Page 2 of 3 FirstFirst 123 LastLast
Results 26 to 50 of 55

Thread: The demo of AMD's Barcelona...

  1. #26
    Xtreme Addict
    Join Date
    Nov 2006
    Posts
    1,402
    i think more and more that A2 is a barcelona one, and A2 will reserved for server.

    If this is true, that will be a socket F, AMD will release one super FX on QuadFather with 8 cores. 2*quad.

    The desktop version may release 2 or 3 months after.

    The socket F is fully compatible with Quad core, the FX is not i heard, and AM2 no information about.

    AMD sayed it'is but today i have some doubts. I don't remember which web site give me some doubts sorry.

    The demo is very impressive, but i'm always hungry of benchs

    The QuadFather may perform well in vista ultimate i heard.

    XP can't use NUMA ( AMD use ccNUMA ), toms don't use node mode in bios

    I think that AMD will release the Barcelona in end of Q1. And the desktop version ( AM2 ) in end of Q2.

    When i contacted two moths ago AMD press service they said me "quad core is for H1".

    I send a mail one hour ago, i asked about 15 questions for 4x4 quadcore and 65nm future. If i have any information, i will write for evrybody here

  2. #27
    Xtreme Addict
    Join Date
    Sep 2005
    Location
    Las Vegas, NV
    Posts
    1,771
    Quote Originally Posted by Lightman
    Did he mention in what application??
    BTW 170% when 100% means 2x faster?? or when means equal speed??

    PS. You must buy a pack of beers for your friend! Mayby then he will tell you more
    I believe he means equal speed.. so I guess only 70% faster than Fx-62

    ~Mike
    Main Rig: Intel Core i7 7700k @ 4.2GHz, 64GB of memory, 512GB m.2 SSD, nVidia GTX1080Ti
    NAS: QNAP TVS-1282, 8 x 4TB WD Golds(Main Storage Pool), 4 x 960GB M4 Crucial (VM Storage) , 2 x 512GB M.2 Caching
    Private Cloud: 4 Nodes (2 x Xeon 5645, 48GB DDR3 ECC/REG, 1 x 1TB HDD, 1 x 960GB SSD/Each)
    Distributed Encoding Cloud: 4 Nodes (2 x Xeon x5690, 24GB DDR3 ECC/REG, 1 x 128GB SSD/Each)
    Feedback
    EBAY:HEAT

  3. #28
    YouTube Addict
    Join Date
    Aug 2005
    Location
    Klaatu barada nikto
    Posts
    17,574
    twice the number of cores, one would assume it would perform 200% of the dual cores assuming perfect scaling
    Fast computers breed slow, lazy programmers
    The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay.
    http://www.lighterra.com/papers/modernmicroprocessors/
    Modern Ram, makes an old overclocker miss BH-5 and the fun it was

  4. #29
    Xtreme Mentor
    Join Date
    May 2005
    Location
    Westlake Village, West Hills
    Posts
    3,046
    Quote Originally Posted by Teroedni
    And he also got a sens of humor

    Don't expect many fireworks though, we are told the demo consists of a four socket, 16 core beast running, wait for it, task manager! Yes, task manager, that hugely transactional FP heavy system thrasher will be shown off working away on all 16 cores. We are told it scales flawlessly across the cores, probably the crowning achievement of HT in action.
    I got a good laugh from that too.
    PC Lab Qmicra V2 Case SFFi7 950 4.4GHz 200 x 22 1.36 volts
    Cooled by Swiftech GTZ - CPX-Pro - MCR420+MCR320+MCR220 | Completely Silent loads at 62c
    GTX 470 EVGA SuperClocked Plain stock
    12 Gigs OCZ Reaper DDR3 1600MHz) 8-8-8-24
    ASUS Rampage Gene II |Four OCZ Vertex 2 in RAID-0(60Gig x 4) | WD 2000Gig Storage


    Theater ::: Panasonic G20 50" Plasma | Onkyo SC5508 Processor | Emotiva XPA-5 and XPA-2 | CSi A6 Center| 2 x Polk RTi A9 Front Towers| 2 x Klipsch RW-12d
    Lian-LI HTPC | Panasonic Blu Ray 655k| APC AV J10BLK Conditioner |

  5. #30
    Xtreme Enthusiast
    Join Date
    Oct 2006
    Posts
    896
    Your math is wrong nn_step. 200% would be quite an achievement This is quite silly for some people to get hyped over some AIM conversation Too bad though, the best numbers AMD could give us was 70% over dual core in OLTP, barely above Clovertown, and 40% over dual core in SPECfp_rate2000, which seems to be more of an architecture thing rather than a quad thing. Wish AMD would give us more concrete info.

  6. #31
    Xtreme Guru
    Join Date
    Jan 2005
    Location
    Tre, Suomi Finland
    Posts
    3,858
    Quote Originally Posted by red
    Quote Originally Posted by nn_step
    twice the number of cores, one would assume it would perform 200% of the dual cores assuming perfect scaling
    Your math is wrong nn_step. 200% would be quite an achievement (...)
    Come on, red.
    It's common sense...

    If "A" performs 1% of "B", "A" is worse.
    If "A" performs 100% of "B", "A" and "B" are equal.
    If "A" performs 200% of "B", "A" is better than "B" by 100%.
    You were not supposed to see this.

  7. #32
    Xtreme Addict
    Join Date
    Nov 2006
    Posts
    1,402

    The demo of AMD's Barcelona points the way downhill

    Demo of AMD's Barcelona points the way downhill

    A lot more than you think

    By Charlie Demerjian: dimanche 03 décembre 2006, 15:43
    AMD'S demo of Barcelona on Thursday told us a lot more than you might think, and none of it looks good.

    OK, I lied, the demo of a working quad core part does look good, but it is downhill from there. If you recall, AMD laid the gauntlet down 'demo before the end of the year', and it did indeed do so with a full month to spare, so mad props to the peeps in AMD engineering or somesuch.

    A couple more data points to keep in the back of your head, AMD showed off wafers of Barcelona at the Microsoft Global High Tech Summit - cool event, but unless you like supply chains, avoid it - on September 29. I was told it had chips up and running, and the wafer pictured was not the only one.

    Backtracking a little, wafers take around 12 weeks to wiggle their way through a fab and get from shiny blank disk to shiny patterned disk chopped up into expensive rectangles and stuck to a piece of fibre glass. In the process, they pick up more frequent flier miles than I got last year, so you're talking a lot. If you 'hot lot' this process, basically giving them priority at every step of the chain, you can shave two or so weeks off. Let's just say that from tape out to chips, counting travel time, you are at three months.

    That puts the tape out of Barcelona at the end of June if one presumes the AMDers didn't tape it out, get blindingly drunk, and forget about it for a few weeks while the hangover subsided.We first got word in August, one story short of the magic 33700 mark and something that is not a coincidence, but it had to have happened about two months before that.

    It then got another stepping back about a month ago, let's say October 29 to make the numbers easier. Backtracking 10 weeks says these changes went into the oven in the middle of August. I am not enough of a chips designer to know how many bugs you can work out without silicon to physically test, or why it would make sense to do the first rev if you knew of bugs, but this is what the data shows.

    If it put another batch in the oven, which it undoubtedly did, it had to have been after the middle of October because it would have to get A2 silicon in September and late October A1 parts back before the end of the year. In fact there are other reasons why it would have had to have it in the oven before mid-October, the main one is the Chrismakwanzakuah holidays, you lose a few weeks there. An analyst demo does not do much good if NY City is empty of said folk, they are probably in a place where the spit does not freeze on the sidewalk.

    Basically, AMD set a deadline of 'before the end of the year' and was then stuck with A1 silicon, good, bad or ugly. This is why you get a demo of task manager, and the press aren't even allowed close enough to play minesweeper. On the other hand, AMD got it to boot Windows, and that is no small task.

    If you want to play devil's advocate, Intel showed Tigertown running at speed over a month ago. This is a patently unfair comparison because Tigertown is a Conroe x 2 with a new chipset, a very different proposition from a new CPU. Still, that chip is what AMD will be up against in the market so it may not be technically correct, but you don't buy tech, you buy products. Intel let a bunch of hyperactive monkeys that masquerade as press beat a system at will vs running task manager with a stick wielding goon ready to break fingers should you touch. Again, patently unfair but still, that's the way the Chipzilla cookie crumbles.

    What it boils down to is that AMD had an internal time table that it planned to meet. You can be pretty sure that AMD was not planning six months out to run task manager in order to awe a group of industry analysts. It was planning on being far ahead of where it is now, and possibly running Solitaire or something else useful.

    It didn't happen, so I think Barcelona silicon is behind where it hoped it would be. This does not necessarily translate into a schedule slip, but it sure doesn't add a positive spin. Add in the change in rhetoric from Q2 to mid-07 and you have one hell of a conspiracy, or at least a convergence of messiness.

    To toss in a positive tidbit, I hear the AMD 65 nanometre process is coming out on the upper end of its expectations. This may not translate into earlier releases of parts, but it does bode a little better. So, with a single demo, AMD did what it said and signalled a lot more, most of it not terribly good. µ
    Source

    i need some help to understand well, English is not my first language. Thx
    Last edited by Cooper; 12-03-2006 at 10:14 AM. Reason: please add source link next time ;)

  8. #33
    Xtreme Member
    Join Date
    Jun 2006
    Location
    Arizona
    Posts
    329
    Ok.. rabble rabble rabble, AMD may, or may not, be behind schedule a bit.
    stuff goes here.. etc etc.. i win.. lalala.. pwn etc

  9. #34
    YouTube Addict
    Join Date
    Aug 2005
    Location
    Klaatu barada nikto
    Posts
    17,574
    AMD is doing what they need to do, end of discussion
    Fast computers breed slow, lazy programmers
    The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay.
    http://www.lighterra.com/papers/modernmicroprocessors/
    Modern Ram, makes an old overclocker miss BH-5 and the fun it was

  10. #35
    I am Xtreme
    Join Date
    Apr 2005
    Location
    Upstate, NY
    Posts
    5,425
    Yes, now DONT LET THIS turn into a flame thread.
    Core i3-550 Clarkdale @ 4.2GHz, 1.36v (Corsair A50 HS/F) LinX Stable
    MSI H55-GD65 Motherboard
    G.Skill 4GBRL DDR3-1600 @ 1755, CL9, 1.55v
    Sapphire Radeon 5750 1GB
    Samsung F4 320GB - WD Green 1TB
    Xigmatek Utgard Case - Corsair VX550

  11. #36
    Xtreme Addict
    Join Date
    Nov 2003
    Location
    happy place
    Posts
    2,337
    is Barcelona a four-issue core?
    --===== proud owner of new razor tarantula gaming keyboard =====--

  12. #37
    Xtreme Enthusiast
    Join Date
    Oct 2006
    Posts
    896
    http://www.extremetech.com/article2/...2027635,00.asp
    Unlike Intel, AMD has chosen not to widen the instruction pipe, so the CPU is still a three-issue architecture.
    In case anyone missed it..


  13. #38
    Banned
    Join Date
    May 2006
    Location
    Skopje, Macedonia
    Posts
    1,716
    Quote Originally Posted by Kanavit
    is Barcelona a four-issue core?
    No, it is 3 issue.

  14. #39
    c[_]
    Join Date
    Nov 2002
    Location
    Alberta, Canada
    Posts
    18,728
    Anyone got benches handy to compare the 2220 SE, Xeon 5180, and 5355 to see if they are close to "on par" with the graphs?

    Either way, 4x4 is for the enthusiast server market.

    All along the watchtower the watchmen watch the eternal return.

  15. #40
    Xtreme Enthusiast
    Join Date
    Apr 2006
    Location
    Brasil
    Posts
    534
    K8L Quad-Core at "God only knows" clock speed 40% faster than 2.66GHz Clovertown in SPECfp? Sounds really good.

  16. #41
    Xtreme Enthusiast
    Join Date
    Apr 2006
    Location
    Brasil
    Posts
    534


    According to the image Core2 Quad is only 25% faster than Core2 Duo.
    And K8L Quad is 40% faster than Core2 Quad!

  17. #42
    Xtreme Enthusiast
    Join Date
    Oct 2006
    Posts
    896
    But if it's 'just' 40% better, that is a result of K8L, not the quadness, no?

  18. #43
    Xtreme Member
    Join Date
    Jul 2006
    Location
    Do
    Posts
    267
    I just noticed in the Graph.. that is says "Estimated".
    Excuse me for my English

  19. #44
    Xtreme Enthusiast
    Join Date
    Oct 2006
    Posts
    896
    Doom, notice that in OLTP, K8L is obviously not 70% better than X5355, it's 70% better than 2200SE. I would attribute the 40% to architectural changes rather than the doubling of cores. Yep, LCN, they're simulated benches

  20. #45
    Registered User
    Join Date
    Oct 2006
    Posts
    56
    LOL "estimated performance"

  21. #46
    Xtreme Member
    Join Date
    Oct 2005
    Posts
    462
    Only because clock speeds havent been nailed down yet.

  22. #47
    Xtreme Mentor
    Join Date
    Nov 2005
    Location
    Devon
    Posts
    3,437
    Charlie is panicking a little here!

    AMD demoed 4S platform with Barcelona which was running Windows! As most users know opening windows OS with faulty components is very hard on the first place! Then they showed Task Manager where CPUs (16 cores) are @100% utilization! Hmmm... when Task Manager shows things like this? YES! Something was eating CPU resources (computing)!!

    This demo showed that at least major bits of Barcelona core (HT, mem controller, ALUs, crossbar, etc) are working properly at this stage. They still can have problems with cache, we know that those parts of CPU can be switched off.

    BTW some informed people on this forum are hinting that AMD already have A2 silicon and that chartered is producing some Barcelona wafers as well.

    Cheers!
    RiG1: Ryzen 7 1700 @4.0GHz 1.39V, Asus X370 Prime, G.Skill RipJaws 2x8GB 3200MHz CL14 Samsung B-die, TuL Vega 56 Stock, Samsung SS805 100GB SLC SDD (OS Drive) + 512GB Evo 850 SSD (2nd OS Drive) + 3TB Seagate + 1TB Seagate, BeQuiet PowerZone 1000W

    RiG2: HTPC AMD A10-7850K APU, 2x8GB Kingstone HyperX 2400C12, AsRock FM2A88M Extreme4+, 128GB SSD + 640GB Samsung 7200, LG Blu-ray Recorder, Thermaltake BACH, Hiper 4M880 880W PSU

    SmartPhone Samsung Galaxy S7 EDGE
    XBONE paired with 55'' Samsung LED 3D TV

  23. #48
    Xtreme Enthusiast
    Join Date
    Apr 2006
    Location
    Brasil
    Posts
    534
    Yes and No, Clovertown and Kentsfield are limited by the shared FSB, Hyper Transport and the IMC allow K8 and K8L to run "free".
    But that wouldn't increase performance by 40% so it's a sum of K8L improvements, native "quadness" and the IMC.

    Note that Clovertown runs on 1333FSB and Kentsfield on 1066FSB.
    Opteron uses Registered Memory DDR2-667 (wich is a lot slower than regular unbuffered DDR2-800), so the gap beetwin a desktop K8L Quad-Core and a Core2 Quad might be even greatter than 40%.
    Last edited by doompc; 12-03-2006 at 01:04 PM.

  24. #49
    Xtreme Member
    Join Date
    Nov 2006
    Posts
    331
    Those estimated benches are always complete bs, plain marketing and propaganda - nothing usefull at all.

  25. #50
    Xtreme Addict
    Join Date
    Nov 2006
    Posts
    1,402
    Quote Originally Posted by Lightman
    Charlie is panicking a little here!

    AMD demoed 4S platform with Barcelona which was running Windows! As most users know opening windows OS with faulty components is very hard on the first place! Then they showed Task Manager where CPUs (16 cores) are @100% utilization! Hmmm... when Task Manager shows things like this? YES! Something was eating CPU resources (computing)!!

    This demo showed that at least major bits of Barcelona core (HT, mem controller, ALUs, crossbar, etc) are working properly at this stage. They still can have problems with cache, we know that those parts of CPU can be switched off.

    BTW some informed people on this forum are hinting that AMD already have A2 silicon and that chartered is producing some Barcelona wafers as well.


    Cheers!
    THx

Page 2 of 3 FirstFirst 123 LastLast

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •