the new chip, but obviously the company has since ruled that out. To break it down: LGA1366 uses a 12cm gap, LGA1156 uses an 11cm gap, and LGA775 uses a 10cm gap. Yes, one centimeter difference and you have to dustbin your pricey high-rise cooler even though it’s capable of handling the thermals of the new chip.

We asked Intel if it was doing this just to piss people off and the company said no, it did it for legitimate engineering reasons. Intel actually lowered the height of the new direct socket load mechanism that clamps the CPU in place, which required moving the mounting holes out. Existing heatsinks capable of the thermal load should work, Intel said, so long as consumers obtain updated mounting brackets from the cooler maker. We have to also note that very new high-end coolers are coming with mounts for LGA1156 too. Still, make sure that if the box says Core i7, support for LGA1156 is included.

Although Intel wouldn’t confirm this, we’ve been told by high-end system builders that certain LGA1366 motherboards and coolers would flex enough to create a gap between cooler and CPU. The new design presumably fixes that problem.

Sockets Compared

Lynnfield Socket

Bloomfield Socket

The new budget Core i5/Core i7 CPUs use a new LGA1156 socket design (top image), which is incompatible with existing LGA1366 Core i7 motherboards and coolers (bottom image). The design eliminates traces for the third memory controller, but builds in wires that will accommodate forthcoming CPUs with integrated graphics cores.

Memory Loss

Enthusiasts will also question the move from tri-channel DDR3 to dual-channel DDR3—why go back if wider is better? Intel’s decision is based on pragmatism and cost. The tri-channel circuitry in the CPU doesn’t add much cost to the processor, but it’s not cheap to implement when building a motherboard. Those added traces from the socket to the RAM slots mean more layers and pricier boards. That’s one of the contributing factors to excessively priced X58 boards this past year.

Should you be concerned about shifting down to dual-channel? Generally, no. For the most part, only the most memory-bandwidth-intensive apps will actually see any performance hit. The fantastic latency and overall bandwidth capacity of the Nehalem design is more than adequate for today’s applications. This doesn’t mean it’s a non-issue. The vast majority of Lynnfield motherboards we’ve seen are opting for four-DIMM slots. That means a maximum of 8GB using affordable 2GB DIMMs (4GB DIMMs are currently cost-prohibitive). The only board we’ve seen with more DIMM slots is Gigabyte’s GA-P55-UD6, which features six, for a total of 12GB, using 2GB DIMMs. The memory controller in Lynnfield tops out at 16GB, while Bloomfield maxes out at 24GB. Realistically, 24GB of RAM is way overkill for 99 percent of us. Our experience has shown us that most apps do not consume that much RAM; 4GB to 6GB is the sweet spot today.

PCI-E at the Core

Another significant change for Lynnfield is that PCI-E comes directly off the CPU core. In X58/Bloomfield, X48/Core 2 Quad, and 790FX/Phenom II, PCI Express is external to the CPU, residing in the north-bridge chipset. Data is pumped out to the chipset where it must then be directed to the PCI-E slots that connect to the GPUs. As you can imagine, the extra hop creates a latency issue—that’s eliminated with Lynnfield. Lower latency is better, right? Yes and no. While latency is better on Lynnfield, the CPUs feature but a single x16 PCI-E 2.0 lane in the chip. That’s fine for a single GPU, but multi-GPU configurations will have to split the bandwidth. That means an SLI or CrossFire X rig will run both cards in x8 mode. Remember, however, that we’re talking x8 at PCI-E 2.0 speeds, which is 8GB/s for each card. From what we’ve seen and from what vendors have told us, only at the highest resolutions with antialiasing do you see any impact.

But, you say, what about tri-SLI? That, too, can be theoretically supported. Even though Lynn-field supports but a single x16 PCI-E 2.0 lane, additional PCI-E lanes are added through the P55 chipset—at the cost of latency. Previous designs that had PCI-E plumbed from both the north bridge and south bridge were dinged for doing just this. Some vendors are working around the lack of bandwidth by adding Nvidia’s nForce 200 chip to their boards. The nForce 200 doesn’t magically add bandwidth but it does manage the available bandwidth across multiple slots. Thus, a board vendor could add three or even four x16 PCI-E slots and have the nForce 200 chip manage the load for multi-GPU configurations. This would eliminate the need to have the GPUs feed off of the higher-latency connection in the chipset.

CrossFire X support is a given in the vast majority of P55 motherboards. And SLI will be as well for the board vendors who pay Nvidia to “certify” that their boards are SLI-ready. Fortunately, the big names are already onboard with that, including Asus, Gigabyte, MSI, EVGA, and even Intel.

If you’re wondering why Intel doesn’t just fix the lack of bandwidth by adding additional PCI-E lanes in the CPU, consider this: Intel would have to add additional traces from the CPU and the socket, and even Intel is loathe to introduce yet another new socket standard so soon. The company also doesn’t think it’s worth it as anyone who truly wants full dual-x16 GPU support should really be buying X58: Remember, folks, Lynnfield and P55 are for the “mainstream.”

The Lynnfield LGA1156 again rejiggers the design of the modern CPU. With the Athlon 64 and the Core i7, the memory controller was moved from the north bridge directly into the CPU core. With Lynnfield, the PCI-Express ports are now handled directly by the CPU as well. Furthermore, relatively low-speed data from the hard drives and USB ports flows though a single DMI link at 2GB/s with Lynnfield.

P55: Downsizing the Chipset

In the old days, new core-logic chipsets were almost as exciting as a new CPU. Not so today. With X58 and Bloomfield, the chipset got fired from its job of managing the memory controller. With Lynnfield, it even loses its responsibilities for managing PCI-E, as both features are now integral to the CPU itself. Today, the chipset is nothing more than a glorified south bridge, managing USB, SATA, PCI, PS/2, and other decidedly unsexy items. The P55 might have been sexier if it had USB 3.0 or SATA 6 functionality, but that won’t appear until next year. And even if it had those features, it’s pretty clear where the P55 stands: Its desk has been moved to the basement and its red Swingline stapler has been confiscated.

Tweakability

There was some initial confusion when Bloomfield was introduced. Early reports had it that the memory multipliers were unlocked on the high-end Extreme versions of the CPU, but locked on the lower-end versions. It turned out that was only the case for samples first sent to the press and system vendors. Retail versions of the lower-end Bloomfields were unlocked as well, making it possible for users to set the RAM at speeds higher than the rated DDR3/1066. This time around, there’s no such confusion. All three new Lynnfields feature unlocked memory multipliers and the chips are actually officially rated for operation at DDR3/1333, with higher speeds obtainable through “overclocking.” Of course, the chips are upwardly clock-locked, so you can’t simply set your $200 Core i5 to run at 4GHz by changing the multiplier.

Though open about memory locks, Intel has been cagier about Turbo mode. With Bloomfield, Intel never revealed to consumers the top clock speeds the CPUs could hit under Turbo mode, but with Lynnfield, it’s now publicizing the top speed that a single core can hit. The Core i7-870, for example, can top out at 3.6GHz by overclocking individual cores based on the thermals and power consumption of the chip. That’s actually far greater than the Bloomfield modes can top out at on default. The take-away is that, with the months it has had to tinker with Lynnfield, Intel has improved Turbo mode (now officially Turbo Boost). However, the feature is locked. Only on the Extreme parts will users be able to tinker with Turbo beyond the default caps.

The Big Decision

All this techno-speak is meaningless if the processor doesn’t perform as expected. We don’t want to give away everything here but let’s just say we’re not disappointed. Lynnfield is everything you’d expect of a new iteration of Nehalem and Intel’s now-very-mature 45m process. As such, overclocking, is also fruitful. By simply boosting the base clock of the cheapie $200 Core i5-750, we were able to take it from 2.66GHz to a very stable 3.5GHz without any additional voltage on our very first attempt using a Gigabyte GA-P55-UD6 board.

System builders have been equally impressed with these budget parts and have achieved overclocking results every bit as good as, if not better than, the most expensive Core i7-975 Extreme Edition parts.

That brings us to the main question: Why even build a Bloomfield LGA1366 at this point if building a Lynnfield system will save you at least $100 on the board and a little on the RAM, and even more money if you opt for the ultra-budget Core i5-750?

Here’s that fork in the road: Lynnfield is cheaper and gets you 90 percent the performance of a Bloomfield system, but early next year Intel will introduce a CPU code-named Gulftown, aka Core i9. Core i9 adds two more physical cores to the CPU and will likely be the first consumer hexacore CPU. With Hyper-Threading, that’s 12 threads available to the OS and enough to make the most jaded enthusiast perk up. Core i9, however, will only be available on the Bloomfield/LGA1366 platform. If you were to build a Lynnfield LGA1156 box there’d be no six-core for you! At least, not at this point. Intel said it has no plans for an LGA1156 hexacore. You see the dilemma. Save money now and build a really kick-ass LGA1156 or spend the extra $200 to build an LGA1366 that has an easy upgrade path to six cores with Hyper-Threading. It’s not an easy choice to make under normal circumstances, but in this economic climate, it’s even harder—that $200 goes a long way toward a better GPU, better PSU, more RAM, or a bigger hard drive. The choice, however, is up to you.

Next, the CPU buyers guide and benchmarks!

CPU Buyers Guide

What’s the best budget chip available today for those interested in getting good performance on the cheap? We’ll walk you through the top five chips and tell you which one to buy.

CPU Specs

CPU	Core i7-920	Core i7-870	Core i7-860	Core i5-750	Core 2 Quad Q9550/Q9550s	Phenom II X4 965 BE
Socket	LGA1366	LGA1156	LGA1156	LGA1156	LGA775	AM3
Price (Volume)	$284	$562	$284	$196	$266 / $320	$245
Price (Street)	$280	N/A	N/A	N/A	$220 / $350	$245
TDP	130Watts	95Watts	95Watts	95Watts	95Watts / 65Watts	140Watts
Codename	Bloomfield	Lynnfield	Lynnfield	Lynnfield	Yorkfield	Deneb
QPI/HT	4.8GT/s	4.8GT/s	4.8GT/s	4.8GT/s	N/A	4GHz
Core Clock	2.66GHz	2.93GHz	2.8GHz	2.66GHz	2.83GHz	3.4GHz
Turbo Boost (Max 1 Core)	2.93GHz	3.6GHz	3.46GHz	3.2GHz	N/A	N/A
HyperThreading?	Yes	Yes	Yes	No	No	N/A
Cores/Threads	4/8	4/8	4/8	4/4	4/4	4/4
L1 Cache	256KB	256KB	256KB	256KB	256KB	512KB
L2 Cache	1MB	1MB	1MB	1MB	12MB	2MB
L3 Cache	8MB	8MB	8MB	8MB	N/A	6MB
Die Size (mm^2)	263	296	296	296	214	258
Transistor Count (million)	731	774	774	771	820	758
Process (nm)	45	45	45	45	45	45

AMD 3.4GHz Phenom II X4 965 BE

It’s bad news for AMD’s recently released quad-core. Its best and brightest could never compete with the Core i7-920 in performance, but it certainly outgunned it in price. Now with Intel’s LGA1156 Lynnfields here, it can’t even compete on price. As of this writing, the Phenom II X4 965 Black Edition had a list price of $245. The Core i5-750 has a list price of $200 and it mercilessly punts the Phenom II X4 965 BE up and down the field. Only in the mostly-single threaded Photoshop CS3 and POV Ray 3.7 did the Phenom’s 700MHz advantage put it over the top. Against the Hyper-Threaded Core i7’s though, the virtual cores vaulted the Nehalems over the Phenom II X4 965 everywhere. The good news for the Phenom II X4 965 is that it isn’t dead ass last. That falls to its old nemesis, the 2.83GHz Core 2 Quad Q9550. There, the Phenom II X4 965 BE serves space-cold revenge to the only chip here not to sport an on-die memory controller. Yeah, so you Intel fan boys say so what? Core 2 is an end of life CPU big whoop, but you gotta take your victories as you can get them.

The real bad news for AMD is that it’s not expected to get its next-generation Bulldozer core out now until 2011 so it can build it on a 32nm process. With Core i5 pushing the $200 range and Core i3 around the corner and expected to push into the ultra-budget range, there’s going to be very little maneuvering room for AMD’s rather dated design.

Still, there is place for the Phenom II X4 965 BE: existing upgrades. The processor will drop into a large assortment of existing AM2+ boards (you should check your mobo maker’s web site first for support) and even though it’s slower than a Core i5/i7, it’s a hell of a lot cheaper and easier to remove your current Athlon 64 X2 part and drop in the Phenom II X4 965 BE. Once you do that, you can thumb your nose at the Intel boys who have to replace their LGA775 boards and likely do an OS reinstall to upgrade to Core i5 and Core i7. Heck, there’s even rumors of a six-core AM3 part which would likely drop into modern AM2+ board so there’s some solace for AMD fans even if the performance isn’t there.

Intel 2.83GHz Core 2 Quad Q9550

We’ve said for months that Core 2 was a lame duck and it gets even lamer with the introduction of the LGA1156 parts. The Core 2 can’t compete with its Nehalem brethren on any front. Even the lowly – and cheaper Core i5-750 – gives the Core 2 Quad Q9550 such a bad beat down, that AMD would likely feel bad for it. And as we said, even the Phenom II X4 965 BE mostly smokes the Core 2 Quad. Sure, the Core 2 Quad surprisingly outperforms the Phenom II X4 965 BE in a couple of places, but it’s still the loser. We could have reached for the highest bin Core 2 Quad, the 3GHz Q9650, but at $319 on the street and list, it makes no sense.

Hell, even at the street price of $220, it’s hard to justify the Core 2 Quad Q9550 over a new Core i5 rig. Even AMD’s Phenom II has a better roadmap as the company will support it through at least through 2010 with newer, faster CPUs. The same can’t be said of Intel which is unlikely to introduce faster Core 2 parts. There is still a place for Core 2 though: If your box is rolling a dual-core and the LGA775 board supports 45nm quads, it would be dumb not to get one more upgrade out of it. Outside of that, it’s clear Core 2’s glory days are long, long gone.

Intel 2.66GHz Core i5-750

Of the troika of new chips, Intel’s Core i5-750 is likely to be the big hit. With wholesale prices of $200, the average joe or jane can now build a Nehalem-based system for almost as much as a Core 2 or Phenom II box. The key price differentiator is the motherboards. Board’s we’ve seen will likely top out about $225 but many will dip into the $150 range. That makes it significantly different from when Intel’s Core i7-920 launched. Sure it was only $280 (and still a year later) but the boards for the processor all pushed the $300 mark. Even today, it’s hard to get an LGA1366 board for $250.

In performance, the lower clocked Core i5 schools Core 2 Quad Q9550 as well as the Phenom II X4 BE. The Phenom II X4’s massive clock difference actually gets close to the Core i5 in a few benchmarks and surpasses it in two but the cheap Lynnfield is clearly the winner. Against the Core i7, as expected, the Core i5 loses but for many, giving up 10 percent to 25 percent in performance is worth it to save, oh, $300 over the price of the top-end Core i7-870. The performance delta between the i5 and i7 is the greatest in multi-threaded apps. In gaming and apps not optimized for quad cores, they’re a lot closer.

In the final analysis, the Core i5 falls right where Intel’s bewildering branding scheme puts it: faster than Core 2 and Phenom II but slower than i7 and really damned cheap to boot too. Buy this chip is the vast majority of your applications are not optimized for quad-core or more. If, however, your work counts on an increased amount of cores, go the extra mileage for a Core i7 with Hyper-Threading.

Intel 2.8GHz Core i7-860

If you use the Goldilocks formula, this new quad-core, Hyper-Threaded CPU is “just right.” It’s just the right blend of performance to price. At $266 for a 2.83GHz, its most natural enemy is the Core i7-920 in LGA1366. The 920’s primary advantage is increased bandwidth thanks to its tri-channel DDR3 support. But since the vast majority of apps really don’t need that much bandwidth, the 860 is generally faster by five percent thanks to its 5 percent clock advantage. That’s pretty much what we saw in our tests too – about five percent difference in the vast majority of the tests in favor of the 2.8GHz 860. And of course, the 860 also gets a boost from its increased Turbo Boost clocks. A stock 920 will only Turbo as high as 2.93GHz under some loads. The 860 will max out at 3.46GHz under some loads. Why is the Turbo higher on the 860? The 920 was a first generation Nehalem processor and Intel played it very conservatively with the Turbo. With the 860, the company feels more comfortable pushing it higher.

So what should you buy? Both procs are priced the same at $266, but each has its strengths. The 920 has true dual x16 PCI-E 2.0 support thanks to the X58 chipset and discrete PCI-E as well as tri-channel DDR3. Frankly, we think both of those features are pretty minor reasons to choose 920. The main reason to pick a 920 is the upgrade path. In a few months, Intel will introduce a 32nm-based six-coire processor codenamed Gulftown. With Hyper-Threading, you get 12 threads as well as any goodness Intel can squeeze out of the new 32nm process (both the 860 and 920 are 45nm parts). A rig built on 860 will also not see Gulftown and probably will not see a six-core part for quite some time if ever. Again, Intel intends for LGA1156 to be for the mainstream and the 99 percent of mainstream users don’t need 12-threads for their apps nor will they pay out $1,000 for a Gulftown CPU.

That takes us back to the Core i7-860 part. If you save possibly $150 to $200 by building an 860 rig, that can be put toward a faster GPU, a bigger hard driver or a larger monitor. In this economic climate, that’s pretty appealing and really the pick of the litter if you ask us.

Intel 2.93GHz Core i7-870

In this comparison of five CPUs, the 870 was head and shoulders the fastest CPU. Generally, we saw a 5 percent advantage in the 870 – just what you would expect from a 5 percent clock advantage. We did see some interesting results though. The multi-threaded Cinebench 10 actually saw the 870 with a 13 percent higher score. World in Conflict also saw the 870 run away with a 24 percent higher frame rate than the 860. So here’s the real question: Would you pay nearly 100 percent more for that? For most folks, probably not. At $555 for the Core i7-870 and $266 for the Core i7-860, it just makes a hell of a lot sense to give up some of that performance to put towards something else in your machine. And if you really are into being on the ragged edge of performance, it makes a lot more sense to jump in with both feet for an LGA1366 platform and the promise of a six-core upgrade in a few month. So we’re a little ambivalent about the Core i7-870. But make no bones about it, the Core i7-870 is the king of the CPUs that battled today, it’s just not really a “budget” CPU is it?

Intel 2.66GHz Core i7-920

If you haven’t figured it out by now, Intel’s 920 is the chip all three LGA1156 Lynnfield’s were gunning for. A huge hit among the performance budget shoppers, it overclocked well and got you into the Core i7 club without paying through the nose.

With its three new siblings here, the 920 is mostly pushed to the back in performance and pricing competitiveness. The other chips run at higher clocks, automatically Turbo Boost to higher levels and can save you from $150 to $200 in associated building costs. So why bother with the 920? There is still some magic to this old favorite. In this roundup, the 920 was generally in third or second place in the vast majority of the tests. And in some benches, the 920 interestingly bubbled up to the top. Even though its superior tri-channel doesn’t always pay off, it apparently does in Photoshop CS3 where the 920 took top honors and we suspect its superior bandwidth helps keep it in hunt elsewhere too. We had heard early rumors that Intel would axe the 920 once the LGA1156 procs were out but the company has since told us it had no plans to discontinue 920 right now. That’s a good decision too. While LGA1136 is definitely an exclusive club, there’s no reason for Intel to make it even harder to get in. If your only choice to getting an LGA1366 platform was to shell out $500 for a Core i7-950 or $1,000 for a Core i7-975 Extreme Edition, very few people would choose that route.

Of course, the biggest reason to buy 920 is for the path to the Core i9 “Gulftown” CPU that Intel will introduce in a few months. Ideally, you could build a 920 box and use it for a year or so until the Gulftown derivatives drop down enough to become “budget” chips themselves.

The Benchmarks

To run our test, we tried to balance each particular CPU’s platform as closely as possible but given the differences in each chip’s requirements that’s almost impossible. We used the same make and model Western Digital Raptor 150GB drive for all four builds, the same 64-bit Windows Vista Home Premium SP2 build, the same make/model/clocked EVGA GeForce GTX280 and the same drivers. Windows 7 was available at the time of our reviews but we felt it would be best to use Vista as drivers were readily available for it.

The only real differences, of course, were the motherboards and RAM configurations. Although the Core i7-920 does unofficially support higher clocked RAM, we stuck with the official DDR3/1066, but 6GB of it. For the dual-channel Phenom II X4 and Core i5/i7, we used 4GB of DDR3/1333. For the dual-channel Core 2 Quad, we used the far more common DDR2/1066 as the vast majority of Core 2 Quad folks are running DDR2. For motherboards, we used three Gigabyte boards: the new GA-P55-UD6 for the LGA1156 procs, a GA-EX58-UDR3 for the LGA1136, and a GA-MA790GPT-UD3H for the Phenom II. For the Core 2, we relied on an Asus Maximus II Formula.

Benchmark Test Setup

CPU	Core i7-920	Core i7-870	Core i7-860	Core i5-750	Core 2 Quad Q9550/Q9550s	Phenom II X4 965 BE
RAM Mode	Triple	Dual	Dual	Dual	Dual	Dual Unganged
RAM Size	6GB	4GB	4GB	4GB	4GB	4GB
RAM Speed	DDR3/1066	DDR3/1333	DDR3/1333	DDR3/1333	DDR2/1066	DDR3/1333
RAM Latency	7-7-7-20-1T	9-9-9-24-1T	9-9-9-24-1T	9-9-9-24-1T	5-5-5-18-2T	9-9-9-25-1T
Hard Drive	WD 150 Raptor	WD 150 Raptor	WD 150 Raptor	WD 150 Raptor	WD 150 Raptor	WD 150 Raptor
OS	Vista HP 64-bit SP2	Vista HP 64-bit SP2	Vista HP 64-bit SP2	Vista HP 64-bit SP2	Vista HP 64-bit SP2	Vista HP 64-bit SP2
GPU	Geforce GTX 295	Geforce GTX 295	Geforce GTX 295	Geforce GTX 295	Geforce GTX 295	Geforce GTX 295
GPU Clocks 3D	670/1458 /1215	670/1458 /1215	670/1458 /1215	670/1458 /1215	670/1458 /1215	670/1458 /1215
Motherboard	Gigabyte GA-MA790 FXT-UD5P	Gigabyte GA-P55 -UD3	Gigabyte GA-P55 -UD3	Gigabyte GA-P55 -UD3	Asus Maximum II Formula	Gigabyte GA-EX58 -UD3R

BENCHMARKS

CPU	Core i7-920	Core i7-870	Core i7-860	Core i5-750	Core 2 Quad Q9550/Q9550s	Phenom II X4 965 BE
MainConcept Reference 1.0	1235	1115	1170	1337	1644	1388
MainConcept Refernece 1.0 Pro AVC	696	635	664	769	986	840
Premiere Pro CS3	671	610	630	620	741	733
Cinebench 10 64-bit	16140	18275	16085	14442	12280	14083
Handbrake iPod Classic	994	939	993	1198	1366	1220
PCMark Vantage 64-bit Overall	6929	7536	7299	7208	6241	6824
POV Ray 3.7	3470	3888	3702	2773	2669	3045
Photoshop CS3	116	119	126	128	132	123
ProShow Producer	636	640	617	700	862	911
Everest 5.0 RAM Read (MB/s)	14387	12997	13641	12867	7807	8154
Everest 5.0 RAM Write (MB/s)	11639	10811	10992	9881	7085	6794
Everest 5.0 RAM Copy (MB/s)	15790	15414	15393	14684	7455	10246
Everest 5.0 RAM Latency (ns)	61	53	52	31	64	54.3
Sandra RAM Bandwidth (GB/s)	22.4	16.7	17.2	16.8	7.2	12.7
Fritz Chess Benchmark	21.4	23.94	22.38	17.38	16.97	17.04
3DMark Vantage Overall	15008	15002	14985	14947	14681	14544
3DMark Vantage GPU	12306	12231	12247	12249	12013	11978
3DMark Vantage CPU	44002	46815	45525	44066	40644	40679
Valve Particle Test	143	159	151	124	99	95
Valve Map Compilation	146	128	133	121	129	125
Crysis	146	156	150	147	119	104
Resident Evil 5 Fixed DX9	114	115	118	109	85	89
Resident Evil 5 Fixed DX10	117.7	106.5	119.7	109.4	83.8	89.2
Resident Evil 5 Variable DX9	145.8	155.9	147.2	155.4	133.7	140.3
Resident Evil 5 Variable DX10	150.4	155	157.8	160	133.9	140.2
World in Conflict	221	282	227	266	159	160
WinRar 3.20	581	588	594	706	868	805

Best scores in bold.

Next, our detailed benchmark analysis.

Benchmark Analysis

Sisoft Sandra RAM

(higher is better)

Sisoftware’s Sandra RAM benchmark lets the Core i7-920 strut its tri-channel DDR3 stuff. Keep in mind, we tested the Core i7-920 at its officially rated speed of DDR3/1066 vs. DDR3/1333 for the five other parts here. That extra bandwidth of the 920, in fact, may help it keep pace with the newer processors despite a general clock deficit. Still, as we’ve said, the vast majority of folks just don’t need that bandwidth.

WinRar 3.20

(lower is better)

Now that we’ve said don’t pay attention to bandwidth, here’s a test that does actually show how much it can help. We used WinRar 3.20 to compress a folder of RAW files shot with a Canon EOS 5D. The files, pretty much, are near uncompressible so we thought it would be a good way to stretch the processors. The Core i7-920 just edges out the Core i7-870 despite a 266MHz gap between the two chips.

World in Conflict

(higher is better)

We ran World In Conflict at low resolutions to take the GPU out of the equation for performance. Some have questioned this but it tells you what you would get if you had the fastest GPU in ze vorld. The results were a bit odd but we saws the Core i7-870 up front and then the budget Core i5-750 next in line. Huh? Why would the results be so whacky? We have two theories: Turbo Boost can play havoc with unpredictability. Afterall, it automatically overclocks individual cores for a given thermal load and power consumption load so it’s possible the Turbo Boost for the 750 just kicked up into high gear in the most single-threaded game. Our other theory is that the oddly super low latency that the 750 exhibitied may have contributed to its scores. The other conclusion you can draw from this is that the Phenom II and Core 2 Quad should have just stayed off the field.

Crysis

(higher is better)

We used the same low quality and low resolution mode for our Crysis runs to show you just what kind of frame rates you’ll get in four years when the newest GPUs finally make Crysis its bitch. If you happened to still be running the same rig in 2013, here’s how your CPU would shape up. Unlike some of the multi-threaded benchmarks, we actually saw very predictable results with the most expensive 870 on top, the budget builder’s favorite, the 860, in second, and the 750 and 920 pretty even.

Valve Particle Test

(higher is better)

Valve’s Particle Test is a quad-core optimized test that measures CPU performance. Here, our results were predictable with the Lynnfields seemingly scaling with clock speeds. The surprise again is how well the 920 does which can only be attributed to its third-channel of DDR3.

Proshow Producer

(lower is better)

Our Proshow Producer benchmark hammers all cores available and again, there’s that pesky Turbo Boost messing with us. The Core i7-870 should have been the fastest, but the Core i7-860 just edges past all contenders. If you’ve wondered if Hyper-Threading pays off? Yes. Notice the roughly 10 percent gain the 2.66GHz 920 has against the 2.66GHz 750. That’s generally what we’ve seen from Hyper-Threading in other benchmarks. One other conclusion: Core 2 Quad and Phenom II X4, why are you here again?

Intel Lists New Low-Cost Processors

Paul Lilly — Tue, 02 Sep 2008 10:10:29 -0500

With ~~Nehalem~~ Core i7 nearing release, that means you can expect to find good deals on what's soon to be last generation hardware. But if you're looking to jump onto the Core 2 bandwagon on the cheap, you needn't wait for Core i7. Intel has updated its processor pricing list and added a new Celeron D model.

Taking its place as the second least expensive quad-core processor in Intel's lineup is the 45nm Q8200 priced at $224 (only the Q6600 costs less). Two and a quarter C-notes buys you four 2.33GHz cores running on a 1333MHz frontside-bus, but only 4MB of L2 cache.

For those content with two cores, the 45nm E5200 priced at $84 is now the least expensive Core 2 Duo processor in Intel's lineup. The E5200 comes clocked at 2.5GHz on an 800MHz frontside-bus with 2MB L2 cache.

And finally, making its debut is the Celeron D 450. Priced at a low $53, the 65nm 450 runs at 2.2GHz on an 800MHz frontside-bus with 512K of L2 cache.

Image Credit: Intel

Intel Introduces a Few New CPUs, Cuts Prices on Others

Chris Moody — Thu, 21 Aug 2008 12:31:08 -0500

Intel adds a few processors and drops a few prices this month in it’s CPU line up. There doesn’t appear to be any shakeups from Intel’s expected plans.

Intel's Core 2 Extreme Quad Core line remains unchanged, but in the standard line, the Q9650 joins the line up at the top, while the Q9550 drops 40% from $530 to the Q9450 previous level of $316. The Q9400 is also new, and enters at the same price as the Q9300 and Q6700 (a 65nm process CPU) at $266.

The only other prices changes were in the Xeon line, with the new X3370 coming out and the X3360 dropping 40% to $316.

All prices are in 1000 tray units.

We will certainly see more changes when Intel ships Bloomfield sometime in Q4.

AMD’s new X4 quad core

Gordon Mah Ung — Wed, 26 Mar 2008 23:02:28 -0500

It’s been a pretty lousy six months in AMD fan boy land. Even the most vocal of fan boys have found themselves ducking for cover as AMD was hit with delays, production problems and errata bugs.

With the release of the Phenom X4 9850, AMD hopes to finally put some of those problems behind it. Dubbed “B3” chips, the new CPU no longer features the infamous TLB bug that hobbled performance of the original Phenom chips. If it isn’t already obvious to you, AMD is also reverting back to the X4 nomenclature that it had toyed around with before Phenom launched. New Phenom’s will now carry the X4 designation for quad cores and X3 for tri-cores (of if you want to be snide, quad cores with one bad core).

We tested the new 2.5GHz Phenom X4 9850 using an Asus M3A32-MVP board with 2GB of Corsair Dominator RAM at 1066MHz data rates, a 150GB WD Raptor, Windows XP SP2, and the new Nvidia GeForce 9800GX2 GPU. For a comparison, we populated a new 45nm Penryn Core 2 Quad Q9300 and a Core 2 Quad Q6600 in an EVGA nForce 790i Ultra motherboard. For those who don’t know, the Q9300 doesn’t have the same relationship that the Q6600 has to the top-end procs. Intel has cut the cache in half from the 12MB in a Core 2 Extreme QX9650 to 6MB in the Q9300. This makes it sort of a “Penryn-lite” chip. Why? There’s probably two explanations: Intel has either decided to do something with its Penryn CPUs with some bad cache (ala AMD) or it has decided to hobble the chip intentionally to give its top quad cores more of an edge. We set the pair of 1GB Crucial DDR3 modules at 1333 and installed Windows XP SP2 on a WD Raptor 150GB drive. Finally, the same reference GeForce 9800GX2 card was used in the same machine and the same public drivers were used for both tests. Is it fair to pit a DDR3 platform against a DDR2 platform? We think so. Afterall, we would have tested Phenom on DDR3 but there’s just no support for it yet. Even though the DIMMs in our Core 2 platform cost more than the CPUs, we think it would be unfair to not put the Core 2’s best foot forward. If anything, Intel is probably grumbling that the test should be conducted on X48 which, we’re sure, the company thinks is a better and faster platform.

For the most part, the Q9300 owned the night. It generally outpaced the old Q6600 quad core and the new Phenom quad core in most of our tests. In a clock-for-clock test between the Phenom and Penryn, we would declare the Q9300.

AMD fans shouldn’t hand their heads in shame though. The Phenom numbers weren’t so far behind as to be dead in the water. Certainly, it doesn’t go quite head-to-head with the Penryn-lite but it is still fastest AM2 chip available today for most apps. That makes it a pretty good upgrade for those with AM2 boards that are capable of supporting Phenom. If you had to start afresh though and you weren’t opposed to Intel for anything other than religious reasons, Core 2 is still the champ in the cheap quad core category.

AMD’s big problem is expectations. When Phenom first launched late last year, the company had built up wild expectations that Phenom would level the playing field with Intel’s quad core. In truth, all it could do was barely compete with Intel’s lowest of the low quad cores – a chip that was a year old. AMD still has no response to Intel’s 3GHz Core 2 Quad QX9650 and the soon to be released 3.2 Core 2 Quad QX9770 chip and it doesn’t look like the company will for several more months. Meanwhile, Intel is already demonstrating fully functional Nehalem CPUs and says they’re on track to launch later this year.

Benchmarks
	Core 2 Q9300	Core 2 Q6600	Phenom X4 9850
Clock	2.5GHz	2.4GHz	2.5GHz
L2	6MB	8MB	2MB
L3	N/A	N/A	2MB
Rated TDP	95 watts	95 watts	125 watts
FSB	1,333	1,066	N/A
Bulk Pricing	$266	$266	$235

3DMark06 Overall	13,990	13,565	13,594
3DMark06 CPU	4,103	3,856	3,807
Valve Particle Test	87	76	72
Valve Map Build (min:sec)	2:42	2:48	2:57
CineBench R10	9,622	8,763	8,065
ScienceMark 2.0	1,572.21	1,393.57	1,606.46

PCMark05 Overall	9,080	8,948	8,547
PCMark05 CPU	8,064	7,746	7,396
PCMark05 RAM	6,012	5,434	4,761
PCMark05 GPU	15,643	15,806	13,972
PCMark05 HDD	7,633	7,647	7,766

Adobe Premiere Pro CS3 (min:sec)	15:45	17:33	18:01
Photoshop CS3 (min:sec)	2:20	2:28	3:35

FEAR (fps)	199	209	200
Quake 4 (fps)	169.7	166.6	165.7
Unreal Tournament 3 (fps)	119	105	87
AutoGK DIVX 6.8 (min/sec)	9:36	10:07	11:44
AutoGK XVID	11:08	11:38	12:42

It's A War Of Words Between Intel and AMD

Gordon Mah Ung — Wed, 19 Sep 2007 16:33:19 -0500

Intel's Pat Gelsinger with the company's upcoming dual Core 2 Extreme "Skull Trail" system.

The CPUs aren’t even out yet but the war of words has already begun between Intel and AMD over whose CPU is bigger, err, better. While Intel showed off early samples of its next-gen Nehalem CPU at its annual developer conference this week, AMD was sniping at the company from the sidelines.

The big splash came from Intel which said that it would have its 45nm-based Penryn family of CPUs out this fall, the company also demonstrated its, next next-generation chip, code-named Nehalem, up and running a year before it's even scheduled to be launched.

Nehalem will be scalable from 1 to 8 cores and will bring back a form of Hyper-Threading, but the big change from previous Intel processors will be the inclusion of an integrated memory controller. The three-channel DDR3 controller will work along with a new chip-to-chip communication protocol dubbed QuickPath. Previous generations of quad-core chips, including the new Penryn processor, continue to use a shared front-side bus toplogy. For one core to communicate with another core on another die, it must use the much slower front side bus. Intel has addressed by increasing the bus to 1,600MHz for its server and workstation Penryns, but Nehalem will have a much faster connection between CPUs at it's beck and call.

That got AMD into the spin game by whispering to anyone that would listen, that they were right!

“What's amazing is that many of (Intel’s) ‘groundbreaking, innovative new technologies’ are close facsimiles of technologies AMD pioneered, is already shipping, and in some cases, has been shipping for years,” AMD crowed to the media. “For example, products that are more than a year away, like Nehalem (compare to native Quad-Core AMD Opteron), and QuickPath (compare to AMD Direct Connect Architecture and HyperTransport) are simply Intel's admission that AMD was right all along about an integrated memory controller being the key to a superior processor architecture.”

AMD first moved the memory controller from the chipset directly into the CPU die with the Opteron and Athlon 64. The chips also utilized a chip-to-chip communication protocol based on HyperTransport instead of a front-side bus topology. CPUs weren’t the only thing AMD pooh-poohed. The company also said Intel’s graphics project, code-named Larrabee (which Intel publicly confirmed this week) sounded pretty damned similar to its own Fusion project that will combine ATI graphics with AMD CPUs.

Obviously sensitive to AMD’s jibes, Intel exec Pat Gelsinger, acknowledged the similarities.

“When you look at that simple system architecture they look pretty similar,” Gelsinger responded rather testily to a media question. “On a gross level they’re very similar.”

But Intel's Gelsinger shot back, if Opteron’s on-die memory controller and chip-to-chip is so bad-ass, why does a Harptertown (Intel’s Penryn-based Xeon) using the outdated front-side bus already outperform Barcelona? Gelsinger attributed it to Intel’s superior cache design and the bus. “When we bring those big caches over to Nehalem, that superior cache architecture, wow.”

"(With Nehalem) we’re going to have a superior systems architecture, dramatically superior system bandwidth with superior caching technology,” Gelsinger booasted. “The instruction per clock of Merom is still ahead (of Opteron), and Nehalm jumps those even further. “So (Nehalem has) a superior microarchitecture, cache technology, and memory bandwidth -- This is a phenomenal platform.”

Translation: We’re going to kick AMD's ass. Gelsinger’s and AMD’s posturing aside, all bets are on Intel for this season. Penryn will make its debut on the desktop and server within the next two months while AMD’s first Barcelona chips were shipped at a disappointing 2GHz. Meanwhile, AMD’s great green hope, Phenom FX, may not debut until December. Too late to make the crucial Christmas shopping season. There’s been speculation that AMD’s problem lies in fabrication problems--with low yields being a common explanation. Matters weren’t helped when the company announced that it would release a triple-core core Phenom in 2008. While some say AMD’s tri-core plans could work because it would fill a price and demand gap between dual-core and quad-core machines, it left an opening for Intel to pee on AMD’s parade.

Is AMD doing this to fill a non-existent gap, or is AMD having problems with its native-die approach that it’s trying to cover up? AMD has made much hay of its “true quad core approach,” which puts all four execution cores on a single contiguous piece of silicon. Intel, meanwhile, still builds its quad-cores by joining two dual-core dies inside the CPU package. While AMD’s architecture is more elegant, it also can lead to lower yields as each execution core on the die must be functioning for it to be made into a CPU. Intel’s yield could be drastically better since the company only has to two find matching dual cores to make a quad core. In other words, AMD must make its quad-leaf clovers by finding perfectly formed ones while Intel can construct them using two, two-leaf clovers. Intel’s pragmatic, but architecturally inferior approach could prove to be right in the end. Intel officials say yield issues aside, it's already selling $200 quad-cores, so why would anyone want a tri-core?

Enthusiasts confidence in AMD also took a hit when rumors surfaced on the Inquirer that AMD’s much touted FASN8 dual-processor uber-enthusiast platform may have fallen under executioners axe. The company rep couldn’t confirm or deny the rumor when Maximum PC asked.

Even if FASN8 is dead, Intel says it plans to roll forward with its V8 platform. Like AMD’s FASN8, V8 is a dual-processor board offering up to 8 execution cores in a single PC. But unlike its workstation brethren, V8 would offer up to four physical and electrical x16 slots and be based on an Intel platform dubbed, Skull Trail.