The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.

    Thoughts on ATI's HD38x0?

    Discussion in 'Gaming (Software and Graphics Cards)' started by MozzUK, Nov 1, 2007.

  1. MozzUK

    MozzUK Notebook Guru

    Reputations:
    8
    Messages:
    60
    Likes Received:
    0
    Trophy Points:
    15
    Just thought it's interesting seeing information crop up about ATI's new RV600 based card coming out ( www.fudzilla.com). Even though this is just speculation, any views on what this might mean for notebooks?

    The 8800GT is 65nm while the HD38x0 will be 55nm and apparently much smaller in size and cheaper to produce. Also, the ATI's card is DX10.1 and should sell a lot cheaper than the 8800GT (even the 256 MB version).

    Seeing as it's power consumption could be significantly less than the 8800GT and it's cheaper overall, might this be a faster or better candidate for laptop gaming?

    It's been stated that it'll be at least 20% slower than the 8800 BUT in conversion to a laptop card wouldn't it scale better, requiring less modification to reduce power consumption and because it's already a small card, while the 8800 would need significant change to make it notebook compatible? Will the ATI retain more power and therefore game better?

    I don't know, maybe I've got it all wrong. Just thought I'd throw it out there. :confused:
     
  2. Jalf

    Jalf Comrade Santa

    Reputations:
    2,883
    Messages:
    3,468
    Likes Received:
    0
    Trophy Points:
    105
    Not sure where you get all this from. Any big rumors I've missed?
    55nm, yes, DX10.1, yes. But cheaper? Why?
    Power consumption lower? Why? (The other R600 cards have been far less power-efficient than the GF8's)

    I might have missed something here, but I don't see a reason why it should be cheaper (marketing reasons aside) or consume less power.
     
  3. MozzUK

    MozzUK Notebook Guru

    Reputations:
    8
    Messages:
    60
    Likes Received:
    0
    Trophy Points:
    15
    I got the impression it'll be cheaper and more power efficient judging by reports from fudzilla. No idea how reliable that i though.
     
  4. link1313

    link1313 Notebook Virtuoso

    Reputations:
    596
    Messages:
    3,470
    Likes Received:
    0
    Trophy Points:
    105
    It costs less to produce because it is smaller, also I believe it has less transistors than the R600.
     
  5. Jalf

    Jalf Comrade Santa

    Reputations:
    2,883
    Messages:
    3,468
    Likes Received:
    0
    Trophy Points:
    105
    Well, keep in mind that the process size doesn't necessarily mean lower power consumption or price.

    Price is determined by size * yield. If the yield is lousy, it doesn't matter that the chip is tiny. On the other hand, if you get excellent yields, you can achieve a low price even on a huge chip. I think it's obvious that NVidia has absolutely no problems with yield on the 8800GT, given the price and the broad availability.

    Also keep in mind that the 8800GT has been through a much bigger process shrink (90->65nm) than ATI's (65-55nm). And if ATI's cards weren't competetive at 65nm vs NVidia's 90nm cards, why should their 55nm do better against NVidia's 65nm?

    Anyway, guess we'll see. I'm not convinced though. But who knows?
     
  6. MozzUK

    MozzUK Notebook Guru

    Reputations:
    8
    Messages:
    60
    Likes Received:
    0
    Trophy Points:
    15
  7. MozzUK

    MozzUK Notebook Guru

    Reputations:
    8
    Messages:
    60
    Likes Received:
    0
    Trophy Points:
    15
  8. Lite

    Lite Notebook Deity

    Reputations:
    342
    Messages:
    1,398
    Likes Received:
    0
    Trophy Points:
    55
    also 55nm means cooler... and less power consumption
     
  9. Jalf

    Jalf Comrade Santa

    Reputations:
    2,883
    Messages:
    3,468
    Likes Received:
    0
    Trophy Points:
    105
    Like the Prescott? :)
    55nm means the *potential* for cooler running and less power consumption. Nothing guaranteed.
     
  10. narsnail

    narsnail Notebook Prophet

    Reputations:
    2,045
    Messages:
    4,461
    Likes Received:
    1
    Trophy Points:
    106
    yeah, they usually make it more powerful, but consume the same amount of power. its usually never lower..
     
  11. mujtaba

    mujtaba ZzzZzz Super Moderator

    Reputations:
    4,242
    Messages:
    3,088
    Likes Received:
    515
    Trophy Points:
    181
    Well, the Prescott was kind of a major change under the NetBurst architecture increasing the total pipeline changes to 36 along with the other changes.

    It's true that architecture upgrade and tweaks can increase the power consumption in cases like the X1900XTX which consumed slightly higher power consumption than the X1800 (both at 90nm), but it leads to lower power consumption in many cases too.

    But the die-shrink alone (in the same architecture) can be very effective and always leads to lower power consumption and a cooler chip (for example the Die-shrink of the CELL processor reduced the power consumption to 135 from 200watts and IIRC the power consumption of the 90nm CELL processor was 90watts)

    But the mobile 7800GTX at 110nm the power consumption was 66watts compared to the 45watts at 7900GTX, while the 7900GTX had a better performance too, so the R670 does not necessarily consume more power.

    Putting the more geekish parts aside , according to the leaked pictures, it has a single slot cooler compared to the dual-slot cooler on the R600, and also theInquirer claims that the power consumption of two RV670's in crossfire would be around a single R600.

    I think the time tells, both about things such as the power consumption and performance.