http://anandtech.com/cpuchipsets/int...oc.aspx?i=3326
Enjoy
Printable View
:worship::worship::worship::worship::worship:
If those numbers hold up and are indicative of the final silicon, which seems plausible, AMD is cooked. They will have no where left to hide, unless they plan on selling their top end CPU's for $50.
Maybe I'm over reacting.
I'm amazed at how Intel is able to improve upon Conroe so fast. Conroe is already a high IPC beast. To go above that by leaps and bounds is simply amazing. I'm speechless.
too bad they didnt test any single threaded bench
Why would they do that? That would be like testing a high performance V8 with 7 spark plugs removed.Quote:
Originally Posted by GoThr3k
yerp. thx for article.
the performance is good, now we just need some overclocking results!
AMD is looking more and more cooked everyday. If intel positions their current releases as mid to low end and nehalem as the top end, AMD will become Cyrix(uncompetitive chips at all ends). And with AMD current financial situation, that could mean death for both AMD and ATI. With Intel's direct attack on NV works and if larabee's a crazy monster, it will have basically a monopoly on everything. I am probaly overreacting(especially the Larabee thing), but core 2 decimated AMD into a shell of its formerself. Another attack like that AMD will never survive.
Hail to the new monopoly of INTEL:(
Anyways why is Larabee supposed to be such a monster. On XS its hyped to be the second coming of christ. I haven't heard anything about it for performance.
Yeeeez, oh god! Ill take two to go please!
OMG it's incredible look at raw power 3d rendering from 25-40% WHOA
I hear a giant sucking sound, the sound of AMD's hopes of survival going down the drain.
I for one don't welcome my new Intel overlords.
Smoked and baked! I hope the power consumption can be tweaked, even tho its still alot better performance/watt.
Intel's new toy is even faster then i thought.
The ATI branch will get sold off and keep truckin just fine.
2 Things. This is not final silicon or platform, so in theory performance should improve. Secondly, why is there an entire tube of TIM on the CPU. :shrug:
All tests run with high latency and singlechannel? :xQuote:
We had access to a 2.66GHz Nehalem for the longest time, unfortunately the motherboard it was paired with had some serious issues with memory performance. Not only was there no difference between single and triple channel memory configurations, memory latency was high. We know this was a board specific issue since our second Nehalem platform didn't exhibit any issues. Unfortunately we didn't have access to the more mature platform for very long at all, meaning the majority of our tests had to be run on the first setup (never fear, Nehalem is fast enough that it didn't end up mattering
So Hector, whatcha gonna do now? AMD has a few rabbits to pull out of their hat, right? :shocked:
looks awesome. I was a convert to Intel only 2n half years ago. I hope to see AMD throw in on this withbobcat for the mobile arena and whatever wild animal name chips they have coming up..forgot the names..
tomcat?
AMD get it together:rolleyes:
Well well, Intel certainly wants AMD gone...don't they?
I'm not suprised about this , it's been 2years sincs conroe so, normal speed boost and not impressed at all :)
The Emperor has arrived, well almost, good to see that we have a solid speed bump incomming.:clap:
I just pray that (like alot of ES chips) performance dosen't go down, it's happened before!....all this early stuff gets us all excited but then you get retail and performance is half what early numbers suggested.
This is really more than I dared hope for!
Incredible!!!
:shocked:
Even if it goes down, im sure the performance will go up when memory latency aint high on a faulty board and running in singlechannel mode.Quote:
Originally Posted by Richard Dower
Performance can only go up.. :D
I think a picture is in order.
http://shintai.ambition.cz/pics/crimageht5.jpg
as if this matters most in nehalem and k10 architectures..... btw none of these benches are high memory bandwidth related. cache latency on nehalem is very good and that is exactly what k10 is missing, the 3th memory channel will only be added value in server platform.Quote:
Originally Posted by Shintai
same for all the tests, scores are very good but then again only high tests that in old p4 days ran very well with hyperthreading
btw you already changed your statement from nehalem will consume less/equal against current york to better performance/watt.... :D
Quote:
Originally Posted by eleeter
Seeing as most software isnt multi threaded yet possibly? Im interested in seeing some single threaded performance also
Did you read the article? Also what part is the hot thing currently? CPU or PCIe NB? :rofl:Quote:
Originally Posted by duploxxx
Plus everything is still a good 3 months atleast from release.
Looks great :D, however Anand said:
This is not true. I ran some of the benches he lists just seconds ago and was able to beat his Nehalem scores with a QX9770 set at stock 3.2Ghz. (not by much though, although I am stuck atm to also run my RAM way below stock settings due to MB issue)Quote:
Note that Nehalem continues to be faster than even the fastest Penryns available today, despite the lower clock speed of this early sample.
Maybe he talked about release frequency. Perhaps 3.33Ghz.
I didn't convert to Intel in 2006 or later - I haven't even tried a Core 2 platform, Conroe or Penryn.
What do you guys think it'll be like going from a 90nm AMD Athlon to this? :rofl:
Should be pretty damn fast, but at a certain point (maybe starting with Nehalem) the performance gains pretty much become imperceptible for everyday use..Quote:
Originally Posted by bowman
I.e. you can tell the difference between 2 seconds and .5 seconds easily
Then things go from .5 to .25 seconds.. okay you notice.
Then things go from .25 to .125.. hmm less likely to notice.
Soon enough it's imperceptible, and you don't have enough workload to justify/exhibit the 2x performance gains anymore.
Quote:
Originally Posted by Shintai
clock for clock Nehalem is going to be huge :D
what I took issue with was the statement that this ES was faster than any Penryn today.
Quick example:
Nehalem result vs. stock qx9770 result (with terrible RAM settings)
this is not a clock/clock comparison.Quote:
Originally Posted by ceevee
let them shut down SMT and see if it is clock/clock faster and for cinebench with a mem controller it doesn't matter how good or bad the settings are it is cache related where nehalem puts a nice score.
:rofl: Hurting? 2.66 VS 2.66 is not clock for clock?Quote:
Originally Posted by duploxxx
:up: 4 physical against 4 physical + 4 virtual in benches where smt is actually working....Quote:
Originally Posted by gallag
thx but i prefer real world unless proven that this time SMT is actually good for daily use.
Duploxxx is on stage one of the kübler-ross model :rofl:
If you do the Cinebench math based on my stock result then 2.66Ghz Nehalem = 2.91Ghz Penryn. If you do it based on Anand's result then 2.66Ghz Nehalem = 3.29Ghz Penryn. So maybe this is why he said it was faster than any stock Penryn? :confused:
However it begs the question what the hell kind of screwed up mess did he use to get those Penryn benchmarks. They are terrible but he obviously knows what he is doing, so I could only think he lowballed the Penryn results on purpose to make the story more sensational and dramatic.
Hyperthreading was less useful back then when every desktop CPU was single core, nearly all apps were written for that, and the architecture was gimped to begin with..Quote:
Originally Posted by duploxxx
This is a vastly improved Core 2 with SMT, and it'll debut in a market where even every single low to mid-end laptop and desktop PC has a dual core CPU.. So I'm thinking it'll see much more use.
wants to see ocing!
just hopes these lga1366 chips and mobos won't be too expensive :D
Quote:
Originally Posted by s0lid
I wonder if the memory controller will introduce another dreaded cold bug like the old AMD chips had.
This makes me wonder, now that they finally implemented everything AMD invented so far what will they do next?
Intel is like Apple they take loads of good ideas, combine them together and then they do them right.
Performance leadership: Intel > AMD
Innovation leadership: AMD > Intel
Them two working together makes things so much better, if only Intel wasnt so touchy about market share.
Sorry man, I just don't see the point in not allowing advantages. Would you want to disable sse4 as well? I mean if they have put smt on there cpu's to make them faster then why b1tch about it?Quote:
Originally Posted by duploxxx
i see people saying the end is nigh for amd....
well guess what? thats bad for us if its true - with amd gone why would intel be bothered improving? i dont see any other competetor around...
Hmm interesting point there.Could be the case,but we need to see the actual independent confirmation(or negation) with the retail production units.At least we now roughly see what kind of numbers Nehalem produces@2.66Ghz,so users with Penryn can compare and see for themselves,as you did.Quote:
Originally Posted by ceevee
:rofl::rofl::rofl::rofl::rofl::rofl::rofl::rofl::r ofl:Quote:
Originally Posted by Syn.
Man are you for real. Did Intel not use an imc first anyway.
Looks like there is going to be a lot of sore fanboys about here before the nehalem dust settles .
what did you do? joined the IT world in 2006? in most cases you were better of shutting down hyperthreading performance wise... and in worst cases it was even mandatory unless you wanted locked databases.Quote:
Originally Posted by gallag
sure keep on posting.... and most of it keep on dreaming, i am confident that i will work on this new architecture before you can even pre-order them, hence in your case it will always be dreaming and posting.Quote:
Originally Posted by Shintai
I hope Mr.Anand was sleepy...
CPU / Everest Ultimate 4.50 Memory Read Memory Write Memory Copy Memory Latency
Nehalem (2.93GHz) 13.1 GB/s 12.7 GB/s 12.0 GB/s 46.9 ns
Core 2 Extreme QX9650 - Penryn - (3.00GHz) 7.6 GB/s 7.1 GB/s 6.9 GB/s 66.7 ns
That would really really suck... I want 50GB/s read.... So he should have got 25GB/s before me being satisfied..
I don't follow you, You said that it was not a clock for clock comparison and when asked for your reasons why, you went back it time to the pentium4 architecture. Whats up with that?Quote:
Originally Posted by duploxxx
The testingboard only have SINGLECHANNEL working. Plus higher latency than another board. But they couldnt keep the other board long enough. SO all tests are done on the faulty.Quote:
Originally Posted by M.Beier
Yea, so you are saying that Nehalem isnt following in the same footsteps that AMD started with K8?Quote:
Originally Posted by gallag
Please for future references i am not a fanboy and as such dont feel bad about anything. It actually made me quite happy to see Intel bring out another great performing architecture. I just thought i might bring in a bit more constructive post to this thread then another "ROFL!!! GO Intel!!! AMD SUXS ASS!!!".
Here are some quote's from the article that i think you might find interesting after finding my post so funny.
Quote:
Nehalem implements a very Phenom-like memory hierarchy
Quote:
While Intel did a lot of tinkering with Nehalem's caches, the inclusion of a multi-channel on-die DDR3 memory controller was the most apparent change. AMD has been using an integrated memory controller (IMC) since 2003 on its K8 based microprocessors and for years Intel has resisted doing the same, citing complexities in choosing what memory to support among other reasons for why it didn't follow in AMD's footsteps.
and finallyQuote:
It was at the same dinner that Pat mentioned Intel may do a chip with an integrated memory controller much like AMD, but that an IMC wouldn't solve the problem of idle execution units - only indirectly mitigate it. With Nehalem, Intel managed to combine both - and it only took 6 years to pull it off.
Quote:
Pat also brought up another very good point at that dinner. He turned to me and said that you can only integrate a memory controller once, what do you do next to improve performance? Intel has managed to keep increasing performance, but what I really want to see is what happens at the next tock.
K8 copied the IMC from 386SL and 486SL.
Phenom/Barcelona copied its cache hierachy from Itanium and other big tin.
The only thing "new" is basicly the HT.
In short, you make no sense Syn...unless you only look over a tiny tiny period of time and exclude other products.
Unless you're doing distributed computing or something that requires 100% load for a few hours at a time, you'll notice nothing.Quote:
Originally Posted by bowman
Come on mate, You make it sound like Intel does nothing but copy AMD and have never thought of anything themselves.Quote:
Originally Posted by Syn.
What did they copy exactly?
I see you're living up to your nickname, "ShINTEL" as usual. Spread some more FUD if you wish.Quote:
Originally Posted by Shintai
What part is fud? Please elaborate.Quote:
Originally Posted by freeloader
Would it be this part?
http://en.wikipedia.org/wiki/History_of_laptops
Or would it be this part?Quote:
introduction of the Intel 386SL processor, designed for the specific power needs of laptops, marked the point at which laptop needs were included in CPU design. The 386SL integrated a 386SX core with a memory controller and this was paired with an I/O chip to create the SL chipset. It was more integrated than any previous solution although its cost was higher. It was heavily adopted by the major notebook brands of the time. Intel followed this with the 486SL chipset which used the same architecture. However, Intel had to abandon this design approach as it introduced its Pentium series.
http://en.wikipedia.org/wiki/Itanium
Quote:
From 2002 to 2006, Itanium 2 processors shared a common cache hierarchy. They had 16 KiB of Level 1 instruction cache and 16 KiB of Level 1 data cache. The L2 cache was unified (both instruction and data) and is 256 KiB. The Level 3 cache was also unified and varied in size from 1.5 MiB to 24 MiB.
Quote:
Originally Posted by Shintai
http://www.xtremesystems.org/forums/...8&postcount=51Quote:
Originally Posted by gallag
I am limiting my references to Nehalem only seeing as this is a thread about Preview Article of Nehalem ES CPU's. Nehalem is a combination of everything that AMD and Intel invented in the last decade, if you disagree please explain why.
Has anybody somekind of idea about the mobo and proc pricing?
Nehalem looks like a product I would want.:yepp:
Depends what type. For X58+Bloomsfield I would expect it to cost as X38 boards and CPUs from 400-500$ and up.Quote:
Originally Posted by kiikkuja
Lynnsfield starts under those 4-500$ and the Auburndale for alot less.
first consider this - the tri chanel nehalem setup had one extra stick of RAM. that's what 5-10 W right thereQuote:
Originally Posted by Shintai
they have 6 months to improve manufacturing...
do you really think that nehalem will launch using more power at a given clock speed? chanches are the chip itself will launch with TDP 5-10W lower than penryn.
this is quite a leap, it's what, 30-50% faster than conroe...
Quote:
Originally Posted by ceevee
Your score is based on vista 64, Anand results are on vista 32, it is a known fact that Cinebench 10 runs faster under 64bit OS.
OT: looks like Anandtech is running out of bandwidth, pages take forever to load or won't load at all...
I'm not moving to Nehalem until it hits 32nm.
The bloomfield one as it is the poweruser/enthusiast model/package is it not?Quote:
Originally Posted by Shintai
So the overall cost with 3 sticks of RAM would be around 1000€ or over.
Man, if you're NV this has to hurt. Imagine the biggest party of the year, and you get a note saying: "sorry, but the invitation we sent to you earlier was a mistake. Please accept our apologies."
Look at the table with 3dsmax score breakdown
http://anandtech.com/cpuchipsets/int...spx?i=3326&p=7
Only ~10% faster in SpaceFlyby?
I guess that represents Nehalem single-threaded performance improvement :/
Could be better with mature mobo's, but I guess we shouldn't expect too much.
Damn that's some insane performance. Too bad it won't be oc'able. Hopefully someone will figure out some mod so we can oc. 20x multi would be crazy.
But it can OC....who told you it cant? Fuad?Quote:
Originally Posted by awdrifter
About a 30% perf increase is what I would have expcted. Nothing exceptional.
It will take one year before a nehalem can beat a 4gig wolfie for gaming ;)
Unless you OC the nehalem too :DQuote:
Originally Posted by RealTelstar
that + power saving features on the nehalem boards where most likely not working properly yet.Quote:
Originally Posted by xlink
Is it just me, or does this so-called "discussion" sound a lot like those fights we had when Conroe ES was reviewed for the first time ?
why? because most applications today are single threaded? :rolleyes:Quote:
Originally Posted by eleeter
its nice to see nehalem own in multithreaded benches, but we all knew that was about to happen....
But i want to see single threaded performance also
This quote from a month ago stands my point.Quote:
Originally Posted by Metroid on May 3
This is a great news and more important it came from a very trustful source. The issues were very well explained.
Single threaded applications should not really be taken in account as most of the applications use 2 cores, unless somebody is still living in the stone age.
Nehalem only tends to get better and better with new bios. What Intel really wants to show to developers is that they could adopt more than 2 cores example(4 to 8 threads scalar).
It is the Multithreaded era, AMD and Intel are together, adapt or die soon or later.
Metroid.
Same here. I didnt get Core 2 untill it hit 45 nm. The earlier releases tempted me greatly, but my S939 system was still running everything out there back then and I didnt need to upgrade.Quote:
Originally Posted by Carfax
I'm sure that my E8400 will be good enough for another 12-24 months, and I can wait for Nehalem to fully mature and improve before getting one.
As for AMD dying, dont be too sure on that. The enthusiast community only makes up around 1% of the market. Average Joe buying a PC doesnt know which is better out of Intel and AMD, and also there are plenty of AMD fanboys that wont touch Intel CPU's no matter how much better they are.
There are far more people that play PC games that still buy Phenoms over a Core 2 Duo or Quad for whatever reason they have then you would expect, and refuse to listen to advice or accept that Intel's CPU's are better for games. Fanboys will be fanboys, regardless of how much faster one companies products are over another.
I myself used AMD for 10 years. My last Intel CPU was a Pentium 100 and since then I have been using AMD. But I wasnt going to buy an AM2 setup over my previous S939 when there was virtually no difference at all other then the type of memory, whereas the Intel setup geve me a huge performance increase over what I previously had.
They will be great in single threaded applications also, they arent going to perform any worse then the Core 2 Quads. Hopefully we will start to see much more use of multi threading and multi core CPU's now. They have been around for ages and are used in most PC's today, whether it is 2 or 4 cores. Now is the time for software developers to make the step towards multi threading and to take advantage of the technology we have available.Quote:
Originally Posted by GoThr3k
I need more cores. That is for sure.
Rendering is my hobby, and a good render used to take over 12 hours on Pentium 4. :p:
(Try Vue full scene rendering with Ultra High quality.)
It was reduced by literally half when Athlon X2 was introduced.
It was further reduced by Core 2 duo, and then Core 2 quad.
Still, 4 cores take more than an hour which is still a PITA to work with. My dual Barcelona system scored similar to Q6600, so, I didn't bother using the platform except for WCGing.
Now, my dual E5410 is in making, I should be able to get rendering done in less than an hour. :)
Though it will only be a matter of time that I become dissatisfied with 8 cores. I suppose that's where Nehalem comes in, and perhaps I can get rendering done in a half hour?
Oh..., dear lord. My dream is perhaps within my reach.
Oh look. Yet another fast CPU.
That's just great. Now, if there just was atleast some use for them...
You gotta buy a couple of GTX 280 and you'll do it in a lot less time.Quote:
Originally Posted by alucasa
This test would be commonly used, if you were only allowed to use one cylinder on 90% of roads.Quote:
Originally Posted by eleeter
It's CPU task.Quote:
Originally Posted by metro.cl
Though the rig does have a Gefroce 7800gs on it for watching DVD.
However, Vue software requirement states no video card, and video card has little to do with rendering at least with Vue, Poser 7, and Carrara 6 Pro.
Quote:
Originally Posted by alucasa
You don't know how to do it really right. ;)
(Ex Vue enthusiast passing by)
That is my point. Intel has done more than enough, now is up to developers. The benefits are very well shown in this great article by Anantech, also 3D benchmarks are sweet and that will attract high end gamers who look for the best system.Quote:
Originally Posted by largon
Servers are the breakthrough here, transactions done using ERP systems that have implemented EDI technology eliminating the bottleneck or deadlocks.
Metroid.
I missed the Pending release date?
Thanks
:rolleyes:Quote:
Originally Posted by Macadamia
P.S. I need to upgrade my Vue to 6... Not this month though.
Q4 2008Quote:
Originally Posted by Think
Bulldozer cannot come fast enough for AMD.
And even then, with numbers like this, it may be too little to late.
THESE TESTS ARE TRIPLE CHANNELQuote:
Originally Posted by Shintai
read thorugh the article, they use DDR3 1066, MAX theoretical BW on a single channel is 8533 MB/S, there is NO way they get 12 gb/s with a single channel :shakes:, maybe it still isn't tweaked enough but AMD barely hits 11gb/s on DDR2 1066 so these tests have to run at triple channel, maybe dual if intel found a miracle way to get close to 100% efficiency :rolleyes:
looking good, AMD'S going to be in the same position they are today if they manage to launch it at the same time amd'S going to release K10 45nm (K10 done right) with higher NB frequencies, higher clocks (maybe 3ghz).
intel's going to keep the lead for a longer time, however we still don't know the performance of their dual channel parts and how fast their NUMA arch is (we still don'T know this for amd either as 45nm finally brings HT3 to the server market...)
going to be some interesting months with all these new CPUs and GPUs, can'T wait until this stuff gets released :D
Lets try again...Quote:
Originally Posted by generics_user
http://anandtech.com/cpuchipsets/int...spx?i=3326&p=2
Quote:
We had access to a 2.66GHz Nehalem for the longest time, unfortunately the motherboard it was paired with had some serious issues with memory performance. Not only was there no difference between single and triple channel memory configurations, memory latency was high. We know this was a board specific issue since our second Nehalem platform didn't exhibit any issues. Unfortunately we didn't have access to the more mature platform for very long at all, meaning the majority of our tests had to be run on the first setup (never fear, Nehalem is fast enough that it didn't end up mattering).
Benchmarks=singlechannel. Not the synthetic everest. The everest is also done with a 2.93Ghz nehalem.Quote:
The motherboard implementation of our 2.66GHz system needed some work so our memory bandwidth/latency numbers on it were way off (slower than Core 2), luckily we had another platform at our disposal running at 2.93GHz which was working perfectly. We turned to Everest Ultimate 4.50 to give us memory bandwidth and latency numbers from Nehalem.
Note that these figures are from a completely untuned motherboard and are using DDR3-1066 (dual-channel on the Core 2 system and triple-channel on the Nehalem system):
Actually it may work out the other way around.Quote:
Originally Posted by Richard Dower
If Nvidia does not get the ability to use SLI with these chips then how many of you will go for ATI GPUs so you can max out your setups ?
why do people think that after Nehalem there will be no other CPU on the market!? I haven't noticed that BMW M3 is the only car selling out there?Quote:
Originally Posted by mstp2009
omg!!! I have been waiting for these a long time, so on the overclocking side? cant wait to see how is it going to be since there is no FSB anymore:up:
That BMW M3 got quite cheap didnt it? :DQuote:
Originally Posted by Nedjo
Alot more people are going to buy intel when they can get this type of performance than get an AMD so they can run two cards.Quote:
Originally Posted by Buckeye
SLI and crossfire are so niche, I'd imagine less than 15 percent of cpu game to begin with, and probaly 1 percent of that runs dual cards.
But computing power, everyone wants that. Just look what happens when some cpu has crazy performance on it compared to the competition(core2 = nearly bankrupt AMD).
You haven't? D:Quote:
Originally Posted by alucasa
Now something's really wrong. The only long renders I have was on bad spectral configs. :yepp:
Hence: 6 is uber slow (5+hrs for default high, 800*600 on my Brisbane), Vue 5 only took 24 hours on my Barton with high supersampling for a 2k res render. :ROTF:
don't know, you tell me?Quote:
Originally Posted by Shintai
Nehalem doesnt equal a BMW M3, perhaps some of the top bins. But Nehalem strecthes all the way down to the lowest CPU.Quote:
Originally Posted by Nedjo
Building a 8 core Harpertown rig this month. I am gonna be on tight budget throughout this month.Quote:
Originally Posted by Macadamia
Gamers will never know our pain.
Too bad that Nehalem is 4 physical and 4 logical though.
Off topic, when is 8 physical-core coming ? What codename is it under ?
Well that just changed my roadmap for this year....
Q4 for servers, not desktop, or only very limitedQuote:
Originally Posted by Shintai
Oh wow I know where my money is going in Q4 fo-sho now :D:D:D