Intel to lose market share to Barcy

i'm not talking about flaming.

in the Integer test, a Barcelona 2.3GHz yields 21% higher score than Clovertown 2.66 GHz, but Floating Point test leaves a staggering 50% performance deficit for Clovertown, and this is not something 45 nanometre Penryn can solve overnight. Unless, of course the clock deficit for AMD is such that Intel speeds past.

AMD's still in the game and price drops will probably be re-calibrated.
intel will suffer market share loss in the server market. it's an impressive score and we'll see everything about it.

i'm still deceived in their desktop max overclocking ability...

i explain. C2D can easily overclock at 3ghz with 15-25% more efficient/clock

while AM2 CPU can go up to about 3ghz too...

i feel intel platform to be more evolutive now. i think amd will need another platform to push another better CPU...

intel got C2D out and lowered the prices to regain market share in the desktop market. this will hit harder for intel that it did for amd on the desktop market... phenom will perform with about the same efficience/clock but needs less stability, un-buffered memory, will be clocked higher... ( at the cost of limiting to 1 cpu socket ). by that time intel must aim an higher spot and hope it's high enough...

competition is good for business, i hope they will ever stay with similarly performing technologies.
68 answers Last reply
More about intel lose market share barcy
  1. You have benchmarks to share or are you just dreaming up numbers?
  2. Sounds like dreamed up numbers to me. Yea AM2 can clock to 3.0 but not much more than that. Core 2's have been know to hit well over 3.0. Not to mention clock for clock core 2's outperform AM2.
  3. The inquirer has a new article with data taken from a Russian AMD powerpoint:

    http://vd.verysell.ru/files/ie/252_10_DOCUMENT_Server_Section_PUBLIC.ppt

    However, unsurprisingly they focus on the less relevant SPEC rate tests, as opposed to TPC-C, SPECjbb2005 and SAP-SD which are important enterprise benchmarks. Using AMD's values for Barcelona, the X5355 should beat the Barcelona system in TPC-C by a bit, easily win SPECjbb2005 and be slightly behind in SAP-SD.
  4. Yea I read about the multitudes of AMD fanboys on here. Guess the rumors were true.
  5. Quote:
    Yea I read about the multitudes of AMD fanboys on here. Guess the rumors were true.


    There are Intel Fanboys claiming the same doom and gloom. It gets really old.
  6. Quote:
    Sounds like dreamed up numbers to me. Yea AM2 can clock to 3.0 but not much more than that. Core 2's have been know to hit well over 3.0. Not to mention clock for clock core 2's outperform AM2.


    first, that's exactly what i said...
    UP TO 3ghz for A64 X2
    and EASILY 3ghz for C2D

    i'm not on intel's side
    i'm not on amd's side
    they're all the same and will always outperform each other in terms of performance... it's not even a question of ressources. it's a question of aiming to the right point with the best timing.

    then, this is my first source. and the least reliable.
    http://www.theinquirer.net/default.aspx?article=40749

    same kind of information intel has made giving only the results.

    from what i've seen, these results are 95% correct...
    more will follow
  7. *sigh*

    Okay, first, take a Cloverton 2.66GHz CPU and test it. Then take the Barcelona 2.3GHz CPU and test it.

    What? You can't. Why? Oh, that's right. There is no Barcelona CPU at 2.3GHz to test against. How about 2.0GHz? No?

    So, the real answer is that there is nothing for Intel to lose against.

    It's also funny that they are claiming to run a 2.3GHz Barcy chip, when AMD announced that it was going to release Barcelona up to 2.0GHz. Yes, UP TO.

    Plus, what accord99 has linked is the presentation that the Inq. based their story on. Let's see. When was that dated? Feb. 2007? Hmm... Hasn't there been a lot of things happen since then? Like the no-show in Tunasia? The lack luster launch of R600? The disappointing showing at Computex?

    Quote:
    from what i've seen, these results are 95% correct...
    more will follow

    Based on what? The Inq. article or the months old slide? Or are you basing it on the non-existent 2.3GHz Barcelona? Did you happen to get your hands on one of the few binned out at 2.3GHz? Cause those are rare. So rare, that they didn't even have one running at Computex.

    You can call me a fanboy or not. I'm just tired of seeing all these facts based on estimated projections (yes, ESTIMATED PROJECTIONS) of supposed top speed binned CPUs, and people screaming how it's fact, and how they have proof. This isn't proof. It's a marketing slide show.

    When Barcelona does arrive in Q3, and after some 3rd party testing, then you can claim all you want. Until then, it's just more hot air.
  8. AMD the king of Slides. 135 Slides MY GOD.

    You gotta take everything from the INQ with a bucket of salt.
  9. These are old numbers. According to this article - http://www.xbitlabs.com/news/cpu/display/20070702235635.html - the comparison was for "performance per clock". Even if the numbers are correct, by the time Barcelona gets shipped we will be comparing 2.0 Barcelona with 3.0 Clowertown (50% faster clockwise) and Clowertown will be on top in all benchmarks. And we are not even talking about Penryn.
  10. Quote:
    in the Integer test, a Barcelona 2.3GHz yields 21% higher score than Clovertown 2.66 GHz


    http://multicore.amd.com/us-en/AMD-Multi-Core/Products/Barcelona/Performance.aspx

    It's 2.6GHz K10 besting an OLD X5355 result.
    http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070528-01175.html
    http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070612-01275.html
    That's not a 21% advantage anymore now is it?

    Quote:
    but Floating Point test leaves a staggering 50% performance deficit for Clovertown


    That's going to be useful. In supercomputers, with thousands in volume lol.

    Quote:
    and this is not something 45 nanometre Penryn can solve overnight.


    I'll be damned, you sound just like the INQ.

    Quote:
    Unless, of course the clock deficit for AMD is such that Intel speeds past.


    2 vs 3.33. Hmm.

    And the rest of what you said made no sense.
  11. *sigh*

    Quote:
    Okay, first, take a Cloverton 2.66GHz CPU and test it. Then take the Barcelona 2.3GHz CPU and test it.


    well, done, i just have to pump up the fsb a little and i have the equivalent.
    engineering samples aren't real barcy so i give you the point.
    i don't have a real barcy and i can't be certain barcy performance will be limited to the performance my sample will give. but there's something weird about it. 38 degrees under full load seems a bit low... while 33 on idle seems possible... with the speed at which it passes between 33 and 38 degrees, it can't be the socket temp. ...


    Quote:

    It's also funny that they are claiming to run a 2.3GHz Barcy chip, when AMD announced that it was going to release Barcelona up to 2.0GHz. Yes, UP TO.


    you misunderstood something. barcelona up to 2ghz are those which are power-efficient cpu's. and all have a TDP of 68W at launch using 17W/core
    this is their feature product from the beginning and will cost up to 390$

    the same product with a TDP of 95 watt will be available in versions of 1.9/2.0/2.1/2.2/2.3 ghz and will be up to 795$ will be released between august and november ( probably october )

    lastest of the line, 2.4 and 2.5ghz products will have a TDP of 120W and be released in Q1 2008 unknown date.

    Quote:
    So rare, that they didn't even have one running at Computex.

    Oh and if you had one, you would've shown it at computex ? an authorisation is needed to publicly perform benchmarks over an engineering sample.

    no benchmark will be seen before they want to... any announcement will only be denied.

    Quote:

    When Barcelona does arrive in Q3, and after some 3rd party testing, then you can claim all you want. Until then, it's just more hot air.


    22nd july intel price cut... you'll see more until that date
  12. and no, i'm running at 2.3ghz and it's a 65W lightly overclocked from
    1.9ghz... the 320$ part.

    of course, the performance gap to the xeon 5355 is the same as the months old paper comparison for the 2.6ghz.

    i'm tired of this,

    maybe more to come next week
  13. copying word for word from the Inquirer doesn't mean you know something...
  14. Quote:

    well, done, i just have to pump up the fsb a little and i have the equivalent.
    engineering samples aren't real barcy so i give you the point.
    i don't have a real barcy and i can't be certain barcy performance will be limited to the performance my sample will give. but there's something weird about it. 38 degrees under full load seems a bit low... while 33 on idle seems possible... with the speed at which it passes between 33 and 38 degrees, it can't be the socket temp. ...


    You have an engineering sample, yet you don't believe engineering samples are real Barcelona CPUs? Why not? They are based on the Barcelona core, so they are, in fact Barcelona CPUs. Engineering samples aren't based on Opteron cores, so why not consider them real Barcelonas? I didn't say final spin of silicon Barcelona, but they are real Barcy CPUs, nonetheless.
    Plus, the fact that I don't believe you have an engineering sample.


    Quote:
    you misunderstood something. barcelona up to 2ghz are those which are power-efficient cpu's. and all have a TDP of 68W at launch using 17W/core
    this is their feature product from the beginning and will cost up to 390$

    the same product with a TDP of 95 watt will be available in versions of 1.9/2.0/2.1/2.2/2.3 ghz and will be up to 795$ will be released between august and november ( probably october )

    lastest of the line, 2.4 and 2.5ghz products will have a TDP of 120W and be released in Q1 2008 unknown date.


    Really?
    Please point out what part of AMD's official announcement I misunderstood.
    AMD Barcelona Announcement
    Oh, and while you're looking over it, please point out where it said that it was releasing it as an energy efficient CPU, and where the higher freq. processors are to be released. All I saw was this statement: "AMD expects its native quad-core processors to scale to higher frequencies in Q407 in both standard and SE (Special Edition) versions." Hmm. No frequencies, no TDPs, no real set dates. Pretty much nothing claiming what you wrote.

    Quote:
    Oh and if you had one, you would've shown it at computex ? an authorisation is needed to publicly perform benchmarks over an engineering sample.
    no benchmark will be seen before they want to... any announcement will only be denied.


    Yes, I would. Computex isn't a small gathering for the semiconductor industry. It's almost on the same level as CES, but for computer related products. It's probably the biggest show after CeBIT. So, why not show your best product? And that's complete bullsh*t if you think that AMD wouldn't have shown an engineering sample, if they could have. It would've helped show that they did have product running at speeds they claimed they could achieve. Authorization? It was their own DEMO.
    Quote:
    In the end, performance was absolutely terrible. We're beginning to understand why AMD didn't let us test Barcelona last month. It's not that AMD is waiting to surprise Intel; it's that the platform just isn't ready for production yet.

    The CPUs themselves work, once again, as we've seen from AMD's own demos. But the real question we've been asking has been "at what clock speed?" The chips we saw at Computex ran at either 1.4GHz or 1.6GHz, the latter being the latest B01 stepping.


    So, if AMD had a CPU that could run stable and fine for Computex, you believe they wouldn't have dropped it in a system, if only to not get that kind of negative press?

    Quote:
    22nd july intel price cut... you'll see more until that date


    I have no idea what the July price cut is suppose to prove or disprove.
  15. Baron? Is that you? :lol:
  16. Quote:
    Baron? Is that you? :lol:



    Give Baron a little more credit. He wouldn't steal, word for word, from the Inq. He makes up his own BS.
  17. Quote:

    Really?
    Please point out what part of AMD's official announcement I misunderstood.
    AMD Barcelona Announcement
    Oh, and while you're looking over it, please point out where it said that it was releasing it as an energy efficient CPU, and where the higher freq. processors are to be released. All I saw was this statement: "AMD expects its native quad-core processors to scale to higher frequencies in Q407 in both standard and SE (Special Edition) versions." Hmm. No frequencies, no TDPs, no real set dates. Pretty much nothing claiming what you wrote.

    Oh realy ?
    This table was on several tech websites week ago

    Models freq TDP 2P 8P
    2P 8P Price Release date
    2360 2,5 GHz 120 W – ? –
    2358 2,4 GHz 120 W $ 1 180 xx.10.
    2356 8356 2,3 GHz 95 W $ 795 xx.10. $ 1 550 – ? –
    2354 8354 2,2 GHz 95 W $ 610 xx.10. $ 1 190 – ? –
    2352 8352 2,1 GHz 95 W $ 450 xx.10. – ? – – ? –
    2350 8350 2,0 GHz 95 W $ 390 10. 9. $ 1 025 – ? –
    2348 8348 1,9 GHz 95 W $ 320 10. 9. $ 790 – ? –
  18. Quote:

    Really?
    Please point out what part of AMD's official announcement I misunderstood.
    AMD Barcelona Announcement
    Oh, and while you're looking over it, please point out where it said that it was releasing it as an energy efficient CPU, and where the higher freq. processors are to be released. All I saw was this statement: "AMD expects its native quad-core processors to scale to higher frequencies in Q407 in both standard and SE (Special Edition) versions." Hmm. No frequencies, no TDPs, no real set dates. Pretty much nothing claiming what you wrote.

    Oh realy ?
    This table was on several tech websites week ago

    Models freq TDP 2P 8P
    2P 8P Price Release date
    2360 2,5 GHz 120 W – ? –
    2358 2,4 GHz 120 W $ 1 180 xx.10.
    2356 8356 2,3 GHz 95 W $ 795 xx.10. $ 1 550 – ? –
    2354 8354 2,2 GHz 95 W $ 610 xx.10. $ 1 190 – ? –
    2352 8352 2,1 GHz 95 W $ 450 xx.10. – ? – – ? –
    2350 8350 2,0 GHz 95 W $ 390 10. 9. $ 1 025 – ? –
    2348 8348 1,9 GHz 95 W $ 320 10. 9. $ 790 – ? –

    Yes, really.

    So, where are they basing that information from? AMD? The announcement I linked is from AMD. So, that is all speculation of releases, not facts.

    Plus, having "?" in the release dates doesn't really help your argument much.
  19. How many watts does it take?
  20. Quote:
    How many watts does it take?


    More than 10, but less than a 100K.

    Depends on the heat spreader on CPU.

    :wink:
  21. :lol: True that!
  22. Thanks noob for wasting my time on this stupid thread.

    If it were possible, I would sue you for the time wasted by clicking on this stupid thread and the time for typing my response.
  23. Quote:
    Baron? Is that you? :lol:


    no, i'm not that guy. and i can't be sure about the release date but i do know how to clock my 1.9 ghz cpu on a socket F board and squeeze 2.3ghz out of it... it doesn't seem to get up to 2.4ghz easily. i feel it's not an overheating problem, as i only have 42 degrees with a standard opteron heatsink from the first gen series at the unstable clock speed of 2.8ghz... (400x6). and it's a socket F CPU.

    i can't tell if my idea is correct but it is like a quad BE-2XXX with shared cache and some load balancing optimisation.

    performance is between 145%-180% from actual opteron performance. more specifically into floating point benches as suggested.

    i didn't test side by side with a clover-thing... i don't have any. i compared with an actual opteron and clocked it almost the same way i did with the unknown CPU.

    How many watts does it take? i don't know. it's supposed to use 68W at regular speed... performance got really high really fast while i upgraded the bandwith...
  24. Quote:
    Baron? Is that you? :lol:


    blah blah blah blah... at the unstable clock speed of 2.8ghz... (400x6). and it's a socket F CPU.


    You can't even do basic math - post CPU-Z screenshots and validation and photo's of your mystery CPU or you're full of $hit....
  25. Floating points are not the CPU holly grail, post a cpu-z shot. 8)
  26. it's always been interesting to me how many people in "forumz" brush off fpu results as something only useful in large scientific settings. Guess these people have never used their computers to encode an MP3 or have done any video editing/creation... or played an FPS heh.
  27. Quote:
    You can't even do basic math - post CPU-Z screenshots and validation and photo's of your mystery CPU or you're full of $hit....


    Spot on! :trophy:
  28. Quote:
    Thanks noob for wasting my time on this stupid thread.

    If it were possible, I would sue you for the time wasted by clicking on this stupid thread and the time for typing my response.


    Yea, sue'em! That's the American way!

    Or maybe you could step into a thread like this knowing up front that you might be knee deep in crap. Why? Because this is Tom's and it's another thread about chips that aren't yet available.
  29. Quote:
    it's always been interesting to me how many people in "forumz" brush off fpu results as something only useful in large scientific settings. Guess these people have never used their computers to encode an MP3 or have done any video editing/creation... or played an FPS heh.

    Except SPECfp_rate isn't so much FP as it is memory bandwidth. An FX-74 would beat a QX6700 by about 40% in SPECfp_rate 2006, but the QX6700 is faster in MP3, video editing/creation and games.
  30. Wouldnt a higher IPC plus higher FP be good on those apps?
  31. For a different take... Overclocker's article; Barcelona due for re-do It's Ed's opinion that AMD may have serious yield and speed problems. I need a desktop chip not a server chip, when will AMD get the Phenom out for us. This year or next year?
  32. Quote:
    post a cpu-z shot. 8)


    i'll try to know what i can or can't post...
    cause we have instructions not posting any benchmarks due to hardware specs and beta bios state

    Quote:
    I need a desktop chip not a server chip, when will AMD get the Phenom out for us. This year or next year?


    next year... i'm deceived too... they should've been faster on this one...
  33. Quote:
    post a cpu-z shot. 8)


    i'll try to know what i can or can't post...
    cause we have instructions not posting any benchmarks due to hardware specs and beta bios state

    I'll back up morg a bit.


    While I don't have any engineering samples, I do know people that do.

    I've been told of a 40-50% increase over a similar clock (IIRC) clovertown - but only for 3/4 threaded apps (must be the monolithic design).
  34. Quote:
    Guess these people have never used their computers to encode an MP3 or have done any video editing/creation... or played an FPS heh.


    The former two probably use SSE, and the latter should if it's a recent game.

    The performance of the legacy floating point unit isn't terribly important to most people anymore; I presume that's what the SPEC test is measuring?
  35. geez... the only info shown by cpu-z is the frequency the cpu is running at and the core selecting from 1 to 4...

    i was told to hide a few other infos given. the only thing i can still show is the working frequency... and the core 4 / thread 4 which doesn't even shows frequency for core 3 and 4...

    there's no fsb/multi shown.
    only a stepping/familly/model which i must remove from the shot...

    i don't feel like cpu-z would be of any help...
    i hoped i would at least see bus and fsb speeds...
  36. THIRD
    PARTY
    BENCHMARKS
    PLEASE


    And no, the Inquirer does not count. I don't even think they'd know how to install a processor, let alone benchmark one. I look forward to THG's review.
  37. i must not validateand send infos to cpu-z!
    a cpu number is shown on cpu-z. by validating, i'll publicly post that we're giving infos about it... it's not safe if we want to have future test samples.

    i could make it post was 2.8ghz being unstable...
    it may run stable at 2.4ghz but require 1.1volt which is way over the stock voltage. which cpu-z is unable to show... you want a censured screenshot of cpu-z showing a cpu running at 2.3ghz without anymore info ? i already told most info boxes were empty... also, there's infos i can't reveal as for cpu voltage, stepping info and the name which shows an 8 character serial. want a shot ? it's deceiving...

    where is it possible to upload a picture...
    uploaded there... http://img30.picoodle.com/img/img30/8/7/5/f_cpuzshotm_532be6a.png

    we managed to make some infos available. switched the motherboard. there's more infos i can show now... still i blacked out some details... this mobo is more permissive to higher fsb... however, the cpu voltage might be wrong as no modification have been made to it... didn't do a stability test on this one. it might damage the cpu if the voltage isn't safe. stock voltage shown on previous motherboard was 0.7v. we pumped up to 1.1volt to get the 2.4ghz but we lowered the fsb and got +.5 on multi. it got up to 480 mhz fsb with 7/8 divider
  38. Fake photo, and you're full of it...sorry.
  39. It's obvious this guy is full of shit. I mean, there is a lot of evidence as of now that K10 won't be a Clovertown/Kentsfield killer. George Ou hit it right on the spot.

    Just ignore pathetic fanpois like him, maybe they will stop posting these made up stories.
  40. what makes you think it's a fake...

    cause just telling it's a fake don't convince me... validation won't be made i told that already. i'm not supposed to post anything about it but i asked and got approved but i must hide some specs... then, i MUST save it into paint and put black squared on the spots with ID numbers.

    this is a sample for ATI to make correct drivers for a better platform. the CPU is unbranded and has a serial where it should have a name.
  41. Quote:
    what makes you think it's a fake...

    cause just telling it's a fake don't convince me... validation won't be made i told that already. i'm not supposed to post anything about it but i asked and got approved but i must hide some specs... then, i MUST save it into paint and put black squared on the spots with ID numbers.

    this is a sample for ATI to make correct drivers for a better platform. the CPU is unbranded and has a serial where it should have a name.


    If youre not supposed to post anything about it, then why are you posting about it? If you cant provide proof of anything, then maybe you shouldnt discuss it.
  42. morg might just be telling the truth. look at the cache structure in the screenie.
  43. yeah, and what is chronicles of narnia out in dvd format all over the internet while it's not out yet... of course, they hide the provenance of it... the frames with dot-marks have been removed ... pfff

    man, ppl want to know things... amd telling their architecture is lower-ended and apologies for the max 2ghz... yes it's true... about what i see, it can't go past 2.4ghz, works at low voltages... but hey, from what i've seen, at 2.3ghz, (while bandwith way over the stock 266mhz, it goes way faster... about the same specs the 2.6ghz based on what many told me... thinking i was only talking about the month's old specs )

    this sample comes from ATI... they're giving stress-tests to new chipsets... and it doesn't go very well... i didn't see the thing... they're trying to get some 4 socket quad core chipset and squeeze some extra performance from it... but the chip works well on dual socket motherboard... at 2.3ghz, the chip seems stable enough but i can't tell if it's safe to use as server at this frequency... however... the chips comming after august will feature up to 2.3ghz and i bet they will be as strong as what i have now.

    why would they put 4 cpu on the board ? to have 4 independant memory controller with more facility that intel quads will do... that's why the chip starts at 320$... hey... 4x4 cores, ram will follow the specs... 4 ram slot / CPU makes more ram... i don,t know if this could support it but i think 4 cpu x 4 slot/cpu makes 16 slots of 4 gb making up to 64 gb into a not so costy server... 16 1.9ghz cores for 1280$ ... reducing TDP also goes into that direction. 4x68W makes 272 watt for cpu.

    i still have several questions...
    is a native quad more prepared to quad socket compared to a 2x2 architecture ?

    is there real advantage of using 4 active memory controller instead of 2 ?
  44. Quote:
    *sigh*

    Okay, first, take a Cloverton 2.66GHz CPU and test it. Then take the Barcelona 2.3GHz CPU and test it.

    What? You can't. Why? Oh, that's right. There is no Barcelona CPU at 2.3GHz to test against. How about 2.0GHz? No?

    So, the real answer is that there is nothing for Intel to lose against.

    You are right, its easy not to lose to unreleased product :wink:

    Quote:

    You can call me a fanboy or not. I'm just tired of seeing all these facts based on estimated projections (yes, ESTIMATED PROJECTIONS) of supposed top speed binned CPUs, and people screaming how it's fact, and how they have proof. This isn't proof. It's a marketing slide show.

    When Barcelona does arrive in Q3, and after some 3rd party testing, then you can claim all you want. Until then, it's just more hot air.


    You are not fanboy Dante, you work for Intel 8)

    I also think its hot air, same as most Intel benchmarketing claims. Its nice at least Intel started new path with more or less reliable testing of cpu's way before launch, - new tradition since C2D, though not as open and wide-spread at Intel as we would like. Do you remember what was before that? Exactly the same as underdog AMD doing.

    On the side note, I think Barcy has a good shot, at least better chance than desktop parts. Why? Lets take a look at current Opterons. Although C2D kills X2 in high-end, the picture isnt the same in servers market. New spanking Xeons are barely faster than old-core Opterons, depends on tests, in few AMD still manages to win. Why its so? IMO its because of the way cpu's vision is based on. Intel makes desktop cpu's and then "upgrades" to Xeons, while AMD makes server cpu's and then "downgrades" to desktop.

    Now, we all know Barcy is significantly improved over Opterons. AMD fans says its best thing since sliced bread, Intel fans says its a flop before even seeing one. I think the truth is somewhere in the middle :) If current Opterons are losing to Xeons by a small margin, its not unreasonable to expect new Barcy to hold pretty good on Penryn based Xeons.

    The major issue is the same as I was saying half a year ago - cpu speed ramping. AMD probably will release with up to 2GHz in the month or so, then reach 2,4-2,6GHz by the end of the year (maybe). The question how it will fare against 2,93GHz x7350 (Tigerton)? If Barcy somehow would manage to compete as even with Tigerton, Intel would have no problem to release speadier version, while it seems AMD have problems with reaching 2,6Ghz, what to speak of 3GHz.

    My prediction: Intel will own one socket low-end quad cores servers, will be slightly faster or even on 2P, and AMD (in best case) having an advantage in 4P+.
  45. Quote:
    i'm not talking about flaming.

    in the Integer test, a Barcelona 2.3GHz yields 21% higher score than Clovertown 2.66 GHz, but Floating Point test leaves a staggering 50% performance deficit for Clovertown, and this is not something 45 nanometre Penryn can solve overnight. Unless, of course the clock deficit for AMD is such that Intel speeds past.


    This was a simulated 2.6 ghz Barcelona tested vs. an older Clovertwon 2.66 Ghz. AMD has already publicly stated that those numbers are wrong currently.
  46. Quote:
    You are not fanboy Dante, you work for Intel 8)

    I also think its hot air, same as most Intel benchmarketing claims. Its nice at least Intel started new path with more or less reliable testing of cpu's way before launch, - new tradition since C2D, though not as open and wide-spread at Intel as we would like. Do you remember what was before that? Exactly the same as underdog AMD doing.

    On the side note, I think Barcy has a good shot, at least better chance than desktop parts. Why? Lets take a look at current Opterons. Although C2D kills X2 in high-end, the picture isnt the same in servers market. New spanking Xeons are barely faster than old-core Opterons, depends on tests, in few AMD still manages to win. Why its so? IMO its because of the way cpu's vision is based on. Intel makes desktop cpu's and then "upgrades" to Xeons, while AMD makes server cpu's and then "downgrades" to desktop.

    Now, we all know Barcy is significantly improved over Opterons. AMD fans says its best thing since sliced bread, Intel fans says its a flop before even seeing one. I think the truth is somewhere in the middle :) If current Opterons are losing to Xeons by a small margin, its not unreasonable to expect new Barcy to hold pretty good on Penryn based Xeons.

    The major issue is the same as I was saying half a year ago - cpu speed ramping. AMD probably will release with up to 2GHz in the month or so, then reach 2,4-2,6GHz by the end of the year (maybe). The question how it will fare against 2,93GHz x7350 (Tigerton)? If Barcy somehow would manage to compete as even with Tigerton, Intel would have no problem to release speadier version, while it seems AMD have problems with reaching 2,6Ghz, what to speak of 3GHz.

    My prediction: Intel will own one socket low-end quad cores servers, will be slightly faster or even on 2P, and AMD (in best case) having an advantage in 4P+.


    Yes, I do work for Intel (right now). But I have never pushed the fact that Intel was better or a better choice for anyone, unless they asked for a comparasion between 2 different CPUs (ie. C2D vs. X2 for certain applications).

    Working for a company does not make me a fanboy, it makes me a bit jaded in seeing and reading supposed "factual" claims, that aren't facts. I've seen enough propaganda during the Netburst years to realize when a company is trying to polish the apple to hard, so it's hard for me to swallow a lot of the "facts" being pushed.

    For example, this recent "benchmark" results, that are actually dated in February. Also, the claims of "newer" benchmarks, dated in April. Those are misleading, and pushed FUD, to new levels. The sad part is that it's not the normal circles pushing it (The Inq or Fudzilla), but some well known tech sites. Even after AMD themselves have said that those slides were using older Intel benchmarks, some of those same sites are still pushing those slides as hard "facts" of Barcelona's power, and that's not all together true.

    Now, I'm not going to say that Barcelona isn't going to be a good CPU for the server market, nor am I going to say that it's a flop, cause I haven't really seen enough data to come to either conclusion. Whether or not it will be the savior CPU that AMD is hoping, only time will tell, but right now, there is an air of disappointment in the release of a CPU that will hit up to 2.0GHz only, especially in light of the "estimated" powerpoint slides that were shown, and words spoken by AMD management.

    I will end my rant about this, but I personally believe that morg's CPU-Z picture is not 100% credible (why does it look like crap? Even if I censored a CPU-Z picture in MS Paint, it wouldn't save in 256 bit color scheme, unless I planned to do so).

    @Boduke - was the BS bag for me? If it was, thanks, I guess. If not, okay. :D
  47. First, it is true that 'people want to know things', however they want to know 'things' from reliable sources, such as third party test sites, for example Anand or THG.

    I'll go out on a limb here, and say 'most people' do not want to know things from AMD(except availability and price), or 'some guy' on the internet who claims to have something that might or might not be a K10, because he cant say and cant provide proof of anything about it. Niether of those information sources can be described as trustworthy, and why? Because AMD has already told everyone so many 'things' in relation to K10, that their 'things' have long since been conflicting and confusing, making them untrustworthy. As far as people on the internet go, they've made more bogus uninformed claims than AMD has.


    Oh, BTW, CPU benchmarks are not comparable to a pirated movie floating around the internet.
  48. Edit - that wasn't supposed to be a quote.

    Working for Intel doesn't make you Biased? I think you lack a little credibility but what the hell lets give you the benefit of the doubt.

    misleading benchmarks??? BOTH sides (that would be Intel and AMD) do it, it's dirty and screws the consumer more than the competitor but what the hell business is business, that's the American way....
Ask a new question

Read More

CPUs Intel AMD Product