The missing Conroe benchmark
Right after FACE Intel's field day, Intel fanbois had their field day as 32 bit Conroe benchmarks popping up yet again for the 100th time. Nor surprisingly, Conroe's lead over Athlon 64 diminished a notch again. A previous 20% lead is now reduced to 5-15% lead. But the noise this time gets louder.
Where are the 64 bit benchmarks? Vista is coming in just a few months, and there is no 64 bit benchmarks for Conroe? Or, only Japanese know how to do 64 Conroe benchmarks? (shows Conroe 40% slower running 64 bit code, EM64T really stands for emulated 64 bit?).
The Intel fanbois will tell you: "I am happy with 32 bit, I will never need 64 bit, I will stick to IA32 for the rest of my life".
The new Conroe benchmarks only proved one thing: the Netburst P4s Intel was pumping are total crap. But Intel is stuck with $7 billion of that crap and is still producing majority of its chips with that architecture. By the time Conroe reaches 40% of Intel's volume in five months, AMD's 65nm products will be out. Before that, AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform.
According to this, AMD's 65nm parts will out for testing in August.
posted by Sharikou, Ph. D. @ 7/14/2006 01:44:00 PM
103 Comments:
"AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform."
This will only happen in your fucking wet dream, you sucker.
What a shame. Trying to relate Conroe to a decace old Face Intel website.
Dude, stick with current tech, not five year old crap. Your know, pervasive 64 bit stuff... Wake up. Your blog might survive.
"I want to give everyone an update on how things are going with AMD's 4x4 project.
As you know, we are developing a new platform for the PC enthusiast market. We see the future as decidedly multi-core and multi-threaded.
In our technology development, we are seeing an 80% average performance uplift from four core over two core AMD64 processor-based systems on multi-threaded benchmarks including Futuremark 3DMark 06, Maxon Cinebench R9.5, and WinSAT running on the latest beta release of Windows Vista.
So results are promising! Looking forward to sharing more as things develop." - David Schwarzbach, AMD.
The 4x4 platform will place two physical sockets on a motherboard, connected by AMD's Direct Connect architecture. Mounted on each socket will be an AMD Athlon 64 X2 processor, for a total of four cores.
While Schwarzbach declined to comment on when the 4x4 platform would formally roll out, an AMD spokesman implied it could be in time to steal some of Intel's thunder: "You'll be hearing from us soon."
~~~
AMD has confirmed to us earlier today that indeed they will be bringing price cuts in response to Conroe. Their goal is to price the CPUs so that they retain the performance per dollar leadership.
Schwarzbach declined to comment further, citing the quiet period mandated by the Securities and Exchange Commission in the period immediately preceding an earnings call.
Could you give us an example of the type of 64 bit benchmarks that Conroe should be tested with? What programs should be used to evaluate the 64 bit prowess of Conroe. What makes you think that Conroe will have bad 64 bit performance?
Also, the general consensus of the benchmarks show that Conroe is over 20% faster. Read some of the reviews will you? I'll link you to the exact pages.
Final thing. Isn't 4x4 supposed to be a FX exclusive thing?
"Before that, AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform."
And this is based on what data source? And how much will this solution cost comapred to a $500 Conroe chip which won't need top of the line DDR2 timings? Looks like with 4x4 AMD will really be holding to their whole price/performance value proposition.
And just how much 65mn volume will AMD be "cranking out" in 5 months (initial samples anyone?) How long before AMD converts 50% of their production to 65nm? (end of 2007? 2008?) Based on AMD's own roadmaps aren't all of the INITIAL 65nm products targeted for mid-stream speeds and same power envelop as 90nm (with possible exception of quadcore)
You keep complaining about benchmarks (and did the same on Woodcrest) - why not go out and buy one like you said you would and benchmark yourself - this would ensure no bias. :)
Could you include some actual sources/facts/links in some of these blogs?
""AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform."
This will only happen in your fucking wet dream, you sucker. ""
You sound like a 70 year old blind and deaf man who purchased Intel stock in 1984 ..dude can you raise your neck to stare at a computer monitor even. Take your post elsewhere in those million websites owned and controlled by Intel ..perhaps u can help jack up Intel shares by a cent ..but for how long ha ha !!
AMD is going to price 4x4 out of the market. 4x4 is only being created to win a few benchmarks with a few games (those that scale up well on multiple threads) and for super rich gamebois (i.e. "PC enthusiast market") to have the most elite hardware. 4x4 will not be available with entry-level dual core chips.
So we will have Conroe dominating 95% of the market and AMD struggling to hock their top-end 4x4 solution to trust fund gamers.
The sad thing is that AMD *used to be* that company that was bringing great performance at a reasonable price to the masses.
Today's AMD is just a drunken greedy company that has no reasonably priced performance chips. AMD's insane pricing of many of their processors is a profound betrayal of the very people who took a risk and bought AMD instead of Intel.
AMD has definitely let their momentary success go to their heads and unless they get back to what made them successful, they are in deep deep trouble.
There needs to be reasonably priced Opterons across the board and 4x4 must work with entry level dual core chips.
If not, even though I have an investment in Opterons and AMD Athlon x64, I am ditching AMD as they are becoming an exact clone of Intel. So why not buy from the big evil (at a lower price) than the small evil?
I am curios how Anad was able to reach 4ghz on air where as all other review site's pooped over 3.5ghz on phase? Only one review site was able to do 4ghz on LN2. The tuniq tower is not a great heatpipe and notice it is standing up, not the way it is in PC. And w/ tuniq he says was stable @ 4ghz, but w/ stock was only stable @ 3.4ghz. This does not make sense. heatpipe does not give you another 600mhz boost. Yes the temp improves during load, but will not give you stability. Maybe 100 MHz boost can be believable, but not 600.
I didn’t read all the reviews, but just wondering if any (or anad) says anything about temp during load, or did anyone do raid test?
"Before that, AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform."
4x4 will not be available for any entry level chips. 4x4 will not be available on reasonably priced motherboards.
The real name of 4x4 is $+$.
It is AMD's ploy to make you buy twice as many processors as you really need.
So you may be able to play some pervasively multithreaded games better than Conroe on your $+$, but mostly you are just buying the most expensive platform so you can be elite.
AMD is going to price $+$in such a way that a $+$ motherboard will be $100 more than a normal motherboard. And then only special top of the line processors will work with $+$.
Overall, a $+$ solution will run you $400+ for a special motherboard and $1000+ each for two special $+$ processors. Not to mention a bigger power supply, more fans to move the extra heat, etc.
For that kind of money, $2500 in motherboard and processors alone, you might as well just get x6800 and have a system that will run cooler, quieter, and cheaper vs. AMD's $+$.
So the smart and savvy buyer who used to buy AMD is going to avoid $+$ like the plague and go with a Conroe system instead.
What $+$ means is that AMD's architecture/micro-architecture team HAS FAILED. It means AMD has no ability to make a better processor.
It means AMD is going to focus on the cash-rich world of high-end gamers and cede the mass market to Intel.
It is the end of AMD. Instead of offering innovation to the market, we see AMD's greedy hand reach out for $+$.
I think planetx64 did a good review of Conroe/AMD64 without bias.. looks like AMD only needs a larger cache and the results would go the other way.
I am sorry to say this, but AMD is admitting defeat... AMD and this 4x4 concept is clearly saying that if you want to use AMD and beat Intel you need to use two chips.
They are clearly behind on 65nm, in relation to Intel, and are going to only get 5-10% performance increase when they come out. Still leaves Intel with about 5-10% in relation to AMD's current clock speeds, which Intel may decide to raise the speed again... 3.2Ghz is what I think I read and then what, you would need a 3.6Ghz AMD chip to compete with that which would make it a FX 70, and when do you think that will be out.
Intel is in the driving seat for a little while, AMD is in the back kicking and screaming.
While neither of these two companies are sitting idle, they both have projects comming, and next year will be very interesting.
So the 4x4 is just a sisterboard plug into a cpu socket? Then the part should be fairly inexpenseive.
Wonder if they can disable the second cpu for non-demanding apps just to save on watts, like having an on/off "Turbo" switch in computers of old.
The Conroe performance looks good kudos to Intel, I hope 4x4 is out soon I think it will be cost effective because AMD need to make this work on any of the x2 chips?
You want 64bit bechmarks? Well here ya go-
http://akiba.ascii24.com/akiba/column/latestparts/2006/07/14/663447-002.html?
translated here-
http://www.babelfish.altavista.com/babelfish/trurl_pagecontent?lp=ja_en&trurl=http%3a%2f%2fakiba.ascii24.com%2fakiba%2fcolumn%2flatestparts%2f2006%2f07%2f14%2f663447-002.html%3f
Oh, you want more? Then just head on over to techreport.com here-
http://www.techreport.com/reviews/2006q3/core2/index.x?pg=1
And to top it off a setup just like you'd set it up (no pun intended) yourself-
http://www.bit-tech.net/hardware/2006/07/14/intel_core_2_duo_processors/1.html
I'm awaiting your responses and this time read and analyse them throughly and try not to be rash.
Cheers :)
"/*
Q1. Final thing. Isn't 4x4 supposed to be a FX exclusive thing?
Q2. ...4x4 will not be available with entry-level dual core chips.
A. "Mounted on each socket will be an AMD Athlon 64 X2 processor, for a total of four cores." - David Schwarzbach, AMD.
Where do you see the word 'FX' ???
~~~
Q. And how much will this solution cost compared to a $500 Conroe chip which won't need top of the line DDR2 timings?
A. Two Athlon 64 X2 3800+ at 169$ plus a premium for the motherboard.
~~~
Q. How long before AMD converts 50% of their production to 65nm?
A1. AMD can convert their production now because AMD starts getting chips out of Chartered:
http://news.zdnet.com/2100-9584_22-6093601.html
A2. "In the fourth quarter, we will start to ship chips based on the more advanced 65-nanometer process from our own factories." - Thomas Sonderman, director of Automated Precision Manufacturing (APM) Technology at AMD.
A3. "We expect that 65nm wafer starts will account for more than 50% of total production at Fab 36 by early 2007, and we expect substantial conversion to 65nm by mid-2007." - Damon Muzny, AMD representative.
~~~
Q. The sad thing is that AMD *used to be* that company that was bringing great performance at a reasonable price to the masses.
A1. They have a FAB 36 to pay...
A2. http://www.techreport.com/onearticle.x/10153
~~~
That's all for now...
I'm beginning to think that Mr. Sharikou's "Ph. D" stands for "piled high and deep".
Looking for 64-bit C2D benchmarks? As noted by Jackall, check out Techreport's review. Running on XP Pro x64, they included 64-bit versions of SiSoft, POV-Ray, Cinebench, LAME, WME 9, picCOLOR, and UT 2005.
Their conclusion? "Both the Core 2 Extreme X6800 and the Core 2 Duo E6700 easily outperform the Athlon 64 FX-62 across a range of applications—and the E6600 is right in the hunt, as well. Not only that, but the Core 2 processors showed no real weaknesses in our performance tests."
Since I possess basic reading comprehension skills, presumably this makes me an Intel fanboy.
"AMD will sell two Athlon X2 AM2s frag the Conroe by 70% with the 4x4 platform."
This will only happen in your fucking wet dream, you sucker."
You are sick. Don't like read this this blog? just fuck out. or say something reasonable.
I like Sharikous blog often come here and read the news. A lot of alternative points thats is intresting for me as I;m chip dealer too. I guess for AMD would be much better To prize X2 chips compare with netbirst analogs. that would give much advantage and popularity for this chips. and somehow rais overclockers ability ( 65 nm will help I guess ) I thinq to use the same arqitecture wich is used in opterones will make new processors much more attractive. exscuse for my bad english.
4x4...
Will be doulbe the silicon area. Charge what double the price. But it won't be doulbe the performance.
So match Conroe and maybe beat it on one or two benchmarks.
It amounts to AMD giving away two cores for free and pricing 90% of the rest of their volume at 50% less. They go from slim profits to large losses.
PhD please tell me what I missed.
You really would get more credit if you weren't beating your meat so hard.
Why don't you read the reviews first jackass? Techreport/bit-tech don't do 64 bit reviews and the ones from the Jap site show Core 2 losing by 25%. I guess Sharikou was right. You dumbass Intel fanbois are all living in ancient history. Go clutch to your 32 bit mommy and cry moron. Then maybe you'll learn some manners and not lie so publicly and obviously.
You disgust me.
"I think planetx64 did a good review of Conroe/AMD64 without bias.. looks like AMD only needs a larger cache and the results would go the other way."
AMD is dropping the cache in the AM2 processors from 1MB/core to 512K/core. At the same time, they are dropping the number of HT links on the chip from 3 to 1. It truly is a crippled chip.
There will be no $+$ with low end chips because the HT links will not be there to make it work.
The only $+$ (4x4) implementation will be top-end FX chips that will keep the 3 HT links. AMD is making sure that the customer will not get good performance at a reasonable price. If the customer wants to have top $+$ performance, they will be paying enterprise server prices for processors to make it work.
Clearly AMD knows it cannot compete anymore vs. Intel on performance, even at the high-end.
It is a sad day when AMD requires two processors to perform as well as one Intel processor.
Smart buyers will not buy AMD until 2008 when K8L (or whatever) comes out. Until then, it looks like Intel is the good choice.
"Q. And how much will this solution cost compared to a $500 Conroe chip which won't need top of the line DDR2 timings?
A. Two Athlon 64 X2 3800+ at 169$ plus a premium for the motherboard."
WRONG.
AMD's current 512K cache AM2 chips have only 1 HT link.
They cannot be used with 4x4.
They are also $300 each
It will be $1000+ FX chips. Because otherwise AMD cannibalizes their highly lucrative SERVER market.
And there is no way AMD is letting their golden goose of overpriced chips, Opteron, get displaced by some glorified gaming machines.
Right here you can see the true AMD at work -- the highly cunning anti-consumer, anti-value attitude that is THE NEW AMD.
"AMD is going to price 4x4 out of the market. 4x4 is only being created to win a few benchmarks with a few games (those that scale up well on multiple threads) and for super rich gamebois (i.e. "PC enthusiast market") to have the most elite hardware. 4x4 will not be available with entry-level dual core chips.
So we will have Conroe dominating 95% of the market and AMD struggling to hock their top-end 4x4 solution to trust fund gamers."
Please, tell me how conroe will nominate 95% of the market if even Intel said it will only be hable to provide 20% of market coverage for the rest of the year and only 40% by the end of 2007?
dude, they WANT TO SELL OLD P4, thus making Conroes VERY HARD TO GET..
dream on with your 95% domination.
some people are just blind when they have the information right infront of them and they refuse to see it.
"I am curios how Anad was able to reach 4ghz on air where as all other review site's pooped over 3.5ghz on phase? Only one review site was able to do 4ghz on LN2. The tuniq tower is not a great heatpipe and notice it is standing up, not the way it is in PC. And w/ tuniq he says was stable @ 4ghz, but w/ stock was only stable @ 3.4ghz. This does not make sense. heatpipe does not give you another 600mhz boost. Yes the temp improves during load, but will not give you stability. Maybe 100 MHz boost can be believable, but not 600.
I didn’t read all the reviews, but just wondering if any (or anad) says anything about temp during load, or did anyone do raid test?"
because
1.- Anandtech always loves to exagerate the diferences and overinflate intel's products, plus since they're the first and most paid pumper.. obviusly they got the best revission ever ( Iwouldnt be surprised if that cpu they got would end being the ExtremeEdition X6900 or higher ( wich would appear by the end of the year if not later ))
A few guys over at xtremesystems.org could get their chips to or over 4GHz on air, so it's not like some major milestone.
I doubt whether it was stable at 4GHz though. Getting a screenshot @ 4GHz and being totally stable is a different story.
For what it's worth here are some UT2004 64bit numbers:
http://techreport.com/reviews/2006q3/core2/index.x?pg=6
Nothing out of the ordinary, Conroe scales just as well as the A64s and P4s, gaining a few % from 32bit to 64bit, nothing spectacular.
"/*
WRONG.
AMD's current 512K cache AM2 chips have only 1 HT link.
They cannot be used with 4x4.
It will be $1000+ FX chips.
bla bla bla...
"
AMD ATHLON 64 FX-62 PROCESSOR TECH SPECS:
Frequency / Cache Sizes: ...
L1 Cache Sizes: ...
Memory Controller: ...
HYPERTRANSPORT LINKS: 1 <----- ONE!
HyperTransport Spec: ...
Effective data bandwidth: ...
ETC: ...
Both AM2 X2 and FX have ONE HyperTransport Link.
I repeat: X2 and FX have ONE HT Link.
Do you copy ?
Next you will tell us that 4x4 will need two Opterons ???
Go here and look at number 5 and 7
I remember at one point some released the first 64bit game along with its 32bit version. This was to demostrate that games written in 64bit got better performance a lot better textures because of 64bit coding.
Just remember the success of a chip relies on 2 factors, perforamce and avaibility.
Just because your precious Conroe beats the FX in a bunch of 32bit does not mean it is a success unless 80% of you who have ordered it received it instead of substituting for another chip like an AMD64 or the FX or even one of those electric stoves you call a P4 because you couldn't wait that long for your chip to arrive.
Now considering you waited that long and AMD released a better performance chip that matches or surpasses your Conroe. And they happen to be able to produce in fairly good numbers. Would you still wait for your ordered Conroe to arrive or go with AMD?
Another Core 2 Duo bug evidence ???
Look at the C2D E6700...
http://techreport.com/reviews/2006q3/core2/bf2.gif
Maybe a new Intel "throttling" trick !?
To the joker who posted:
"A1. AMD can convert their production now because AMD starts getting chips out of Chartered:"
**These are currently only 90nm chips, it is also cheaper for them to manufacturer chips themselves then BUY them from someone else (unless Chartered is giving it to the without making any profit tehmselves**
A2. "In the fourth quarter, we will start to ship chips based on the more advanced 65-nanometer process from our own factories."
**Start...my question was 50% of total production - AMD fans keep saying well Core2 won't account for 40% of Intel's production until beg Q1'07 **
A3. "We expect that 65nm wafer starts will account for more than 50% of total production at Fab 36 by early 2007, and we expect substantial conversion to 65nm by mid-2007." - Damon Muzny, AMD representative.
**This represents probably <25% of AMD's total production as F36 will not be completely built out by then, and they still have a 200mm cranking out 90nm chips (at a higher WSPM for 200mm, before you start with the whole you get 2X more die/wafer out of 300mm - there will be fewer wafer starts in the 300mm fab)
Thanks for the non-answer
"Who on earth hasn't made the switch yet to a 64 bit OS?"
I haven't.
I would estimate the majorityof all business PC hasn't...
Most laptops haven't...(last I checked Poor-ion had a pretty small market share and all other major mobile chips are only 32bit).
I would also venture to say most non-enthusiast home users have not either...
Found something that might interest you.
Was checking PlantAMD64 and it seems they reviewed Conroe. What's even better they did it in both 32-bit and 64-bit Windows.
http://www.planetamd64.com/index.php?showtopic=24551
Yes you heard right Planetx64.com (that's the same people that run PlanetAMD64.com) has had the chance to test Intel's new flagship the Core 2 line.
They compare the Core 2 Duo E6700 and the Core 2 Extreme X6800 against the AM2 X2 5000+ and FX-62.
"The evaluation as it stands is still enough to show that Intel has finally pulled their collective heads out of the sand and gotten back into the performance game… and with a vengeance. They were not satisfied with meeting the AM2’s performance, they surpassed it significantly."
They used DDR2 800 with 4-4-4 timings and 2 Raptors in Raid 0 so there can be no arguing. No Conroe RAID problems were reported. Conroe is definitely conclusively number one whether in 32-bit or 64-bit. And this is from an AMD fansite supporting 64-bit technology too.
It's hard to read in the Japanese site, but even though 64-bit seems weaker on Core 2, the X6800 still generally leads the FX62 in 64-bit mode. I don't know if the 2nd set of charts is on Vista, but Core 2 seems to respond better in 64-bit there. The site seems to mention that some compiling and optimization for Core 2's microarchitecture from the software side is still needed so it may not be entirely the hardware's fault.
The Core 2 Duo E6700 outperforms the Athlon 64 FX-62 more often than not, yet the E6700-based system draws 74 fewer Watts under load. --quote from http://www.techreport.com/reviews/2006q3/core2/index.x?pg=16
wow.
allow me to use my SUPER mathematical skill to match the performance, say E6700 is 1.15 perf of the FX-62. then if the FX is OC (assume it is capable of it) to match it and the POWER increase linearly (which is not the case, the power up a bit exponentially over freq). then the E6700 consume 74*1.15 = FX-62 with 85Watt lower.
Then allow me to use my SUPER economially calculation (guess who always do this? :) ) .... 85 * 24 * 365.25 = 745kWh ... errr i do not know how much this cost in US ... ut in my country this is about USD50 ...
oh, someone is going to balme me on missing out the chipsets ... the intel high end MCH+ICH is about 25 watt in total ... guess how mnay watt an NVIDIA AMD chipset consume?? go google it ... ( i cannot find the number actually, but bas eon one site which state that the chipset for AMD run quite hot, even though it is just one)
mmm, let's create one conspiracy here, the energy bill is targeting AMD indeed ...
64 bit benchmark -- no difference
http://www.planetx64.com/index.php?option=com_content&task=view&id=283&Itemid=14&limit=1&limitstart=5
Many commenters here do not understand what 4x4 is and what it means.
It is wrong to compare AMD's 4x4 with Intel's large cache, or the latter's Extreme Edition, which requires lots of $$$ for a little improvement. It's also wrong to say AMD's 4x4 means its defeat because it is requiring two CPUs to achieve higher performance.
Two important things that make the differences are behind 4x4: 1) AMD's Direct Connect, 2) AMD's modular design. With these two, AMD is able to add additional processor, connecting through ccHT links, and achieve (80%) better multithreading performance. OTOH, Intel CANNOT do so. Core 2 will have difficulty to scale, unless some powerful North Bridge is put in place that adds $$$, system complexity, power consumption, and one more single point of failure.
So if you want a powerful workstation or gaming station, if you run multiple programs or multithreaded programs, if your applications need high memory/disk bandwidth, then 4x4 is the way to go. You won't get the same performance with two Conroe, just like you won't get that with 200x 80486. All x86 processors before K8 simple do not scale. Even Core 2 do not scale. (That's why IBM put so much efforts & money on its special x86 chipset.)
That said, Conroe does have superior single-threaded, 32-bit performance. About 15% faster clock-for-clock than K8. Plus, it seems to draw less power, too. I can't say it's not a good job. Yet, it's definitely not comparable to 4x4.
For what it's worth here are some UT2004 64bit numbers:
Since they can do 64-bit why do they only have ONE benchmark in 64-bit? It is kind of odd not to do a full 64-bit performance review when they apparently can. Perhaps they cannot show the results of other 64-bit benhcmarks?
"I can't say it's not a good job. Yet, it's definitely not comparable to 4x4."
If you could provide some links to the 4x4 performance that would be appreciated...I haven't come across any yet.
"Go here and look at number 5 and 7"
What I see is this: the dual-socket single-core Opteron box slightly outperforms the single-socket dual-core Xeon box.
I'd expect a single-socket dual-core Opteron box to outperform the Xeon box even more.
Maybe I'm a bit biased... what do you see, though?
I can’t believe what I’m reading. To some, Intel is a religion. Get a grip on reality. When 4x4 is main stream you’ll be able to afford it. Continue to use your Conroe system on 32 bit for awhile, until the time comes when you can step up. If your into games, Conroe is a fine cpu and cheap. You always get what you pay for.
"Both AM2 X2 and FX have ONE HyperTransport Link.
I repeat: X2 and FX have ONE HT Link.
Do you copy ?
Next you will tell us that 4x4 will need two Opterons ???"
In a way, YES, AMD will make you buy the gaming equivalent of an Opteron.
That "Opteron" will be a new FX chip that has an additional HT link for this so-called "4x4".
Except to see AMD's version of SLI marchitecture, but for dual CPU vs. dual video card. Maybe AMD will call it "FX2". Maybe something fancier.
And it will not be cheap! With the move to Socket 940 for consumer AMD processors, it is easy to put in an Opteron 2XX style consumer FX chip in that socket and double up on sockets.
AMD is looking at this "4x4" SLI-for-processors play as a way to sell twice as many processors to each high-end gamer. It is a gold mine for AMD.
It will be right around the same price as an Opteron 2XX solution. So for two fast consumer processors to win game benchmarks, you are talking $1000-$1200 per chip.
The pricing will be at this level to avoid cannibalism of real Opteron server processors.
Only the most loyal and fanatic AMD fanbois will buy that sort of system. You can get two very nice SLI video cards instead of that extra processor!
Smart buyers will go Conroe.
"That said, Conroe does have superior single-threaded, 32-bit performance. About 15% faster clock-for-clock than K8. Plus, it seems to draw less power, too. I can't say it's not a good job."
Intel hit a home run for 95% of the market. Including games -- the vast majority of which still need maximal single-threaded performance to really scream.
This hokey "4x4" dual-socket stuff that AMD is hocking may be a status purchase for rich gamers running a few fancy games, but there is no real need in the consumer market for it.
There are very few high-end consumers who go out and buy a dual Opteron system. And there will be very few high-end consumers who buy the consumer version of dual Opterons.
AMD's "4x4" is just a diversion from the fact that AMD's architecture team has failed to improve its processor in the past FIVE YEARS.
As much as Intel gets slammed here, K8 has been sitting still for a long time. And there is nothing on the horizon that is better.
So the smart buyer will go Conroe until 2008 when AMD has the K8L for consumer products out.
Sharikou - C'mon please start saying something neagtive about conroe. You cant be quiet for so long
I don't really know why it's dificult to get everybody on the same page. Conroe and Woodcrest really rocks for the platform they target. AMD has longer term advantages due to ccHT. 2008 and 2009 may prove the toughest years Intel has ever had to face. But it should rebounce in 2010 (or late 2009) when it ships CSI based systems.
And for all those peeps who come here to make a decision to buy a desktop/2P server - GET YOUR HANDS ON CONROE/WOODCREST!
You don't have to care what the share price of Intel is to make a choice. You have to only care about performance per dollar!
"I can’t believe what I’m reading. To some, Intel is a religion. Get a grip on reality. When 4x4 is main stream you’ll be able to afford it."
Was it not AMD who said this product is targeted at the ENTHUSIAST market? To me the enthusiast markt does not mean mainstream (this means the FX-62, Conroe x6800 audience who have cash to burn on $1000+ CPU's)
Found the quote:
"The 4x4 initiative will be directly targeted at enthusiasts, according to Moorhead" (Extremetech article)
a little bit about benefits of having large 4 MB cache. Im not sure if i like this core or hate it....
MD's "4x4" is just a diversion from the fact that AMD's architecture team has failed to improve its processor in the past FIVE YEARS.
...err and how is intel diferent from that? i mean do you realy need pentium 3 with 4 MB of cache? (and yes it is A pentium 3 with 4 MB of L2)
To some, Intel is a religion
is it? they must have been worship at the wrong place. Here is the Temple Of AMD
Thou shall listen to my messenger Sharik-ou, for whatever he said, is true. For I love him so much, I have given him the power of prediction.
This is our holy site. Others sites are evil, with the exception that INQ sometime being evil, somtime being holy
When INQ said good thing about me, it is holy, else it is otherwise. Occassionally other evil sites are holy on the particular line that protraying Me beating the evil.
OMG!!!
Even PlanetAMD64 guys have turned into paid pumpers!!!
What is the world coming to?
Your prediction of Intel colapse could come true because of Conroe, I mean that all these Intel fanbois should spread the word that theo nly Intel chip worth a damm is the "new ones "don't, never buy any of those crap P4's"
(7 billion down the toilet)
Try give credit where it is due. Learn from your frens out there.
http://www.planetamd64.com/index.php?showtopic=24551
planetamd64 and planet64 are run by the same group of people.
Trust me, it would be real funny if you label this site as intel pumper as well. Try do that and see what is other's response. :)
"Conroe and Woodcrest really rocks for the platform they target."
The target is single-socket desktops and servers, then yes.
"AMD has longer term advantages due to ccHT. 2008 and 2009 may prove the toughest years Intel has ever had to face."
You might want to change the 'longer term' to 'architecture'. Also, I'm not sure K8L in later 2007 would not give Intel some pressure.
"And for all those peeps who come here to make a decision to buy a desktop/2P server - GET YOUR HANDS ON CONROE/WOODCREST!"
Conroe and Woodcrest are certainly very attractive for single-threaded or single-process workloads, although the former is a better desktop chip than the latter as a server chip.
However, real-world multi-thread and multi-process workloads are very complicated, so much that there's not yet a good benchmark for them. The large shared L2 SmartCache helps simple benchmarks but could hurt real-world performance in common but complex scenarios. Other tricks such as code fusion and speculative loads also could behave alike. They could well be like the HyperThreading that to be found harmful for heavy workloads.
One thing is for sure: Conroe/Woodcrest were designed to perform on benchmarks. I do not say that they are not good choices for desktops & single-socket servers. Only that they might not 'rock' as much as those simple benchmarks suggest.
Given AMD's coming price cut to match Conroe equivalents, I'll stick with Athlon64 X2, which is 1) right-away available, 2) proved to perform both 32-bit & 64-bit real-world apps well.
4x4, if not just a quick remedy to narrow the performance gap before K8L, would be a nice early experiment for a coprocessor setup for desktops.
The coprocessor needs not to be a full-blown cpu, just elements of it to boost whatever is needed.
Then when K8Ls tape out, 4x4 will be even more
powerful.
So if you want a powerful workstation or gaming station, if you run multiple programs or multithreaded programs, if your applications need high memory/disk bandwidth, then 4x4 is the way to go.
are you aware that this will kill the opteron workstation market?
http://www.theinquirer.net/default.aspx?article=33048
Conore problems or am I AMDfaing my perception?
After reading the post and all the comments to date, I would like to thank Sharikou for his blog.
I wonder if there is similar Intel fanboy site somewhere out ?
It would be nice to open a fight club.
[Intel] won't be able to supply more than 27% of Core 2 Duos to the market by the end of the year.
...
However, there is still one thing that disturbs me about the Intel Conroes. The load times on our Far Cry tests simply took ages, regardless of whether we used a NetWurst craptecture or Core marchitecture. AMD loads FarCry in a matter of seconds, and in this particular game, we feel like watching Athlon 64 to complete SuperPI after running Conroe for three years. B-o-o-o-ring. A big surprise for me was the fact that Quake 4 gameplay also experienced more hiccups on Conroe platform than it did on a competing AMD platform.
...
Looking into the crystal ball
If the claim that "Intel makes great CPUs" is true, the claim "Intel sucks at motherboards" is true as well. Sadly, the testing was conducted with third revision of the very same motherboard. If you're investing in Conroe right now, bear in mind that your machine just might not support the future quad core Kentsfield CPUs, which are bound to use a 1.33GHz front side bus. To us, the safest bet for a Conroe based computer would be to hold out for an nForce4-SLI-16X or nForce 590 SLI based motherboard, but then there's ATI's plans, of course. We would also avoid betting your shirt on any of the 965 chipsets, since the lack of IDE controller and current zerg-rush bolting of UATA-133 support does not add to the corporate stable image Intel is trying to convey to its customers. Also, there is the small matter of Windows Vista compliance. But you do not want to buy these CPUs with integrated graphics.
...
Intel has regained the performance crown. It only took five years, but the company did it. If you are planning to buy a new system, however - we are putting up a warning flag. Today, the processor is just one of the parts that need to work in harmony in order to achieve good results. For ATI lovers, RD600 is getting more and more fans each day, and our friend and die hard enthusiast Rahul from VoodooPC stated numerous times that we should watch out for ATI. The fact that DirectX 10 graphics is coming in late September or early October certainly should also have considerable impact on your spending decision.
theinquirer.net
Core V2 : x86 vs x64
****Your prediction of Intel colapse could come true because of Conroe, I mean that all these Intel fanbois should spread the word that theo nly Intel chip worth a damm is the "new ones "don't, never buy any of those crap P4's"
(7 billion down the toilet)****
Thats all great but who will be buying any am2 or 939 when the conroe is available for much cheaper. Only fan boys will be making that decision and Intel has deeper posckets than Amd go look at their loss in sales, low revenue prediction for the next quarter and slashing prices to compete before you start looking elsewhere FANBOI!!!!
"Your prediction of Intel colapse could come true because of Conroe, I mean that all these Intel fanbois should spread the word that theo nly Intel chip worth a damm is the "new ones "don't, never buy any of those crap P4's"
(7 billion down the toilet)"
No, Intel will just donate most of the inventory to various 3rd world projects. They will have plenty of processors to put in the "$100 PC". Intel will be in write-off heaven for many years to come.
I'm very much looking forward to when Conroe ships. I am really tired of AMD's high prices and low performance/watt.
And just when will cornrow be stable enough to attract gamers?
Not soon enough.
Well I just want to be up front before I write comments, I am partial to AMD. I want to make this clear because I don't like to act like I'm completely neutral(which for well over 90% of the opinions out there, they are written with prejudice, even though they act like they are neutral). In spite of my lack of neutrality, I'm still capable of forming a reasonable opinion.
In the short term, I think it is going to be difficult to argue that Conroe and company have the performance advantage. Unless there is some sort of "unforseen" hiccup of course.
However, what annoys me is the forced hype with this launch. When AMD initially launched K8, there was a big focus on what AMD was bringing to the table as far as technology. Pre-release benchmarks weren't the name of the game. The focus with this launch has been, "see look", we are way better then what we had and we smoke the competition. The technology behind the processor has taken the back seat in all of this.
You might ask ok, so what? Well, when all these benchmarks were "officially" released, it was rather anti-climatic. If Intel had waited till recently, I think I would have been shocked. I would have been "wow, where did this come from?" Instead, with this "official" launch, we are only confirming what we basically knew 4-6 months ago(except arguably the percentages were a little higher a few months ago). This whole hype campaign feels like someone marketing to me and saying, "hey, don't buy a new processor because we are going to have a really fast processor soon, and you don't want to get stuck with a slower processor, do you?"
People have these rose colored glasses when they remember the AMD K8 launch. The reality is, K8 when first launch was not this great overwhelming event in favor of AMD. Do people remember "synthetic" benchmarks running faster on the older P4? It took time for people to accept the idea that games ran faster on K8. Sites like tomshardware were questioning whether K8 was much of a processor at launch.
Now, we have this new launch, and without a doubt this processor is the fastest and best thing since slice bread. Benchmarks aside, no one has highlighted any weakness to this processor. Unlike the K8 launch, there is no debate to all the review sites out there. I thought there was always debate? Wait a sec, that is unreasonable because I'm a fanboy who can't see the truth for what it is. Right.
"/*
Q. That "Opteron" will be a new FX chip that has an additional HT link for this so-called "4x4".
A. You tell me that with 2 FABs, AMD will produce distinct CPUs for socket 939, 940, AM2, F, and 4x4, some with 2 cores and some with 4 cores, all at the same time !!!
~~~
Q. The pricing will be at this level to avoid cannibalism of real Opteron server processors.
A. AM2 and FX use standard memory while Opteron use ECC memory. Servers need ECC for maximum data integrity, so the Opteron will not be cannibalized.
The reality is, K8 when first launch was not this great overwhelming event in favor of AMD.
Doing business base on Technology alone doesn't get you elsewhere. although you might wanna make fun of the marketing team, but they are the one that one that help you sell technology. On your comment, it would be either AMD marketing team had failed to do their job nicely, or juts that AMD did not has good reputation at that time (just look at K5 and K6). Intel has a very strong Marketing team and the reputation for its technology.
Benchmarks aside, no one has highlighted any weakness to this processor. Unlike the K8 launch, there is no debate to all the review sites out there.
again, refer to my answer above. Besides that, the Core 2 duo is indeed a very good CPU design. Well, some might wanna argu on the IMC and the FSB stuff. but indeed this are really some sort of engineering trade off. No matter how one argu on this, it is clear that the current Core 2 Duo will beat AMD in 1P and 2P, even it has no IMC and still use its FSB. AMD will have its niche at 4P and above for sometime. Nevertheless, not all server job can use x86. Some required extreme RAS features, and that where the Itanium (and may be other non x86 CPU) comes in.
+++TO EDWARD AND C2++++++
Stop your crying for the last three years AMD has been on top of the benchmarks and no one here complained. All of a sudden Intel came and whooped AMD at its own game there is nothing that the X2 and the FX chips are superior to conroe in and suddenly conroe is a "benchmark fitted" product.
Go look at any tech website (except the joke of the tech community hardocp) and tell me if they share your opinion obviously you being the expert in such things and all....FANBOI!!!!
The important thing is that Intel is now the OPEN company. Intel shares processors, specs, performance months before ship.
Intel does this because they are damn sure the chip rocks. Even AMD has admitted this.
On the other hand, no one knows what is going on at AMD anymore.
And that is a sign of a company that has no technology, no innovation left in its gas tank. K8 was it.
AMD is now out of gas and has to think of dumb stuff like "two crappy chips > one good chip".
When AMD is a real company they will start telling their customers what they are doing. And they will have the strength of character to say "Intel did a really good job on that processor."
"I'm very much looking forward to when Conroe ships. I am really tired of AMD's high prices and low performance/watt."
Conroe is still NOT YET OUT..
inshort words.. AMD might have high prices, but atm it still leads the performance/watt
or you still want to believe these monster p4 are efficient?
The important thing is that Intel is now the OPEN company. Intel shares processors, specs, performance months before ship.
Like Intel is selling unannounced chips to Tier 1 OEMs at hugh discounts weeks after Intel told its channel partners that no special
pricing for big OEMs to level the playing field?
See Digitimes story.
On the other hand, no one knows what is going on at AMD anymore.
It's because AMD is in the middle of product transitions. Intel had its.
To be fair,AMD has presented some future plans.
AMD is now out of gas and has to think of dumb stuff like "two crappy chips > one good chip".
Like you have real insight of what's gonna happened. I think not.
Coprocessor is more flexible in terms of
upgrading, adding functions and performance.
When AMD is a real company they will start telling their customers what they are doing. And they will have the strength of character to say "Intel did a really good job on that processor."
Reminds of Intel spewing FUDs about Athlon and Opteron. Will people who bought Netburst P4 and Xeon feel a bit duped now?
AMD bashing aside, AMD has achieved some of its
goals:
Getting product lines in previous Intel loyalists.
Get more recognitions with businesses.
Strengthened channels and partnerships.
Yes, there will be "swing votes" going to Conroe.
And if some think AMD won't have anything competitive in the future, let'em crow after 2+ years of lagging behind.
"All of a sudden Intel came and whooped AMD at its own game there is nothing that the X2 and the FX chips are superior to conroe in and suddenly conroe is a "benchmark fitted" product."
You really got the facts reversed. Please note that not only was Conroe designed for benchmarking, so was NetBurst. Intel isn't defeating AMD at AMD's own game with Conroe, it was AMD who defeated Intel at Intel's arena for the past 3 years.
If you didn't know, Intel tried so hard to optimize NetBurst for all types of benchmarks, went as far as to use CPUID for the trick, yet wasn't able to beat K8 on those benchmarks they optimize for. Now finally Conroe with its P3-based core could do that fairly, and I do not contest it at all. Just congrats to Intel that it finally can beat AMD by 10-15% on those Intel-favored benchmarks! Good job indeed!
(Okay, I was a bit biased toward AMD. It seems on some games Conroe would beat AMD by 30%+. But I stand by my prediction that Conroe's performance would be less in real-world multi-threaded apps, just like NetBurst's HyperThreading was.)
OTOH, AMD's K8 from the beginning was not designed with as much benchmarking in mind as its Intel counterparts, simply because AMD didn't have (any) sway on software developers 3 years ago. If you look at it, K8's design was very practical, aimed at scalability & modularity, average performance, and (frankly) the heavy server workload.
Like it or not, the sole purpose of Core 2 is to perform well on benchmarks. The current Core 2 chips were meant to live short, period. Even Intel admits that. Intel is a better marketing company than a design house or engineering firm.
"And that is a sign of a company that has no technology, no innovation left in its gas tank. K8 was it."
That seems to apply more soundly to Intel for the past 3 years. ;-)
In any rate, I'd withold such comments as yours until rev.G is released later this year.
"AMD is now out of gas and has to think of dumb stuff like "two crappy chips > one good chip"."
No, two crappy chips doesn't outperform one good chip. Two Intel P4s only perform even more poorly per watt. OTOH, it's not rocket science that K8 has excellent socket scalibility (except to AnandTech, of course).
"When AMD is a real company they will start telling their customers what they are doing. And they will have the strength of character to say "Intel did a really good job on that processor.""
AMD's management did say Core 2 is a good job. They also express confidence in K8's open platform approach & direct connect. I guess some people weren't bright enough to read properly.
Also, everyone knows what AMD's going to do: rev. G, socket F, K8L, you name it. People only don't know HOW AMD's could make them better - and I believe there's good reasons for AMD to keep the lid on.
Of course, we cannot say AMD's future products will rock before seeing them at work, but we also cannot dismiss them just because we know little of them. We simply have to wait until they are released.
Here are a couple of comments from previous post...
1. Wirmish said...
Another Core 2 Duo bug evidence ???
Look at the C2D E6700...link
Maybe a new Intel "throttling" trick !?
Answer...
Did you read the article? BF2 caps the frame rate at 100.
2. Edward said... “I can't say it's not a good job. Yet, it's definitely not comparable to 4x4.”
Answer...
What would happen if they redisgned Bensley to run two Conroes? Still not comparable?
3. Anonymous said... "When 4x4 is main stream you’ll be able to afford it."
Answer...
Well when will that be? Prices for the initial launch? We need facts not opinions. FX or AM2?
My comments...
Every thing being said by AMD fanboys is an opinion of a hurt soul who's company has now lost the performance crown, this is new ground for them, as they have not been here for a few years.
Sucks doesn't it.
As for Intel hyping it up, they definetly should, everyone who has been playing PC video games has understood AMD in gaming was where you wanted to be. Long time Intel users have a chance to rejoyce, they are back on par or slightly beating AMD.
Comments for the Dr....
Dear Dr. it would be nice if you could point us to some links in regards of your claims that K8L will be out in a few months.
Thanks.
"What would happen if they redisgned Bensley to run two Conroes? Still not comparable [to 4x4]?"
First, lets wait till they redesign Bensley to say that, will we? My guess is it'll be easier & more efficient for Intel to go quad-core than to glue two Conroes with a chipset. For one, the cache coherency problem will crap the chipset design all over the place.
Still think 4x4 & K8's direct connect are "opinion of a hurt soul"? Come on, I wasn't even denying the superiority of Core 2 as single processors! How can you Intel's truly fanboy be so blind to see the other facts, while still calling others fanbois?
To be fair, 4x4 does have some problems, but those are not chip performance per se. First, it'll be very pricy unless AMD executes 50%+ pricecut. Second, its NUMA nature will need some software optimization. Third, it'll need at least 2 HT links for all processors that support it.
Edward everytime I comment on your post all of a sudden its not a discussion it turns into this...
"How can you Intel's truly fanboy be so blind to see the other facts, while still calling others fanbois?"
Have I ever said that P4 was superior to AMD? NO.
I do use Intel, and acknowledge that AMD was superior in gaming and would have to say better at multitasking than P4. Fanboy is not a term I would normally attribute to myself, except here if your not pro AMD all of the time all of a sudden you an Intel fanboy, but to all the people who can not look at the benchmarks and see Intel winning, those are the fanboys in which I reference.
Did you read the whole comment... "Every thing being said by AMD fanboys is an opinion of a hurt soul who's company has now lost the performance crown, this is new ground for them, as they have not been here for a few years."
Did C2D beat AMD FX62? Yes.
Hence lost the perormance crown, it was not directed at your comment, nor was it meant to be mean, it was just a comment in which it should there denial of the benchmarks.
Instead of jumping to conclusions, and reading things out of context, lets try and be civil.
I hate the fact that every post I make after yours ends up like this, me having to defend it.
To be fair, 4x4 does have some problems, but those are not chip performance per se. First, it'll be very pricy unless AMD executes 50%+ pricecut. Second, its NUMA nature will need some software optimization. Third, it'll need at least 2 HT links for all processors that support it.
Frankly speaking, it has quite some baerier to overcome at the software side, i.e. the NUMA nature. It is not just the 'some' that you mention.
And secondly, the cost. While people has been talking about the cost at the hardware end, what they forgot to see is the software end. You would have to pay about double for all the software that you own. Windows charge per CPU, not per PC. If you have the MS Office, and that's it. It will be too costly unless you go for piracy. Even if AMD is indeed working with multiple Software vendor on renewing the license terms, the changes won;t be fats and the first adopter has to pay the cost. Besides, i think most vendor will not change the license term to per PC(i know there are some that is already per PC, but not MS, etc). Else the hardware provider start to put more CPU into a single motherboard and the software company would lose money.
The software cost has been in my mind for quite sometime, but i have been reluctant to say it out. Why? because it is against my interest. :). Why i say it out now? Because i now thik that the AMD is most likely smarter than me and they have a solution. And perhaps, YOU know the answer. Can you please share with me on this?
"Conroe is still NOT YET OUT..
inshort words.. AMD might have high prices, but atm it still leads the performance/watt
or you still want to believe these monster p4 are efficient?"
I love these types of comments, let's look at the five stages of AMD fan denial:
1) Intel release benchmarks at IDF - "Yeah , well Intel built the system and probably just handicapped the AMD system; that performance isn't real" (if you recall Intel even overclocked it 200MHz)
2) Engineering samples start floating around the web with good benchmarks - "well those are just engineering samples; Intel probably just cherry picked them"
3) "Well Intel will probably make these ridiculously expensive at launch" - XE edition is cheaper then FX62; midperformance parts which are comparable/better to AMD tope end are priced $300-500.
4) Real benchmarks come out all over the place on July 14th. "Well if you look at certain games which are dominated by GPU, Conroe is only comparable to the top end AMD chip" (nevermind all of those other benchmarks like video/audio encoding)
5) "Well it isn't out yet; AMD still has price performance crown". Well I guess I'll have to wait a few days on this last point. Of course there's still the possibility that this could all just be a grand marketing conspircay where Intel isn't planning on actually shipping any of these chips...of course when it does ship AMD will either need to 2-3X the performance or cut proces on FX62 by 50-67% to be competitive on price-performance (to E6600)
"Frankly speaking, it has quite some baerier to overcome at the software side, i.e. the NUMA nature. It is not just the 'some' that you mention."
Now you're criticising 4x4 for its "NUMA" software nature. Please note that 1) multi-threading & multi-processing with NUMA is already supported by Windows system calls today, 2) I bet 2 years later, maybe even closer, Intel will enter the NUMA game, too. THEN you'll have a completely different attitude toward it, won't you?
Intel fanboys have been wrong about microprocessor evolution for the past 3 years. Now they will miss to recognize the new trend again specifically due to Intel's benchmarketing strategy. Note that this is irrelevant to how well Conroe performs in benchmarks. This is only about what high-performance computers will look like 2 years later. I assure you they'll be much more like 4x4 than C2D.
p.s. I'm not saying C2D is not a good single-socket performer. Just that it's designed for the benchmarks, and to have a short life. Even Intel admits the latter, but Intel fanboys do not (wish to) listen. As a marketing product C2D is wonderful, because it beats everything in most benchmarks; as an architecture... I can only say that it'll be gone in the wind before 2008. Again, Intel is more a marketing company than an engineering one.
enumae said: "Fanboy is not a term I would normally attribute to myself, except here if your not pro AMD all of the time all of a sudden you an Intel fanboy,..."
What a spectagular response of yours! 'Fanboy' is not a term you would attribute to yourself, yet you seemed to use it against other people freely and casually. Now you who called others fanboys first without good reasons starts to ask me to be more 'civil'? I'd kindly direct that request to yourself.
As for why I called you a fanboy, here's why: 1) Have you grasped my points that a NorthBridge glue won't be a good match for 4x4's direct connect? You seemed to ignore that, didn't you? 2) Have you recognized that ccHT do have more architectural superiority than AMD fanboism? How about Alpha-EV8 fanboism? How about Intel-CSI fanboism? Yet you dismiss 4x4's direct connect using "Bensley redesign" without much thought.
I thus conclude this guy much be a C2D fanboy who would not accept anything other than C2D rocks. Nice.
Now you're criticising 4x4 for its "NUMA" software nature. Please note that 1) multi-threading & multi-processing with NUMA is already supported by Windows system calls today, 2) I bet 2 years later, maybe even closer, Intel will enter the NUMA game, too. THEN you'll have a completely different attitude toward it, won't you?
Come on. i didn't critisize on the NUMA and i think it is good for MP. i just re-enforce your statement that it takes time for the software to be efficient on it. And here is my pro-intel view: The software tuned for NUMA will happen rapidly when Intel join the game as it has the ecosystem influnce that AMD doesn't has. But, it is going to happen at the server end, long before it happens in the desktop end which the 4x4 target on.
You can blame all you want that the Conroe stuff are for benchmarking. Bear in mind that there are quite some 'benchmarks' are actually real applications : games, SQL transactions, forum, etc.
Btw, i'm very much interested to know you opinions on the software cost end of the 4x4, as what i explain in the previous post.
"as for why I called you a fanboy, here's why"...
Edward you truly are an idiot, and an illiterate a$$.
"1) Have you grasped my points that a NorthBridge glue won't be a good match for 4x4's direct connect?"
Did I ever say anything about that, no, I was actually asking you if it would work, try reading the post.
"2) Have you recognized that ccHT do have more architectural superiority than AMD fanboism? "
Again, I have never knocked AMD's "architectural superiority".
And another one...
"Yet you dismiss 4x4's direct connect using "Bensley redesign" without much thought."
Again illiterate a$$, look at the response to your first question.
"I thus conclude this guy much be a C2D fanboy who would not accept anything other than C2D rocks. Nice."
You really are an amazing breed, to bad you never finished that short story book with all the pictures, it might have helped you later in life.
Move along until you can comprehend what is written in front of you, you are a waste of peoples time.
"Now you're criticising 4x4 for its "NUMA" software nature. Please note that 1) multi-threading & multi-processing with NUMA is already supported by Windows system calls today, 2) I bet 2 years later, maybe even closer, Intel will enter the NUMA game, too. THEN you'll have a completely different attitude toward it, won't you?"
Actually, NUMA is not implemented well in any modern consumer OS. Microsoft does a decent job in Windows Server 2003 (and also Windows XP x64 which is based on Windows Server 2003).
Windows XP (not x64) has a poor NUMA implementation.
Maybe for Vista AMD's 4x4 makes some sense.
But the vast bulk of people over the next couple years who are not using Vista will see little or no benefit.
When it comes to marketing, AMD looks like they are the ones playing the con games these days.
AMD said they would drop prices to compete with Core 2 Duo on performance/$ basis. So far this has been a lie. No price drops.
When are price drops coming on all the walking dead Socket 940 Opterons that AMD is not supporting in the future? Or is AMD going to milk the Opteron 940 market into the grave?
What is 4x4 and when is it coming? So far it is just a VAPOR announcement by AMD.
All in all, AMD really does look like they are falling apart. All it took is one hard punch from Intel. Today's AMD is all about high prices, vapor products, and lies.
"Today's AMD is all about high prices, vapor products, and lies."
No, no and no. They are going to drop prices to compete with Conroe. 4X4 is as far from vaporware as you can get and since when does AMD make a habit of lying?
Did you make a typo and write "AMD" instead of "Intel" in this post?
Intel fanboys should restrict their comments on the single-socket superiority of Core 2 Duo. Because almost everything else they were implying were seriously flawed & brain-washed.
"Actually, NUMA is not implemented well in any modern consumer OS. Microsoft does a decent job in Windows Server 2003 (and also Windows XP x64 which is based on Windows Server 2003)."
I got this from your logic: AMD's NUMA architecture is ahead of the evolution curve. The majority consumer OS is actually waiting for AMD to push for and bring on the next evolutionary processor/memory architecture.
"But the vast bulk of people over the next couple years who are not using Vista will see little or no benefit."
Well, if you don't drive a hybrid car, you certainly don't get the better mpg. Is it Toyota's fault? Does that mean those hybrids are bad?
"AMD said they would drop prices to compete with Core 2 Duo on performance/$ basis. So far this has been a lie. No price drops."
Because Conroe is not out yet. Gosh... I've been waiting for ages for that simple release - and its accompanying AMD pricecut!
It seems only Intel's lying more than anyone else.
"When are price drops coming on all the walking dead Socket 940 Opterons that AMD is not supporting in the future?"
This argument is totall flawed, because there are plenty of other choices from AMD if you want to buy a high-performance server chip but don't like socket 940.
OTOH, does Itanium-based servers give you any choice of that? Does Woodcrest allow you to scale past 4 sockets?
"What is 4x4 and when is it coming? So far it is just a VAPOR announcement by AMD."
So are you now critisizing 4x4 because it's not here and late? I though people from your camp (maybe yourself) just said 4x4 is overrated and not needed?
"Today's AMD is all about high prices, vapor products, and lies."
I don't give a damn to your claim of high prices or vapor products, because Intel's desktop C2D wasn't out, and its server C2D isn't scalable compared to Opterons.
However, I see nowhere that AMD lies, at least recently. Intel OTOH lied about C2D's 60% faster and 80% less powerthan 'competition' (complete bullshit proved by all Intel-pumped websites, which show C2D advantage against 'competition' is at best 30%, and in average 15%).
"Edward you truly are an idiot, and an illiterate a$$."
What a great response from someone who demanded others to be 'civil' when being called merely a 'fanboy'! You are truly the example of yourself.
And all you could say later is something like "I didn't deny what you said was true... I was just asking if its opposite was also true". Tell me how this could come from outside fanboism?
"Btw, i'm very much interested to know you opinions on the software cost end of the 4x4, as what i explain in the previous post."
For OSes, the scheduler needs to recognize memory affinity to different nodes; this is relatively easy (and unaffecting end users).
For programmers, NUMA brings concern on memory allocation when multiple threads/processes are used. That could touch many lines of code on existing multithreaded programs, but for new ones, it is minor compared to multithreading itself.
If what you meant to say was that NUMA-optimized program is few today or within a year, then it is agreed. But I expect that to change before or around 2008. I may be wrong, though, if Intel/AMD could pull other tricks (performance or marketing) out of this multi-core UMA architecture.
Edward what are we going to do with you... I was trying to be civil, but you poke and prod, and make no sense, in turn it feels as though your trying to start crap.
This comment... "What a great response from someone who demanded others to be 'civil' when being called merely a 'fanboy'! You are truly the example of yourself."... WTF?
Look here at my comment in a post directed to you... "Instead of jumping to conclusions, and reading things out of context, lets try and be civil."
Somehow me asking you to be civil, turns into demanded everyone to be civil, and it was not in protest of being called a "fanboy". If you would like to lable me one go ahead, I could careless what you think, you are a complete dumbass.
In regards to your last comments..."And all you could say later is something like "I didn't deny what you said was true... I was just asking if its opposite was also true". Tell me how this could come from outside fanboism?"... WTF are you talking about? I would be much more definitive in regards to a statement than that, it would seem that you are refering someone elses posting.
Trying to be civil with you was obviously a waste of time, you are incapable of thinking with your brain, go and get an etch-o-sketch and keep yourself busy for a month.
"Today's AMD is all about high prices, vapor products, and lies."
Response: "No, no and no. They are going to drop prices to compete with Conroe."
So basically they will milk their consumer base for as long as possible before they HAVE to cut prices? Sounds like the customer centric company everyone's been dscribing....why not drop them now?
(and if you're response is "well because there isn't a competitive product" than ask yourself what will happen if you all get your wish and Intel goes bankrupt...)
"Btw, i'm very much interested to know you opinions on the software cost end of the 4x4, as what i explain in the previous post."
For OSes, the scheduler needs to recognize memory affinity to different nodes; this is relatively easy (and unaffecting end users).
Nope, i'm referring to $$$ here. e.g. you would need 1 windows license per socket. refer to my previous post on this doe detail.
enumae said: "I was trying to be civil, but you poke and prod, and make no sense, ..."
enumae said: "This comment... (skip) ...WTF?"
enumae said: "If you would like to lable me one go ahead, I could careless what you think, you are a complete dumbass."
enumae said: "In regards to your last comments... (skip) WTF are you talking about?
Then (drumrolls) enumae said: "Trying to be civil with you was obviously a waste of time, you are incapable of thinking with your brain, ..."
To be frank, I didn't see you try to behave 'civil' at all! I also didn't see you regret calling others fanbois freely and casually in the first place. I also didn't see you admit that your "Bensley resign" proposal is not comparable to 4x4. I also didn't see you admit that many pro-AMD comments are NOT from hurt fanboy souls as you previously described. All I saw is you asking for civility, but keeping none for yourself.
Maybe I am incapable of thinking with my brain, and I don't need a brain for microprocessor architecture discussions. Tell me how my technical analyses are wrong, and I am all ears. OTOH, I couldn't care less how much or little a fanboy you are, or you think I am.
If you were poked because I called you a fanboy, I apologize, but would like you kindly be reminded that 1) people here on both sides are (mis)called fanboys all the time; literate as you should've known that before joining the club; 2) it might be wiser for you to refute this fanboy mark by recognizing - not just avoiding - direct connect's architectural superiority, rather than spilling out vulgar languages.
I do not wish to continue this conversation futher. Say what you must, and expect no more reply from me. Hope you understand. Thanks.
Edward... "recognizing - not just avoiding - direct connect's architectural superiority, rather than spilling out vulgar languages."
I will definetly recognize that AMD's architecture is superior, I do not recall ever saying differently, if I made the impression, my bad.
"Tell me how my technical analyses are wrong, and I am all ears."
I also do not remember saying you were wrong, honestly I was actually trying to agree with the technical side of the initial conversation.
"I also didn't see you admit that your "Bensley resign" proposal is not comparable to 4x4."
I have no idea if it would be, as I pointed out I was asking you, I read alot about this technology, but to be honest some of it escapes me.
Sorry for jumping around, but I wanted to touch on your points, I have no I'll will towards you. I think we always get off on the wrong foot when posting.
"So basically they will milk their consumer base for as long as possible before they HAVE to cut prices? Sounds like the customer centric company everyone's been dscribing....why not drop them now?
(and if you're response is "well because there isn't a competitive product" than ask yourself what will happen if you all get your wish and Intel goes bankrupt...)"
So what you're trying to tell me is that prices for everything in the world should always be dirt cheap since something better will be out shortly? Should a 60" HDTV be only a few hundred dollars since there will be bigger TVs with higher resolutions coming out in the future? Should cars be sold for 50% of their current value simply because they or a competitor are releasing a better vehicle a month later? Please tell me how your argument makes any sense at all. AMD IS the top dog right now and they deserve the price premium until there's anything out there that can compete. We're lucky they're even dropping the prices as much as they are as soon as they are since people like you and me won't be able to even get any conroes for some time to come.
No...you are absolutely right and just proved my point. AMD is as interested in profits and keeping prices high as much as any other company when they have a performance lead.
So all those folks who are hoping Intel will go bankrupt will just be trading one devil for another, but somehow don't have the foresight to see this.
Even you acknowledge we are lucky they will be dropping prices with limited Conroe availability - what would happen to those prices if there was no Conroe?
"Even you acknowledge we are lucky they will be dropping prices with limited Conroe availability - what would happen to those prices if there was no Conroe?"
I did not prove your point. You really don't have a point. If there was no conroe the prices would stay the same the way it should be. AMD has the best processor and they are charging accordingly (and the prices REALLY aren't that bad.) You can't just sell everything dirt cheap so your company makes almost no profits simply because there is something better coming out in the future. It's simply stupid to think that's how any business works.
"If there was no conroe the prices would stay the same the way it should be."
That's my point! If there is no competitive product there's little reason for a company to reduce prices.
So for the folks who hope Intel goes bankrupt, the rate at which prices go down on processors will slow considerably as there will be no pressure to, so hoping this makes absolutley no sense.
Personally as I own an Audi, I hope all car makers go bankrupt so everyone will own Audi's. I'm sure Audi will remain incentivized to keep lowering prices on new cars and adding features and performance at their current rate despite the fact there will be no more competition...make sense to you?
"That's my point! If there is no competitive product there's little reason for a company to reduce prices."
That wasn't really part of the conversation until now, but you are somewhat correct. However, I highly doubt that AMD would get more and more expensive over time. Only the latest and greatest would keep the high price point. I would still do what I always do and buy cheap and overclock like my Opteron 146 which is 3.0ghz prime stable (2.85ghz in the summer :/ .) The difference between AMD and Intel is that Intel is a horrible company and really does thrive on being a monopoly where AMD actually works with everyone in the technology world to bring out the latest and greatest and keep pushing technologies. I'm positive the innovation would not stop even without Intel due to the fact that EVERYONE would still be working on the technologies like HyperTransport and everything else that runs A64 instead of just intel making everything that runs your Intel system. Intel with no competition has no reason to excel as AMD does.
"That's my point! If there is no competitive product there's little reason for a company to reduce prices."
I just replied to this, but I forgot to add that this has no relevance at all to the subject at hand. Of course no one has a reason to drop prices if nothing else is competitive. Why would you? If you owned the only car manufacturing company in the world would you drop prices to the point where you're making very little/no profits simply because there is nothing competitive? I'd hate to have you running my business.
Now if you want to discuss this in terms of AMD being the last man standing then the situation doesn't change much. As I said AMD has a reason to keep innovating and pushing the technology envelope as Intel does not so there would always be "competitive" products from AMD going above and beyond the previous ones and if you honestly think AMD would not lower their prices on the older technology then you're simply kidding yourself.
I'm not biased on any of this from a hardware perspective and If I had to build a brand new system right this second I'd love to have myself a shiny new conroe. I have a 3ghz Opteron that I'm very happy with so I'm sitting back and seeing what happens with 4x4 before I upgrade. The only problem is I'm very biased against Intel as a company for a damn good reason. They're lying bastards and there's really no one who can deny that. I do, however, have the utmost respect for AMD and trust them in the event that they become a monopoly.
OK I'm done arguing with you aftre thie as you are not making any attempt to understand my comments.
I've been saying the rate of innovation would slow - what would push them? Just like it slowed at Intel until AMD put out a better product(K8) and lit a fire under Intel by grabbing market share.
Once you own a market and have limited competiton (from market share perspective) the only way to grow further is to expand the market or increase gross margin. Ultimately AMD is a business and their goal is to make money and provide return to those who own the business (stockholders). If this is not their goal then they should file for a non-profit organizational status and simply charge enough money for them to break-even, pay their employees and pay for continued R&D work.
"I do, however, have the utmost respect for AMD and trust them in the event that they become a monopoly."
I think this quote says it all; you have already made up your mind. I understand this is your opinion and it is no better/wore than mine, but you are closed to the idea that if AMD were to obtain a dominant market position their attitude toward their customer may change. In a recent interview Hector has already stated they are focusing on server market and mobile moving forward (I believe because this is where the better margins and revenue are vs desktop products wher both margin and vlumes are diminshing and as area that AMD owes most of its loyal fanbase to).
As a further example - why doesn't AMD continue to sell Opteron chips with the same socket as Athlon anymore? (Is their a specific technical advantage/benefit to the consumer to have 2 different sockets? Or is this to protect their FX62 product and margin - most savy AMD fans know Opteron is a better and far cheaper solution, why would AMD not conitnue to offer this as a desktop solution?)
Congrats Sharikou! You now have a single blog with over 100 comments! :)
6
You seem to think that I'm all for Intel dying. I don't think that Intel going bankrupt would have any effect on anything other than ridding the world of a lame company. AMD is going to continue to innovate and push things further and further regardless with the help of the likes of the Hyper Transport Consortium and SUN and a whole list of other companies who are constantly working on new projects. AMD is NOT the only company that works on their products. That is what separates the two companies as well as the fact that Intel is run by marketing people while AMD is run by innovators like Hector Ruiz and others who actually make processors and know what the hell is going on.
I'm really not sure what would happen with Intel out of the picture, but I'd have to say that my personal belief is that things would get a lot better as all of the companies that work together with AMD would all prosper quite a bit more and be able to put more money into R&D and speed up technical advancements which is clearly better for everyone.
"I don't think that Intel going bankrupt would have any effect on anything other than ridding the world of a lame company."
One of the stupidest comments I have ever read here.
Followed by a second one...
"...my personal belief is that things would get a lot better as all of the companies that work together with AMD would all prosper quite a bit more and be able to put more money into R&D and speed up technical advancements which is clearly better for everyone."
You cant make this stuff up, and the sad thing is he/she is not kidding... LMAO!!!
Well everyone knows it is much better to have a monopoly than having competition - isn't that one of the basic tennets of a free market?
Teach us more, oh wise one!
:)
"You cant make this stuff up, and the sad thing is he/she is not kidding... LMAO!!!"
You fail. You still think Intel and AMD operate in the same fashion. There's really nothing similar about them. Intel being a monopoly would be completely different than AMD being a monopoly since AMD is partnered with so many groups/companies that produce their next gen products. If you really think that AMD would cut off it's partnerships then you're a complete retard.
If you think my views are so far off then please tell me how you think it would happen. I'm guessing you're going to spew out some meaningless crap still basing your thoughts off the belief that AMD is remotely like Intel.
Without competition there is no need for innovation, and the 3 or 4 year cycle we are on now would probably be more like 5 to 6.
I am not trying to bash but isn't AMD on its 3rd year with K8?
Do you think Intel would have changed their roadmap to have a new architecture every two if AMD was not around?
Look at ATI and NVidida, what would be the benefit of building a better product if there is nothing to compare it too?
Picture this you just built an extremely dominant processor, and your competition goes bankrupt. Why on earth would anyone allocate money to your research and development, to beat your own design, that has no competition?
Hence going back to the original comment, the cycle will get much larger, thus innovation would slow considerably.
Post a Comment
<< Home