AMD's "Puma" Mauls Intel

Boy, you guys all need to relax.

It doesn't matter which company you like better, it doesn't make you better just because you like it.

 

I never said Intel wasnt good performance, its the other way around. I like Intel's performance, but you Intel Fanboys find every AMD related article and just belittle the company to nothing. Find something better to do.

What words exactly would I be eating? Nehalem is moving to surprise surprise the architecture AMD uses, because Intel maybe decided to stop decieving the masses by saying they have a 1066 or 1333Mhz FSB when its really just 266 or 333Mhz. Which that is why Intel was fortunate enough, dare i say lucky enough that manufacturing technology has granted them the grace to create chips at 65 and 45nm so they could actually get all that L2 cache on there. If they weren't pushing the envelope of manufacturing technology they'd be mediocre.

I don't mind using Intel products, they do very well, but so do my AMD products.

 

Dude, Intel was crushing AMD even with the handicap of FSB and the lack of an integrated memory controller. Add the two, plus an even more efficient microarchitecture design, and try to imagine how badly Nehalem will crush anything AMD has on offer.

Maybe AMD can just skew the numbers again to make it look like they aren't losing too badly.

While the Core 2 architecture does rely on L2 cache to a higher extent than AMDs offerings, it does just fine with lower caches. For evidence, look at the 2MB vs 4MB L2 cache at the same clock speeds, like T7250 vs T7300 or Conroe vs Allendale and note that the difference is not all that much. Cut the cache to 1MB, and the Core 2's still manage decent performance, far better than anything AMD offers at the same price (in notebooks, at least).

 

all is speculation, after all, its not like you work for any of these companies to justify your claims about the future that you're stating now.

 

About Nehalem, you should go look at some of the preliminary benchmarks of the engineering samples. Then come back, and talk about "all is speculation".

As for working for Intel, I don't, but even if I did, you'd never know either way.

 

Well, judging from the performance of the Lenovo X200, it looks like the OP was right to say that AMD beat Intel for the IGPs.

I don't agree with blanket statements made upfront before any data was released, but now that some test results are out, it appears that the OP might be right.

 

Oh, the OP was right about the HD 3200 beating GMA X4500 soundly, but the title is still incorrect - as a platform, Puma does not maul Montevina, or even, for that matter, Santa Rosa either. Looking at the review of the Lenovo X200, it gets 10 hours of battery life on a 9 cell battery, which scales to around 6.6 hours on a 6 cell battery. And the X200 is a fully featured notebook with a 2.4GHz processor and a 7200 RPM hard drive. Compare this to the 2.4 hours that you get out of a Puma equipped notebook, and you start seeing why people (including myself) are disputing the title quite a bit.

There are multiple aspects to performance, and power consumption happens to be one of them.

 

That's my opinon too. AMD's "Puma" IGP Mauls Intel's IGP would have been an accurate title.

Or: AMD Puma "mauls" Intel.

 

Exactly. Which is why I pointed out the IGP portion of the comparison.

Took me a few seconds to figure out the difference in your second suggestion.

 

I run internet explorer fast enough...but playing counter strike 1.6 is impossible on a x3100. Puma rocks Santa Rosa in CS 1.6. Intel + Nvidia GPU is not a Intel Platform. Therefore Puma > Intel's Same Old Crap. My Intel 810 has the same performance of my Intel 965 ten years later, that's great in terms of inNOTvation. Thanks for Puma AMD.

 

Inte'ls platform does not specify a graphics card, only the chipset (different ones for each integrated or dedicated graphics). Hence, the manufacters can choose what to do with their systems. They are not automatically foisted with a dedicated graphics card, unlike AMD and its HD 3200.

 

The age-old debate never ends, sheesh. Just buy what's best for you, be it Intel or AMD. Who cares what you purchase, you (any prospective consumer) are the one stuck with it. Why do people feel the need to justify their purchase to anyone? If you're happy, that's all that matters. If Intel or AMD works for you, great.

 

And how do you know AMD's next processor generation won't be as good as Intel's. Youre just assuming, but AMD is supposed to release a new cpu that will be much better than todays.

The Fusion project, should keep AMD competitive, and possibly more attractive, and is supposed to be released about the same time as Nehalem for laptops, a year from now.

 

Lets see, most of AMD's future roadmap in the next couple of years is based on the K10 architecture, which according to current findings, is a complete failure. You're going to try and deny it, and I'm going to laugh and throw benchmarks at you.

Considering K10 is so uncompetitive, I'm surprised Intel doesn't just milk the Core 2 architecture for a bit longer, no need to replace it anytime soon, really.

 

Since AMD canceled Bulldozer, moved K10 to being used in Montreal (8 core chip), canceled that, and then still decided that K10 would be the core of choice for some time to come.

You're assuming AMD won't delay/cancel whatever processor you're referring to (got code names?), which, is very likely, considering how late Barcelona and Phenom were.

Oh yeah, and AMD isn't to be trusted. AMD knowingly posted deceptive benchmarks. Seriously. Everyone complains about Intel being anti-competitive, especially AMD, and then they go pull this bull.

Enjoying trusting them.

Oh yes, and even they themselves haven't a clue as to what they're working on. See here, and then click on the two top links.

They've been like that for more than a year. Before it used to simply say "Page Not Found."

That's the epitome of AMD and their future products, and Barcelona and Phenom bring us enough truth for such an assessment.

The Fusion project, should keep AMD competitive, and possibly more attractive, and is supposed to be released about the same time as Nehalem for laptops, a year from now.

Fusion is based off the K10 core, which we've already seen Intel kick the crap out of. On top of that, Fusion isn't slated for released until late 2009/2010, while Intel's Nehalem slated for launch as early as this quarter, and with more chips launching within the next quarters thereafter.

And to finally note, AnandTech has tested a very early version of Nehalem on a very early and basic motherboard, and it's up to 50% faster than current Penryn chips.

Bode well for AMD, this does not.

Fin.

 

Nehalem is due out for servers later this year. And laptops in a year.

You all must be Intel shareholders and most likely shorting AMD if you must be this determined to prove some point. Did anyone say Intel has slower CPUs? No, youre just arguing that AMD sucks, and youre petty. AMD is the only reason Intel has such a low stock price today and gives you good products. Intel had to eat into their loads of money, made by selling you mediocre products at high premiums, to finance their manufacturing development and selling you prodocts at reasonable prices, doing die shrinks and really pusing the envelope of fabrication feature sizes to get an advantage. On AMD's budget they cannot compete with that, and could not have expected the technology to mature that fast to aid Intel.

The Opeteron processors have proven to be very good in multi-cpu systems, because of their Hypertransprt and ability to scale. Now Intel may get it right the first time with quick connect but the power draw looks to be 130W initially, and I'm thinking AMD will come in well below that and have HT3.0 at 4000Mt/s under 100 watts at the same time Intel releases Nehalem for servers. The K10 hasn't shown a lot for desktops, but its real nice for multi socket servers and workstations, people that do more than burn DVDs or make MP3s.

In Fusion you may see some greater benefits out of K10 when throwing a GPU on the CPU and gets rid of the Northbridge all together, so the K10 will be utilized more, taking better advantage of the Hypertransport and direct connect.

IF AMD is so useless why even take your time to put them down?

 

Yea, except Nehalem (Havendale and Auburndale cores, I believe) also have GPUs on the CPU, and also have a faster QuickPath Interconnect, so while Fusion will perform better than Phenom (one would hope so, at least), it still probably will end up competing on price and not actual performance.

Which simply does not work, since it costs AMD more to make their chips anyways (less dies to a wafer, due to a larger 65nm manufacturing process - Intel has been producing 45nm chips for a while now) Competing on absolute price doesn't work in many different industries, just ask GM, Ford, and Chrysler.

 

Except Shanghai is set for 45nm later this year for opterons and Fusion is also set for 45nm mid next year. And one review I looked at showed Intels new chips being pretty large.

 

The package itself is large, yes, but the die is still 45nm.

 

Shanghai may be 45nm, but nevertheless, it's the same design as Barcelona/Phenom.

 

lol too bad the topic still says amd puma mauls intel. q.e.d. lol !

 

I wouldnt argue with that, but they have larger L3 cache and will consume 15% less power, and offer higher clock speeds, and also use newer form of lithography that should lower production costs.

So if I have a 2.8Ghz Opteron at 95 watts or a 3Ghz Nehalem at 130 watts, which might I use? Its may be a tough call, and hopefully Intel will have lower power than that, but AMD may still be able to claim a "green" factor initially.

 

Where did you pull a 130W TDP figure? We actually have benchmarked evidence to back up our claims about Nehalem's performance.

Besides, the 3GHz Nehalem will blow the Opteron the hell out of the water performance-wise anyways, so I'm not quite sure that they are comprable.

 

Hehe for some reason it says you posted the previous comment.

 

I have always been a fan of AMD, my first three computers were equipped with AMD procs and I would love nothing more than to see AMD get as fast or faster than the new Intels.

Although it seems that their new IGP is awesome, the proc is still just too slow.

 

Just curious for people like you who say the processor is just too slow, what is it that you are trying to do that the processor is too slow?

 

Encode 5 HD videos at once

 

Yeah right.
But I get what you're saying, that if someone really did try that, sure they'd notice a difference, otherwise not.

 

Well true, you have me their I am not usually doing anything that intensive that I need to get a 2.8 GHz C2d, I guess I just want the fastest. I have my eye set on the new T series which I don't think will have any AMD proc options anyways so probably a moot point.

haha, yeah that.

 

well I pulled that from the same place you posted your knowledge from, the ever reliable wikipedia. Gotta love a place where you can type whatever you want and it will be accepted as fact.

You can compare certain performance criteria for a server cpu, and AMD still does well with floating point, and virtualization. And if they can produce a product that performs 85% as good at 70% of the power, then they could be considered a good investment to me. They wont be the fastest but it would sure be a lot cheaper too. So they will try to compete on price and performance per watt.

45nm feature size, large caches, and HT3.0 set Shanghai well apart from Barcelona.

Maybe when AMD wins their court case with Intel in like 4 years, then AMD can compete on the same manufacturing technology level and become the performance leader, but thats too far to think about.

 

Whoa, you really think that anything on K10 architecture (Barcelona, Phenom, Shanghai, etc) will perform 85% as well as NEHALEM? Wow, my friend, thats...not really close to happening, methinks. Phenom doesn't perform 85% as well as Penryn right now, and Nehalem's a whole 'nother ball game - AnandTech conservatively estimates between 20-50% performance increase for Nehalem over Penryn. Can a 45nm shrink of Phenom/K10 really perform all THAT MUCH better than the 65nm? Better, for sure, but it won't net you at least a 30% jump instantly.

Wow, thats weird. I fixed it now

 

AnandTech's first estimates were totally wrong, based on invalid Penryn numbers, and Anand came back later to correct them. Nehalem is only 5-15% faster than Penryn. Shanghai may be 10-20% faster than Barcelona. I suspect Nehalem will still be fastest overall on desktop, but the gap will be very narrow.

 

I used to always have AMD Processors in my desktop PC's (don't have a desktop at the minute) but the AMD procs used to perform better than their intel competition of the same price.

Now i'm not really up to date on processors at the moment but the general feeling I have gotten from people from both the laptop and desktop world is that these days you will get a C2D that is a lot faster than the eqivilliently priced AMD..

So have AMD lost their appeal of costing less and being as fast or faster?
What would be say the AMD equivillent performance wise to say an Intel C2D Quad Q6600 (I know this isn't top of the range and thats the point). Most people won't pay more just to support a company these days. They will go with whoever can offer them the best performance for their money (Say about £120 for a Q6600 for your average user spend on a cpu??)

 

The AMD equivalent of the Q6600 is Phenom X4 9850 or 9950. Somewhere around there.

 

The very last desktop I built for myself I switched to Intel and got a 3.2 Ghz P4, it was the worst processor in terms of problems I had ever had.

I heard Core 2 Duo is the one to go for now though? Are they noticeably better at a similar price?

 

depends what you do if any cpu choice will have a noticeable difference. I personally think that 80% of consumers would never even utilize any dual or quad core cpu to its 100% utilization.

Yes the C2D has the performance lead, but i just got a 45watt 2.5Ghz Athlon X2 for my 780G chipset desktop, integrated HD3200 motherboard, and you can get that setup for like $150, and I am VERY happy with it. You can get faster X2s for cheaper even, and their only 65 watts.

 

How about some evidence to back up your claims? Especially the later AT correction on the Nehalem numbers.

Also, AT tested a very, very early version of Nehalem silicon. Expect performance of the release silicon to be much faster than what they have tested.

@TommyB0y - we know that AMD can only compete on price, and not performance. I've been saying it all through this thread. Its just that all of their price cuts are going to come back and really hurt them (they already have) since they have less margins than Intel on the products they sell also.

 

And as I already said before no one ever said an AMD cpu was faster than an Intel cpu, but it doesnt mean AMD doesnt offer a good value and good performance. I think everyone exagerates how much better Intel is, seeing as even those people probably dont need another 10% of performance, but maybe a few times a year to save them less than an hour of their life.

You might think that AMD cutting prices can hurt them, but think about how much a few dollars cut would hurt Intel. If Intel sells a billion products and only makes profit of $5B per year, then cutting a few dollars off of their products gets rid of all their profit too, they need the near monopoly marketshare to justify giving you reasonable prices. Thier margins aint so hot either, and they dont like giving you things at reasonable prices, so whats to happen if AMD stops competing and Intel continues doing things to prevent AMD from competing like offering exclusive pricing?

 

See Anand's comment at the top of this page, and the bottom two comments on the previous page.

http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3326&cp=8#comments

Anand mis-reported the Penryn numbers, making it appear that Nehalem was a much bigger improvement than it really is.

 

I just saw the review for the x200, and I find it funny that someone comopared the battery life of it to the DV5z.

X200 is a 12.1" 1280x800 200 nit screen said to not be bright enough to use outside notebook with 9 cell battery and x4500 graphics that would get 7 hours of battery life when actually using it. And you compared battery life to a 15.4" with an awesome 1680x1050 screen and 6 cell battery with graphics twice as good.

Not to mention that it probably costs 3 times as much when it comes out.

 

$1199, and most of what you will be paying for is the quality, reliability and durability of a ThinkPad.

There is so much else wrong with the rest of that post, I don't even know where to start, but I'll try. You concentrate far too much on the screen - the tx2500 has a similar screen to the X200 and struggles to break 3 hours on a 6 cell. The X200 nearly breaks 10 hours on a 9 cell. Yes, the graphics are roughly 1.5 times more powerful, even twice as powerful, but the battery life is less than half of the X200.

As for an "awesome 15.4" screen", uhh, have you actually looked it? Its above average, at best - there are far better screens out there, from Sony, Apple, and Dell (the LED backlit option on the M1530). Awesome implies standout, which the HP is not. Solid, for sure, but not all that extraordinary.

 

Where is it $1199? I cant even find it on Lenovos website and the older X61s still cost that much and an X300 is $2500.

The TX2500z has a touch screen, so its not really the same except they are both 12.1". And battery life was at 3 hours 40 minutes with 6% left with low brightness levels and light work, with the 6 cell it comes with, and there is an 8 cell option.

The X200 achieved 10 hours because it was idling for more than 8 hours of it in the review, no work being done on it, so was the screen even active? What kind of stupid review is that. thankfully they went on to mention that it would last about 5 hours under load and light duty probably 7-8 hours. And that is with a 9-cell battery, which if you look at some examples a high capacity 9-cell can get almost twice as much as a cheap 6 cell in the same notebook.

 

I just love this article, one guy actually gets it, compared to the Intel biased writers.

This is Opteron related though.

http://computerworld.co.nz/news.nsf/tech/2A719E294658D35ACC257458007A61C9

 

You're incorrectly presuming a lot, BSing a lot of numbers, refuse to see the whole picture, and to be completely honest, its getting quite tiresome. You want to think you're right, and AMD is in fact better than Intel, and while I want to tell you that you are sadly mistaken, neither of us will actually change our positions on this matter, so continuing the conversation is pointless.

For the record though, it said multiple times in the X200 review discussion thread that the base price would be $1199. Before you mention it, the 9 cell battery upgrade should only cost around $50 based on the pricing of the extended battery option in the X60/X61. Order your X200 on August 5th

 

as someone who recently placed a purchased on the hp tx2500z,i sincerely hope you're wrong about the battery life of that unit. having owned previous amd laptops before, and getting 3 hours worth of battery with them, i dont know where you got your low numbers from. considering the new puma platform should improve power management substantially over the previous generations.

as far as cpu power is concerned, anything able to a) multitask without noticeable slowdown b) play wc3, wow , hl2, emulators c) not too much heat is okay with me. so therefore anything over 2 ghz is overkill for laptops in my opinion.

one reason i went for the tx2500 is because its the only laptop thats cheap enough and meet the requirements above. honestly what good is a 2.4ghz intel cpu going to do you with crappy intel igp? you arent going to see that speed in games, you arent going to see much speed difference in web usage or general usage. maybe if you do alot of video encoding or photoshop, but honestly not many people care for those things.

i own a e7200 desktop at 3.6ghz and a e4300 @ 3ghz. the e7200 has a gf7900 while the e4300 has an integrated gf7150m. let me tell you from personal experience the extra speed on the e4300 is totally wasted. cant play games, not much application needing more than a 1.8ghz c2d. so i turn on speedstep on the e4300 to actually lower its speed when its in idle. pure waste of cpu power without a graphics chip worthy. now the new puma platform is a godsend in my opinion. 2ghz cpu combine with hd3200 is more than enough to destroy any intel chips with integrated graphics for my usage. and thats why i choose puma platform.

will have to wait until the 24th form my unit to come and then ill find out which is better. but overall i have no regret with puma, although i cant say the same had i gone with intel/x3100.

 

Your problem is, that I have never claimed that AMD was better than Intel, just that AMD is as good as and can be a better value in some cases. Value is more than price, its price for performance.

You are the one that came to this thread blasting AMD, throwing out numbers you thought were good, stretching the truth, making terrible comparisons, and if youre tired then stop typing.

I dont even know what it is I incorrectly presumed, when it was all from reviews.

 

Quote from the battery part of the NBR review of the tx2500:

Thats where my numbers came from. *shrug* If you choose not to believe it, thats not my problem, but having been a member of this site for over two years, I feel inclined to believe the figures they quote. I will attest to the fact that most notebooks are almost unusable in the Power Saver mode, whether they be Core 2 Duo, Pentium M or Turion X2 based systems.

@TommyB0y: I am merely trying to say that AMD is not as good, and their price/performance is not as good as you would like to say, since you're definition of performance is not necessarily anyone else's definition of performance, and regardless of that, is an inherently fallacious notion of performance, since it does not incorporate all areas of performance, and not even some of the more important ones.

What you are presuming to be wrong? You are presuming conditions of reviews that were not explicitly stated, in one occasion on this page. You are presuming the specs of the batteries without knowing for sure about them while trying to state conclusions based on that presumption as fact. There are multiple other examples in previous posts that I don't feel compelled to dig up, but I'm sure you get the idea anyways.

These "terrible comparisons" you speak of are comparisons that would make sense if you bothered to think about them in the slightest. My numbers are good, and are all backed by evidence given by reputable tech sources on the web. I dare you to challenge them. I don't stretch truths, thats AMD's job.

However, as this is getting quite personal, I feel that it might be advisable to discontinue this line of conversation.

 

yes you should discontinue redline, because you are way too motivated to insult AMD.

there are other reviews on this very site where all of my information came from, and as you have found out not all of the reviews for your Intel benchmarks were true.

 

If you bothered to read the link, Anand said that it was an excel error as to why the graphs came out wrong, however, the Nehalem benchmarks and the conclusions of 20-50% performance increase still stand. However, you obviously didn't catch that part, since you still feel compelled to call them falsified. Really? False benchmarks from one of the most reputable hardware review sites on the internet? I don't think so.

 

Actually, you both should stop. TommyB0y, you are just as guilty to throw insults. It is not going to lead either of you to anywhere good. Each will still think his stink smells the best and the other will still deny the other's good points.

So, best to put this argument (for 'tis no longer a debate) to rest.