Intel's Nehalem benchmarked: 2.66GHz faster than 3.2GHz Penryn

Intel's got some Nehalem chips floating around Computex, and the crew from Anandtech managed to snag a couple and run some benchmarks. As you'd expect, Intel's latest and greatest put up pretty solid numbers, performing most operations faster than a 3.2GHz Penryn chip -- while only clocked at 2.66GHz. When the Penryn chip was clocked at 2.66GHz, Nehalem ran 20 to 44 percent faster, mostly due to insanely fast memory access speeds and Hyper Threading. Sadly, due to a wonky test mobo, Anandtech wasn't able to test performance when paired with a GPU, and there were some memory latency issues -- which hopefully means Nehalem will be even faster when it's out for real. Your move, AMD.

[Via AeroXperience]

[MeatPop @ Jun 5th 2008 9:59PM]

Damn, Nehalem is going to be brain-meltingly fast!

[Wesley Burchnall @ Jun 5th 2008 10:43PM]

Is this Nehalem shown in this benchmark picture above a dual core? or quad core? or...single?

[Flashpoint @ Jun 5th 2008 10:49PM]

It will be brain meltingly fast right up until the Crysis Sequel demands a 2.6 Ghz Nehalem as a minimum requirement with a video card that has 2 GB of Ram.

[Mam00th @ Jun 5th 2008 11:02PM]

@Wesley

It's probably a quadcore because the first nehalem chips to hit the market will be quadcores.

[Mam00th @ Jun 5th 2008 11:08PM]

After reading from AnandTech, I can confirm it's a quadcore.

[Abuzar @ Jun 6th 2008 12:06AM]

It's a quad core with core executing two threads so it kinda functions lika a octa core.

[loosely_coupled @ Jun 6th 2008 4:41AM]

This Engadget post makes little mention of a BIG warning by Anandtech saying that the early prototype motherboard and chipset may be impeding performance, especially in memory-intensive tasks.

Intel says Nehalem will have 10-25% higher single thread performance, and 30%-100% higher multi-thread performance than Penryn.

[Thi mam(kris120890) @ Jun 6th 2008 9:53AM]

There is no crysis sequel. So probably "does it play crysis 2 "

[BYOPC @ Jun 5th 2008 10:00PM]

Well as I always say "Knee - all - of - em"

[ajumps @ Jun 5th 2008 11:18PM]

What are you, twelve?

[wickedpheonix @ Jun 5th 2008 11:55PM]

No he's 8 - he's got an octacore :P

Sorry I'll stop cracking bad jokes now...

[Anthony @ Jun 5th 2008 10:01PM]

it would be nice to see AMD come out of nowhere and kick intel in the nuts.

very satisfying indeed. =]

[bc @ Jun 5th 2008 10:09PM]

Lol! You made me laugh. I agree.

[Ryan Trevisol @ Jun 5th 2008 10:28PM]

LOL! I'd love to see that too, but I just don't think it's going to happen.

[Zach @ Jun 5th 2008 10:29PM]

that it would..but considering their lineup is to consist of broken quad core parts I MEAN TRI-CORE! ;)

i say this coming from a turion mobile 2.0ghz..
sadly..AMD has fallen behind much in the same way intel did way back.

IMO, the only way for AMD to come back would be to completely start from scratch (i.e. C2D) and create something completely new, as their current architecture is clearly failing..

[t3_slider @ Jun 5th 2008 10:34PM]

Although I agree that AMD is behind, I actually like their triple core idea. Instead of the consumer having to pay both for the actual quad-core chip and the chips defective in one core, they're only paying for their one quad core processor (and the supreme defects, of course). It also allows people to buy triple core processors for nice and cheap. Good for the consumer in my mind. That's not going to do much for the gamer, or for AMD as a company, but that's one move I actually like.

[PCIV @ Jun 5th 2008 10:47PM]

If only they can pull a C2D. Only problem, they don't have resources. Intel's C2D came from Israel, but AMD DOESN'T have a team in Israel or so...

[ben @ Jun 5th 2008 11:40PM]

I have an AMD in my current computer that's 3.5 years old, and I'll have an Intel in my next. AMD/ATI turned me off when I bought one of their branded TV Tuners and had to deal with their craptastic drivers, of which the install of a new version would cause unrecoverable hangs that I needed to force reboot to escape. It's not even really a matter of price anymore, since you can get an extremely good wolfdale dual-core for about $200.

[kal326 @ Jun 6th 2008 9:34AM]

I expect that to happen about the same time Ed McMann rolls up to my front door with one of those giant checks from Publishers Clearing House. I wouldn't hold my breath....

[Chuckles McGee @ Jun 6th 2008 8:13PM]

"Your move AMD"

Yeah, it's BEEN AMD's turn to move ever since the first Core 2s came out. Still waiting...

[tbone @ Jun 5th 2008 10:13PM]

it begins...

[digitallysick @ Jun 5th 2008 10:13PM]

DO WANT!

[Dan @ Jun 5th 2008 10:23PM]

Bahahaha....
Good luck AMD indeed.

[Mehdi Cheddadi @ Jun 5th 2008 10:26PM]

AMD doesn't care

[torqueo @ Jun 5th 2008 11:00PM]

Yeah, if they cared they'd compete in some way.

[Jon Doe. @ Jun 5th 2008 11:45PM]

Yah same as Palm not caring about the Pocket PC and you can see where that lead them. AMD is doomed if they truly don't care.

[Mark @ Jun 9th 2008 2:34AM]

Engadget, why do you quote your "read" source as an obscure blog, when the actual source is Anandtech (which you have mentioned above). Why not link directly to the source article?

[Miguel @ Jun 5th 2008 10:52PM]

The "[via X]" tag is NOT the read link. It's Engadget's way of crediting their source. Nilay originally found this information on AeroXperience, including the citation of Anandtech, so he credits AeroXperience with his information. The read link will put you through to Anandtech. It's like reading in the newspaper about court proceedings, wherein a spokesperson says "the judge's decision says [blah blah blah]". Technically, the source IS the spokesperson, but they got their information from the judge's decision. So you credit the spokesperson.

Didn't you learn about primary and secondary sources in high school composition?

[Doc @ Jun 6th 2008 3:17AM]

This isn't a school essay/paper. This is the web. The web was designed so you could go from one point to another *without* having to troll through a whole load of crap to get there. That's why we have hyperlinks.

Would be nice in future if we could get the direct source please :)

Thanks

[KarlW @ Jun 6th 2008 7:17AM]

I wrote the AXP article, and I sent the link. They got the information via the AXP post, which is why is says "via AeroXPerience".

Pretty logical, IMO.

[Alex @ Jun 5th 2008 10:40PM]

AMD is screwed.

[Jon Doe. @ Jun 5th 2008 11:46PM]

AMD has been screwed for years. the difference is that there is now double...check that quad penetration.

[MeatPop @ Jun 5th 2008 11:53PM]

Soon to be an unheard of octo-penetration, ouch!

[Minilap @ Jun 6th 2008 12:00AM]

And is that a good thing?

[MeatPop @ Jun 6th 2008 12:05AM]

Put it this way, would you want eight sources penetrating various orifices in your body?

[Jherez @ Jun 6th 2008 12:39AM]

If AMD is screwed then we're screwed...think about it.

[Alex @ Jun 6th 2008 6:25AM]

Very good point, but look at the other competition around here...like that chip that plays Crysis...I think Intel still has some good competition.

The cost of making chips are going up...and Intel has the "junk" to take care of AMD.

[chris fredette @ Jun 6th 2008 12:11PM]

AMD is only screwed when Intel lowers their prices. History has shown Intel doesn't do this.
Before AMD was the faster chips they were the cheapest and a much better bang-for-the-buck. Intel is only faster in the real high end chips and they are expensive. There are a few exceptions and price overlaps but it is generally true. Like the E8400 is fast but it's $220 right now. If you want to build a $400 box that is too expensive. AMD has fast 2 cores for under $100; not as fast as the E8400 but fast enough for most and over $100 cheaper.
In the quad game they are down right now because the Q6600 is pretty fast and about on par with the Phenoms for price. In volume I suspect they are much cheaper though because the lower end quad machines from HP and gateway are cheaper that the Intel machines(at retail).

[DarkLight @ Jun 5th 2008 10:43PM]

Dude..... !!

(Nothing else to say, I really, really wish I had one of these)

(But still I'm a bit curious as to what will AMD come with)

[TGGGD86 @ Jun 5th 2008 11:41PM]

Well Niley Patel, if you write for Engadget you would be smart enough to know that Nehalem has double the interface speed as Penryn, has 8MB shared L3 Cache (Which Penryn does not have).

You should know this is a COMPLETELY new architecture that has tons of new technologies that were not even public at the release of Penryn. Just like your 3.0 GHz P4 is slower than a 1.83 GHz Core 2 Duo.

It's ALL ABOUT THE ARCHITECTURE!!!!!

Wish Engadget had smarter people than you writing there articles!

[Abuzar @ Jun 6th 2008 12:12AM]

Yeah half the time engadget doesn't know what it's talking about when it comes to PC hardware. Although to be fair hyper threading and the integrated memory controller was really a big reason for the performance increase.

[maveric101 @ Jun 6th 2008 12:54AM]

uh, i think he knew and was just confirming... give him some credit... he does write for a gadget blog

[Zak @ Jun 6th 2008 11:56AM]

TGGGD86 - You should be careful about calling people stupid when you don't seem to know the difference between "there" and "their".

[Flashpoint @ Jun 5th 2008 10:50PM]

analyists estimate over 50 million Netbooks will be sold by 2011.

AMD should take the time to make a powerful, energy efficent chip to push into that market. OTHERWISE THEY ARE F*CKED.

[Rubicon @ Jun 5th 2008 11:12PM]

AMD should spend its time making its multicore cpus more stable

[Shola @ Jun 5th 2008 10:54PM]

This is why i'm not even flinching about Penryn/Montevina, I was even gonna stay put at Napa but circumstances forced me to an SR machine.

Its the same ol story with all recent processors/platforms: Powerful processors choked by FSBs

Nehalem is where it's at and i cant wait

[Minilap @ Jun 5th 2008 10:57PM]

If it wasn't for competition with AMD, intel would have never bothered to bring the technology so much forward, and by now they were selling us 90nm Pentium 5 or some crap like that. But anyways Intel 32nm FTW.

[Mark @ Jun 6th 2008 12:22PM]

Actually, it is not true that we would be still running 90nm pentiums now if not for AMD. Some people seem not to realize that if AMD was not around, Intel would still have competition - from itself. In short, since processors are not consumable items (largely), then consumers have the option of not buying a new processor and staying with their current processor. This means that Intel can only generate sales by coming out with something that is better than what consumers currently have. Given that an already purchased processor costs $0, the new processors need to be substantially better to compete against a competitor that is free. This, much more than the competition from AMD, is what has driven the increase in processor speed over time.

[Paganmoon @ Jun 9th 2008 7:52AM]

Well, tbh, Intel doesn't have a competition atm with their penryn out. But are still going to push their Nehalem platform, by your thinking they wouldn't need to...

[James Cameron @ Jun 5th 2008 11:13PM]

I'm glad I didn't wasted money on the Penryn.