Geforce GT 130 Specs - now with 3D Benchmarks!

Discussion in 'iMac' started by rittchard, Mar 5, 2009.

  1. rittchard macrumors 6502

    Aug 12, 2007
    I was going crazy reading through all the threads and seeing people equate the GT 130 to this or that or to the other thing. The short answer is, I believe, it is not *exactly* like any of them, though as people mentioned, fairly close to the 9600GT, 9600GSO and/or 8800GS.

    I just installed Windows 7 Beta and the latest beta drivers from nVidia, here's what they say:

    GT 130
    64 stream processors
    529MHz core
    1323MHz shader
    792MHz memory speed/1584MHz data speed
    PCI-E x16 Gen 2 bus

    Comparing with the specs listed on Techarp, it looks to me to be closest to the 9600 GT overall, but with clock speeds slowed down. Also close to the 8800M GTS on the mobile side.

    But it is definitely *not* the GT 130M as some people were concerned about:

    - - -

    I'll be running 3DMark06 and 3D Vantage tests on it hopefully later tonight.
  2. rittchard thread starter macrumors 6502

    Aug 12, 2007
    This is interesting, I didn't see it yesterday, direct from nVidia website:

    GPU Engine Specs:
    Processor Cores 48
    Graphics Clock (MHz) 500 MHz
    Processor Clock (MHz) 1250 MHz
    Texture Fill Rate (billion/sec) 12

    Memory Specs:
    Memory Clock (MHz) 500 MHz
    Standard Memory Config 768 MB DDR2
    Memory Interface Width 192-bit
    Memory Bandwidth (GB/sec) 24

    - - -

    Now I am fairly confused as these are not the same things I saw this morning, particularly in terms of # of processor cores (48 vs 64) and the memory interface bandwidth (192 vs 256).

    I'm not sure how the computer determines what it shows you (the specs I listed in the original post were from the nVidia control panel, hitting "System detail" button), i.e. if it just spits out something generic from a database or if it's actually accessing the chip. Anyway, the plot thickens lol.
  3. avihappy macrumors 6502


    Nov 15, 2006
    Perfect! I think there was also another forum member who keeps asking for GPU-Z tests. Maybe that will help clear up some more things in THE GREAT IMAC GPU CONUNDRUM. This also makes me wonder now if the 4850 option is NOT the mobility version.
  4. Shivetya macrumors 65816

    Jan 16, 2008
    Really it will be interesting to see where "this" chip ends up. It seems that Apple not only gets special treatment from Intel but from Nvidia as well.
  5. rittchard thread starter macrumors 6502

    Aug 12, 2007
    Got a little extra time in the afternoon so without further ado:

    3DMark06 (1280x1024 - all default settings):

    9645 3DMarks
    4303 SM2
    3960 HDR
    2804 CPU

    This is in Windows 7 using the 181.71 drivers, 3.06GHz CPU.

    Very respectable!! I was actually expecting something closer to the 7000 range. For reference I had an older PC system with a 2.4GHz Core 2 Duo and an 8800GTX 768MB which scored 9690 3DMarks.

    GPU-Z was a bust, the card registered as a 9600GT Pre-release and most of the information was either blank or looked wrong.

    Will run Vantage later tonight.
  6. avihappy macrumors 6502


    Nov 15, 2006
    That is very respectable! Now I'm wondering, if it's not the mobile version, how Apple fit a card requiring 350W minimum into a iMac that draws only 280W maximum...
  7. EmperorDarius macrumors 6502a

    Jan 2, 2009
    And according to this site:

    Mobility Radeon HD 4850 3DMark06 = 9577
  8. gmanrique macrumors 6502

    Oct 22, 2003
    Ottawa, Canada
  9. GSV macrumors member

    Jun 25, 2007
    Does the discrepancy between rittchard's 3DMark06 results and those of mean that the iMac's card isn't the mobile version? Or did they use different testing parameters? I couldn't find on their site any more detail on that matter.
  10. EmperorDarius macrumors 6502a

    Jan 2, 2009
    Remember that we're not sure if it's the mobility version. It could be an underclocked desktop version, and I don't know how it would score. We'll have to wait for someone to get and test the radeon for an answer.

    Myself, I'll just get the gt 130.

    If you see at the test you'll see that the GT 130 got results close to the GeForce 9800M GTS (9692). Perhaps it's an underclocked 9800M GTS?
  11. Moriarty macrumors 6502

    Feb 3, 2008
    That would make sense. Quell people's fears that the old iMac's 8800 GS was better. The iMac's GT 130 is looking to be a very respectable card. Should play most games coming out for another few years yet, but I suspect that it would struggle to play most post-2007 games at 1920*1200 and decent detail settings.

    Because that 350W minimum takes account of all the computer's components (CPU, hard drive, etc). Granted the iMac's 280W includes display also, it uses a laptop CPU which is much less power hungry (hence a bit less powerful). Essentially, everything in the iMac is more power efficient than the average PC that the 350W measurement is made for.
  12. rittchard thread starter macrumors 6502

    Aug 12, 2007
    At this point, yes, I'm reasonably certain this is NOT the GT 130M. It is registering as a "prerelease" 9600GT. I think it's essentially an underclocked 9600 GT, *possibly* with the 55nm process(I thought I read someone speculating that somewhere?) - that might explain the power consumption discrepancy. Or maybe just the underclocking is enough to draw less power. The iMac doesn't seem to be heating up (during 3D Vantage) that much and I didn't hear any fans coming on at all, it's fairly amazing.

    Also keep in mind when you compare the numbers the CPU portion will influence what you see - I believe this is a very nice CPU (E8435) so the scores will reflect that. 3DMark06 in particular seems to favor CPU considerably. However in the Vantage results...

    GPU 2058
    CPU 17892
    3DMark P2643
  13. chewietobbacca macrumors 6502

    Jun 18, 2007
    If the ATI 4850 is non-mobility, it will be much much faster still (25-40%). If it is mobility, it's probably 15-20%.

    FWIW though, 3dMark06 scores CANNOT be compared across different computers, so going off that website isn't enough. This is because 3dMark06 defaults to 1280 x 1024 resolution, which is VERY CPU bound. For instance, I can put a 4870X2 on a Core 2 Duo clocked at 1.86 GHz and get a ridiculously low score of 10k on my PC. Or, if I put it into my QX9650 OC'd to 4.0GHz quad core desktop monster, I score 22k on the same settings. See how a CPU change makes a HUGE difference?

    Like in rittchard's example with the 8800GTX - my 8800GTX would score ~13-15k on 3dMark06 with a 3.2GHz Core 2 Quad. If he put the 8800GTX in the same CPU as the iMac, it'd probably do better than the 96xx he had with the slower Core 2.

    Vantage is a beter indicator because it is less CPU bound at the higher settings. A Vantage score of P2643 is interesting. A desktop 4850, however, will score P6600~7000ish so take that FWIW.

    Also, I'd like to note that GPU-Z is a database that doesn't detect what the actual card is, it just uses what the software-writer knows and puts it in there. It registering as a 9600GT definitely suggests it is a 9600GT, but it could also be that the GPU-Z isn't identifying the car correctly.

    An example of this was before the RV770's were released. GPU-Z pics said 480 stream processors, but we all know now that they ended up with 800 :D

    The power minimum is just what the GPU maker recommends so people don't fry their PC and sue them for not having given limits. It isn't the actual iMac components being more efficient - they use the same components as any other PC now. It's that the actual power draw at the wall that computers use are grossly overrated.

    For example, a 4870X2 might say you require a 700 or 800 watt PSU. In truth, my desktop enthusiast PC, despite having twin water cooling loops, a 4870X2 overclocked, a QX9650 Quad OC'd to 4GHz, 3 hard drives, etc. only draws 450 watts total at load from the wall, using a Kill-A-Watt.

    Add in the fact that I'm at 80-85% efficiency with the power supply, and my system (the components themselves) as a whole is using around 360W at most. Makes that 1000W PSU I have seem a little ridiculous, but it's a safety thing (because you don't want to draw at 80% all the time at load, it's safer to keep it at 50% for longevity).
  14. Hellhammer Moderator


    Staff Member

    Dec 10, 2008
    After reading these posts, I'm really happy that I chose GT130 instead of 4850
  15. Draic-Kin macrumors newbie

    Mar 5, 2009
    I mentioned it in another thread allready.

    GT 130 = 9600 GSO, but NOT 8800 GS

    8800 GS = 8800M GTS !!!

    The comparison/benchmark above is ********.

    The 4850 is much faster then the GT 130!!!

    And you wont be happy at all, because even the "old" 8800GS is faster then the GT 130... ;)

    If you don´t believe me search for "9600 GSO" instead of "GT 130":

    "9600 GSO" vs. "HD 4850"


    "8800M GTS" vs. "9600 GSO"
  16. Shivetya macrumors 65816

    Jan 16, 2008
    Provide benchmarks on the same machine or go home.
  17. Draic-Kin macrumors newbie

    Mar 5, 2009
    Search for yourself, i think even you should be able to use google.

    Knowing that a GT 130 is a renamed 9600 GSO, how can you even think it is as fast, or even faster then a HD 4850... ;)

    The GT 130 is crap, get the "old" iMac with 8800 GS aka 8800M GTS or get the "new" one with HD 4850.

    This is all i will say, believe me, use google to check it, or dont... i dont care.
  18. Shivetya macrumors 65816

    Jan 16, 2008
    Do you know the exact part number/model used in the new iMacs?

    No you don't, so quit talking out the wrong end.
  19. Mike in Kansas macrumors 6502a

    Mike in Kansas

    Sep 2, 2008
    Metro Kansas City
    I wouldn't believe any benchmarks on the web unless they were done on iMac's with the various cards installed in them. Apple has been known to "mis-label" cards on their specs, like labeling an under-clocked HD Mobility 2600XT as an HD 2600 PRO. Until someone like Barefeats runs these various cards all on 2.93's (or a group of users do the same) I wouldn't believe anything.
  20. maxap macrumors 6502


    Sep 20, 2008
    Hmm, I guess the G130 seems to be a faster beast than everyone thought it would be... This is a good news. Maybe I will buy a new iMac with the G130 now instead of wating for the 4850 to arrive...
  21. FoxHoundADAM macrumors regular

    Sep 9, 2008
    Humm, interesting. this certainly throws up another question mark in my purchasing decision.

    I think in the end the 4850 will still be the faster option and for a mear $45 more I'll have to choose it. I would like to see more benchmarks though.
  22. Moblin88 macrumors newbie

    Mar 6, 2009
    That is simply not true. The 9600 GSO has a 192 bit memory interface. The card in the imacs has a 256 bit memory interface. Changing the clock speed does not change the memory interface. This card is NOT the 9600 GSO.

    The specs look almost exactly like a 9800M GS, for what it's worth. Same number of stream processors, same memory interface, same core and shader clock speeds, same memory speeds.

    It also commonly comes with 512 meg memory. Such as in the ASUS G50Vt


    Also, if you take a look at the link above, you will see that this card benchmarks slightly BETTER than the 8800M GTS, which is what apple calls the "8800 GS" in their 24" early 2008 iMacs. So the report that the new card gave better WoW performance that old top-end card is likely accurate. That being said, the cards are comparable.

    Also, for my money, the 4850 will be the MOBILITY version of the card, which you can also see benchmarks for at the link above. Apple has put mobile gpus, cpus, and memory in all the iMacs in recent years since they all have laptop chipsets. Since the new early 2009 iMacs seem to have mobile chipsets since they have mobile ram and mobile cpus and mobile gpus, It is almost certain that the high end card will also be the mobility versions of the card.

    Note though that per the benchmarks, this is likely the BEST card apple could have put in the iMac considering that due to space and power and heat requirements, they must have mobile chipsets and can't have 2 cards in SLI, Crossfire, or simply mushed together into one card.
  23. EmperorDarius macrumors 6502a

    Jan 2, 2009
    If that's the case, it's better than the 4850 according to :

    9800M GS: 13928
    ATI Mobility Radeon HD 4850: 12318

    3DMark 05

    That is, IF the Radeon is the mobility version.
  24. synagence macrumors 6502a

    Jul 23, 2008
    If the 130GT has been demonstrated to be non-mobility then the ATI option simply can't be mobility otherwise the performance would be much slower....

Share This Page