1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

AMD FX Bulldozer CPU Against Intel Core i7-990X

Discussion in 'Frontpage news' started by Guru3D News, Jul 4, 2011.

  1. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    I'll help derail this topic!

    AMD(AMD GPU that is) as of now are all in the Next Gen Consoles

    Wii U(wtf, nintendo, wtf)
    Play$tation 4
    Xbox Whatchamacalllit

    Good riddance to the Green, Hello the Red

    Communism!!!! (Oh wait not that red)
     
    Last edited: Jul 8, 2011
  2. IPlayNaked

    IPlayNaked Banned

    Messages:
    6,559
    Likes Received:
    0
    GPU:
    XFire 7950 1200/1850
    I disagree 100%. The only thing that's kept me an AMD customer is ease of upgrade. There's no reason AMD can't remain competitive and retain some form of compatibility.

    They know they have no ability to compete high-end, so they have to offer other incentives to stay. Price is one, but barely, and compatibility is another.

    It is a fact that a great many people would not own Phenoms if they weren't upgrading their computer incrementally and that was just the easiest way. I want a Sandy Bridge right now, I so want to just buy it, but it's a big investment, it's a motherboard and processor, at least. Never have I had an AMD build where I had to pay all that out at once.
     
  3. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    Last edited: Jul 9, 2011
  4. PhazeDelta1

    PhazeDelta1 Moderator Staff Member

    Messages:
    15,616
    Likes Received:
    13
    GPU:
    EVGA 1080 FTW
    that cpu-z screen is a fake. bd didnt get cpu-z support until v1.58 and if you look closely, its clearly using v1.57.1.
     

  5. sykozis

    sykozis Ancient Guru

    Messages:
    20,988
    Likes Received:
    636
    GPU:
    MSI RX5700
    When the K7 Athlon launched, I went from a Celeron@433mhz to a Slot A 500mhz Athlon, to a Socket A Thunderbird/Barton to a Socket 754 Athlon64 to an AM2 Athlon64 X2. Every upgrade required a new board. Backwards compatibility was not even a consideration when the K7 architecture first launched. K7 got a brand new "socket" (Slot A)...the first major update to the architecture brought another new socket (Socket A). Backwards compatibility within a processor/architecture family is great....but when an entirely new architecture is being designed, it should not be designed based on the primary goal of backwards compatibility. If your goal is to maintain backwards compatibility...you'll never advance. Just imagine where the automotive industry would be....if backwards compatibility was the primary goal each year... We'd be driving cars that were designed back in the late 1800's....
     
  6. deltatux

    deltatux Ancient Guru

    Messages:
    19,051
    Likes Received:
    13
    GPU:
    GIGABYTE Radeon R9 280
    For a consumer, using the same board is great, backwards compatibility is great. What you said there is why the x86 architecture is kinda crappy ... it has to tow along so many backwards compatibility that the architecture family is so bloated. Frankly, I'm no fan of the x86, I like IBM's POWER architecture a lot more and same with ARM, they're a lot more elegant in their designs than x86.

    Anyways, that's besides the point. For the 6 or so years, AMD has been able to maintain backwards compatibility which is pretty amazing people could incrementally upgrade their rigs. AM2 to AM2+, AM3 to AM3+ (but not backwards). AM3 in AM2+ boards. This is quite amazing for AMD to be able to do this while Intel seems to be replacing their sockets every year or 2 years which gets a bit annoying.

    deltatux
     
  7. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    CPU-Z support has been there since 1.56.4 Beta

    It's not fake

    Code:
    CPU-Z version 1.56.4
    
    Processors
    -------------------------------------------------------------------------
    
    Number of processors 1
    Number of threads 8
    
    APICs
    -------------------------------------------------------------------------
    
    Processor 0
    -- Core 0
    -- Thread 0 0
    -- Core 1
    -- Thread 0 1
    -- Core 2
    -- Thread 0 2
    -- Core 3
    -- Thread 0 3
    -- Core 4
    -- Thread 0 4
    -- Core 5
    -- Thread 0 5
    -- Core 6
    -- Thread 0 6
    -- Core 7
    -- Thread 0 7
    
    Processors Information
    -------------------------------------------------------------------------
    
    Processor 1 ID = 0
    Number of cores 8 (max 8)
    Number of threads 8 (max 8)
    Name AMD Processor
    Codename Bulldozer
    Specification AMD Eng Sample, 1D26246W8K44_36/26/22_2/8 (Engineering Sample)
    Package Socket AM3+ (942)
    CPUID F.1.0
    Extended CPUID 15.1
    Core Stepping
    Technology 32 nm
    TDP Limit 149 Watts
    Core Speed 1400.0 MHz
    Multiplier x FSB 7.0 x 200.0 MHz
    Instructions sets MMX (+), SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, SSE4A, x86-64, AMD-V, AES, AVX, XOP
    L1 Data cache 8 x 16 KBytes, 4-way set associative, 64-byte line size
    L1 Instruction cache 4 x 64 KBytes, 2-way set associative, 64-byte line size
    L2 cache 4 x 2048 KBytes, 16-way set associative, 64-byte line size
    L3 cache 8 MBytes, 64-way set associative, 64-byte line size
    FID/VID Control yes
    Min FID 7.0x
    P-State FID 0x14 - VID 0x0B - IDD 12 (18.00x - 1.412 V)
    P-State FID 0xE - VID 0x0E - IDD 10 (15.00x - 1.375 V)
    P-State FID 0xA - VID 0x16 - IDD 10 (13.00x - 1.275 V)
    P-State FID 0x7 - VID 0x1B - IDD 9 (11.50x - 1.212 V)
    P-State FID 0x4 - VID 0x21 - IDD 8 (10.00x - 1.137 V)
    P-State FID 0x1 - VID 0x26 - IDD 6 (8.50x - 1.075 V)
    P-State FID 0x10C - VID 0x30 - IDD 6 (7.00x - 0.950 V)
    
    Package Type 0x1
    Model 00
    String 1 0x0
    String 2 0x0
    Page 0x0
    TDC Limit 96 Amps
    Attached device PCI device at bus 0, device 24, function 0
    Attached device PCI device at bus 0, device 24, function 1
    Attached device PCI device at bus 0, device 24, function 2
    Attached device PCI device at bus 0, device 24, function 3
    Attached device PCI device at bus 0, device 24, function 4
    Attached device PCI device at bus 0, device 24, function 5 
     
  8. deltatux

    deltatux Ancient Guru

    Messages:
    19,051
    Likes Received:
    13
    GPU:
    GIGABYTE Radeon R9 280
    CPU-Z didn't properly identified Bulldozer until 1.58 which is the issue as it properly identified the Bulldozer in 1.57.1.

    deltatux
     
  9. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    Bulldozer has been support since 1.56.4
    1.56.4 called AMD Processor (A1, B0)
    1.57.0-1 added the names(2nd batch of B0, B1)
    1.58 supports retail models(B2)

    http://www.hardware.fr/medias/photos_news/00/32/IMG0032131.png
    I posted the data for this screenshot above in the
    Code:
    
    1.56.4 became 1.57
    February 18, 2011
    
    seronxolitus
     
    Last edited: Jul 9, 2011
  10. sykozis

    sykozis Ancient Guru

    Messages:
    20,988
    Likes Received:
    636
    GPU:
    MSI RX5700
    Why does every mention Intel changing sockets every year or 2? Intel released LGA775 in Dec 2004. It was finally replaced in 2008. That's almost 4 years. Intel released LGA1366 in 2008. It's now 2011....that's 3 years. The average life-span for a computer is 3-5 years. Based on the fact that the "average user" will simply run down to BestBuy, Walmart, etc and buy a prebuilt system, changing sockets every 3-5 years doesn't directly impact the "average user"... BTW...the "average user" greatly outnumbers those of us building custom systems.

    Like I said....designing a processor with the primary goal being backwards compatibility, prevents advancements....

     

  11. TheHunter

    TheHunter Banned

    Messages:
    13,408
    Likes Received:
    1
    GPU:
    MSi N570GTX TFIII [OC|PE]


    yea it may have lasted for 4 years but chipsets weren't.. and you can't use newer cpu on some old chipset even if its LGA775..

    P865?, P915, x925, P945,.. x38..
     
  12. deltatux

    deltatux Ancient Guru

    Messages:
    19,051
    Likes Received:
    13
    GPU:
    GIGABYTE Radeon R9 280
    Fine, it's 3 years, but that's still quite fast. Note that that "average" isn't really much of an average for the masses, that's the average for us tbh. I see many rigs for non-geeks and non-gamers are around 5 - 7 years with them still wanting to upgrade them even when I advised against it. Also, with so many sockets, it may confuse consumers who don't want to go OEM but may want to quote the parts they want and have someone else build it for them. I do this a lot for friends and family, and then they misquote things, and then I have to fix the mistakes and sometimes spend more than they want (sometimes can save money depending on what "oopsies" they did).

    Designing a processor with the primary goal of being backwards is exactly why the x86 microarchitecture family is successful. Every single revision is backwards compatible to the original 8086. That's the whole point of the microarchitecture, that's why it's so bulky lol. This is because of corporate pressures on Intel to do so. This is the same with Microsoft with Windows. Many corporate users are reliant on very old software which would cost thousands or tens of thousands (depending on size) to redevelop, validate, and train their entire staff to use the new software. This is why both the x86 and Windows retain so much backwards compatibility. It's good business but is a strain on innovation.

    Other RISC architecture either forks off or not completely compatible with each other which is why many of these architectures are so forward thinking and pretty revolutionary but we don't see this in x86 because of backwards compatibility. However, due to such breaks in backwards compatibility, causes headaches for corporate consumers. POWER is an example of it. When Apple moved from PowerPC G4 to G5, they had an issue where IBM designed the G5 (technical name: PowerPC 970 family), they used big endianess instead of the little endianess that Motorola kept using in the G4. This caused a lot of headaches for big companies that they were reluctant to replace their Power Mac G4s to the Power Mac G5 because Apple had to do little to big endianess translation on the fly which added performance penalties. That and the fact that the G5 sucked a lot more power than the G4 made it undesirable for Apple and the corporate consumers from continually use the G5 chips.

    Something like this is undesirable for large corporations, this is why backwards compatibility is so important to the success of hardware and software in the corporate space and this is the bulk of the business for both Microsoft and Intel.

    deltatux
     
    Last edited: Jul 9, 2011
  13. sykozis

    sykozis Ancient Guru

    Messages:
    20,988
    Likes Received:
    636
    GPU:
    MSI RX5700
    It's impossible to design a single chipset to fully support nearly as wide a range of products as LGA775 did. I think my total count of LGA775 chipsets was around 16 or so....and I'm sure some were missing. But I was counting full chipsets, not just the northbridges. Most Northbridges had at least 2-3 complete chipset variants.

    I fail to see where having 2 sockets for Intel and 2 for AMD is confusing.

    I've been quoting systems for the last 13 years and have never had an issue of confusion between processors/sockets. Of course, my clients tend to either do the research first or just allow me to configure what they need based on their usage.

    X86 isn't an architecture...it's an instruction set. Conroe was a micro-architecture.... Nehalem is a micro-architecture. Allendale, Kentsfield, Yorkfield, Wolfdale...are architectures based on the Conroe micro-architecture. Lynnfield, Clarkdale, Bloomfield are architectures based on the Nehalem micro-architecture. For Conroe or Nehalem to be backwards compatible with the 8086 processor....they'd have to share a common design, which they don't. 8086, is comparison to Conroe or Nehalem, was a very simply design. Conroe and Nehalem are extremely complex designs with no relation to the 8086 aside from the x86 instruction set.

    With AMD, they seem to start with a socket and build a processor from there, which leaves no room for any real innovation. They're bound by the limitations of the socket design. If the socket design only permits for a dual channel memory controller....that's all the processor can have. Yeah, backwards compatibility between 2 processor architectures is great for consumers financially....but from the point of innovation, it's far too limiting/restrictive.
     
  14. deltatux

    deltatux Ancient Guru

    Messages:
    19,051
    Likes Received:
    13
    GPU:
    GIGABYTE Radeon R9 280
    Well, apparently they get confused, I also question how it's confusing, but they still manage to screw it up.

    That's why I said microarchitecture family, I didn't say just microarchitecture, I know each architecture is different. However, since you have to follow the same instruction set, you'll still have to build your CPU around it which is the problem. The x86 is bulky and both AMD and Intel has to do it just for the sake of backwards compatibility. Like I said, this is the same for Microsoft with Windows. The only reason why Windows is so successful is because of backwards compatibility. There's still code from Windows 3.1 in Windows 7.

    Both AMD and Intel have been incrementally updating their architectures and there hasn't really been any major breakthroughs that hasn't been found on other microarchitecture families like SPARC, POWER and etc. What AMD is introducing is interesting though for Bulldozer with CMP. Personally, I have never seen it before but who's to say other companies haven't already tried it yet.

    As for dual channel, the number of channels doesn't really matter as long as you can create an efficient memory controller as Intel has shown with Sandy Bridge. The controller on Sandy Bridge is much better than Nehalem, even its dual channel beats out Nehalem's triple channel. So if AMD can design a very good memory controller (I hope they do, but wouldn't be surprised if they can't), they wouldn't need to alter the socket to accommodate the new memory controller since the controller is built into the CPU.

    deltatux
     
  15. TheHunter

    TheHunter Banned

    Messages:
    13,408
    Likes Received:
    1
    GPU:
    MSi N570GTX TFIII [OC|PE]
    ^
    well one leak showed poor mem bandwidth.

    [​IMG]

    read: ~14.5k
    write: ~8.5k
    copy: ~10-11k

    I hope its all fake though, otherwise its kinda disappointing:(
     

  16. thatguy91

    thatguy91 Ancient Guru

    Messages:
    6,648
    Likes Received:
    97
    GPU:
    XFX RX 480 RS 4 GB
    These benchmarks use engineering samples, and there's no indication of the actual bios settings used. For all we know, they could have the RAM set at 8-8-8-22, but have every other RAM setting set at its lowest value. As you should know, there are a lot of other timing options for DDR3 (even for DDR2 on older systems etc). Then of course there's the other bios settings which can all add up too poor performance if detuned!

    Until the first reviews come out using retail CPU's, any benchmark could be skewed to show what they want you to see. Even then, the bias of the reviewer should be considered for potential skewness.

    I believe the accepted practice is to leave the bios settings as the default, but this doesn't necessarily mean its always done.
     
  17. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    [​IMG]

    [​IMG]

    [​IMG]

    1.56.4(also 1.57) Beta support Bulldozer it called it "AMD Processor"
    1.57.1 added names support
    1.58 official support of the retail chips

    [​IMG]
     
    Last edited: Jul 10, 2011
  18. TheHunter

    TheHunter Banned

    Messages:
    13,408
    Likes Received:
    1
    GPU:
    MSi N570GTX TFIII [OC|PE]
    ^

    still to slow.. Take a look at SB mem bandwidth in dual channel mode.



    Anyway i hope for the best.. but this low mem bandwidth isn't a very good sign.
     
  19. seronx

    seronx Master Guru

    Messages:
    261
    Likes Received:
    0
    GPU:
    Wani R7 XT
    Engineer Sample but oh well I guess people don't understand ENGINEER Sample lol

    Also, in the aida test 2 modules turn off crippling it to half of the resources being able to be used

    that why in the aida it shows 4.2GHz for the multiplier

    [​IMG]

    vs
    [​IMG]
     
    Last edited: Jul 10, 2011
  20. TheHunter

    TheHunter Banned

    Messages:
    13,408
    Likes Received:
    1
    GPU:
    MSi N570GTX TFIII [OC|PE]
    and that AMD is OC'ed..thus higher mem bandwidth and yes i know what an Eng.sample is.:nerd:
     

Share This Page