Page 1 of 2 12 LastLast
Results 1 to 10 of 13

Thread: Sapphire Radeon HD 4650 512MB OC

  1. #1
    Join Date
    Jan 2007
    Posts
    14,816

    Default Sapphire Radeon HD 4650 512MB OC

    Phoronix: Sapphire Radeon HD 4650 512MB OC

    We previously had looked at the ATI Radeon HD 4550 and Radeon HD 4670, but if you are looking for a graphics card that's positioned between the two and costs less, there is the Radeon HD 4650. The Radeon HD 4650 is clocked the same as the Radeon HD 4550, but is based upon the RV730PRO GPU and is able to provide a bit more processing power than the lesser RV710 solution. Sapphire though manufactures a Radeon HD 4650 graphics card that operates well beyond the reference core and memory frequencies for the RV730PRO and sells it at a very affordable price. In this article we are seeing how well the Sapphire Radeon HD 4650 512MB OC graphics card can perform under Linux.

    http://www.phoronix.com/vr.php?view=13189

  2. #2
    Join Date
    Feb 2008
    Location
    Linuxland
    Posts
    5,105

    Default

    Could that Phoronix approved pic be converted to png with an alpha channel, to get it looking good (now the gif has jagged edges, due to limited transparency)?

    The only stopper is IE6, and a Linux site should not give a damn about IE6 users

  3. #3
    Join Date
    Oct 2006
    Location
    Israel
    Posts
    597

    Default

    The article doesn't include the card's retail price.
    (And I'm not talking about the relative price - I'm talking about absolute price.)

    Was it intended?

    - Gilboa
    DEV: Intel S2600C0, 2xE52658V2, 32GB, 4x2TB, GTX780, F20/x86_64, Dell U2711.
    SRV: Intel S5520SC, 2xX5680, 36GB, 4x2TB, GTX550, F20/x86_64, Dell U2412..
    BACK: Tyan Tempest i5400XT, 2xE5335, 8GB, 3x1.5TB, 9800GTX, F20/x86-64.
    LAP: ASUS N56VJ, i7-3630QM, 16GB, 1TB, 635M, F20/x86_64.

  4. #4
    Join Date
    Jun 2008
    Posts
    162

    Default

    Thanks for the tests.
    Any reason ETQW or Unigine weren't tested? It desperately lacks an intensive 3D engine test (or X-Plane, or anything else), and maybe one of the new wine tests in PTS (maybe the writer doesn't now about those ).

  5. #5
    Join Date
    Aug 2007
    Posts
    437

    Default

    Looking at the performance comparsion between 4650 and 9600GT makes me wanna cry.

    4650 is less than HALF the performance of 9600GT... maybe they should sell it for 50% price of a 9600GT also so it's gonna be a urban bargain

  6. #6
    Join Date
    Feb 2008
    Posts
    28

    Default

    I'd really like to some quantitative comparisons of the noise made by graphics cards. My GeForce 6800GT cranks up the fan noise during intense sessions and I'd really appreciate knowing that card X is louded than card Y under load.

  7. #7
    Join Date
    May 2007
    Location
    Third Rock from the Sun
    Posts
    6,584

    Default

    Quote Originally Posted by Nexus6 View Post
    I'd really like to some quantitative comparisons of the noise made by graphics cards. My GeForce 6800GT cranks up the fan noise during intense sessions and I'd really appreciate knowing that card X is louded than card Y under load.
    If you want quiet cards the simplest thing to do is look at cards with non-reference coolers. As a rule of thumb they tend to be quieter.

  8. #8
    Join Date
    Aug 2007
    Posts
    437

    Default

    If you want a *REALLY* quiet card I suggest you buy a passively cooled card, like my EAX1600XT Silent.

    Of course you have to pay a premium for that, though.

  9. #9
    Join Date
    Dec 2008
    Posts
    3

    Default

    I don't get it at all.
    We have two computer running Ubuntu Intrepid. I saw the test on the internet so I bought a sapphire 4670. I played some urt, openarena and serious sam then I noticed the games were unplayable if the flares option was true. And I didn't get the max fps in urt and serioussam was nearly unplayable so I thought further investigation will be necessary. There you are phoronix-test-suite 1.4. I ran some test :

    ####################################
    OpenArena:
    1680 x 1050

    150.6 Frames Per Second
    144.0 Frames Per Second
    143.5 Frames Per Second

    Average: 146.03 Frames Per Second
    ####################################

    ####################################
    Urban Terror:
    1680 x 1050

    59.5 Frames Per Second
    61.0 Frames Per Second

    Average: 60.25 Frames Per Second
    ####################################

    fglrx 8.12 + mesa 7.2
    ======================
    ####################################
    OpenArena:
    1680 x 1050

    147.4 Frames Per Second
    144.7 Frames Per Second
    141.2 Frames Per Second

    Average: 144.43 Frames Per Second
    ####################################

    ####################################
    Urban Terror:
    1680 x 1050

    58.9 Frames Per Second
    60.6 Frames Per Second

    Average: 59.75 Frames Per Second
    ####################################

    The newest Catalyst (8.12) and the same goes for mesa (7.2) were slower than before. The other PC has an old and cheap 7300gt. It gets nearly as much point as my brand new 4670. I think you already noticed that in the phoronix test the 4670 had much better performance than what I got.

    Some glxgears:
    7300gt: 12072
    4670: 7380

    If someone has any idea it would be welcome.

  10. #10
    Join Date
    Oct 2007
    Location
    Toronto-ish
    Posts
    7,458

    Default

    The glxgears benchmark tests only the most basic rendering functions; coloured triangles with no textures and no shaders. As a result the numbers end up being misleading in a couple of ways :

    1. Older cards often score higher than newer cards because modern cards spend much more of their transistor budget on shaders and texture engines rather than ROPs. It's not unusual for an older card to outrun a newer card in the same range; usually the older card has more ROP throughput than the new card.

    2. In many cases the benchmark runs CPU limited rather than GPU limited; in that case you end up measuring is the driver efficiency on a trivial path which is mostly used by glxgears and screen savers.

    In the case of 7300gt vs 4670 both cards have 8 ROPs, memory width/type is the same and the clock rates aren't that different, so their theoretical performance on glxgears should be relatively similar. Shader power is a diferent story; the 4670 has somewhere between 5x and 8x the shader power of the 7300gt depending on how you do the math. Something like the Unigine demos (in the Phoronix test kit IIRC) would probably show off that difference.

    By the way you mentioned "fglrx 8.12 and mesa 7.2" which I didn't understand; fglrx doesn't use mesa for 3d, it has its own 3d driver stack.
    Last edited by bridgman; 12-13-2008 at 10:17 PM.

Posting Permissions

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