Is compromising on dedicated video memory right while upgrading to a higher series of a graphics card?

Osama Javaid September 15, 2012

Is it a good decision to upgrade from a 4GB DDR3 GT series card to a 1 GB GDDR5 GTX series card ? I mean are there any cons?

  1. Dalsan
    September 18, 2012 at 5:13 am

    The obvious con would be lower memory, not that it would matter that much unless you actually use programs that can utilize more than 2GB, but you would gain better compatibility and performance with newer games and other software that are GPU capable. The comparison would be like having a 2.0GHz dual core Pentium (Core Duo) with 8GB of ram versus a 2.0GHz Core i3 with 4GB ram. In this case, most times the lesser amount of ram doesn't equate to worse performance unless you use programs that can utilize the extra ram, but still, performance drops are less noticeable than if it were the same processors with less ram than the other. In other words, more horsepower and torque with less storage capacity will most times beat less horsepower and torque with larger storage capacity. Only suggestion would be to try to get a GT model over the GTX model.

  2. Paul Girardin
    September 16, 2012 at 1:38 pm

    From here, it does not look like a compromise!

    Of course, your mileage may vary depending on the card's specific details (Vendor, quality, ETC ...)

    GDDR5 can for all purpose be considered at least twice as fast than GDDR3.

    GDDR3 Bandwidth: up to 9.6 Gbytes/s with a 1GHz clock.

    GDDR5 operates at up to 5Gbps at 1.5 volts.

    GDDR5 operates with two different clock types. A differential command clock (CK) to where address and command inputs are referenced, and a forwarded differential write clock (WCK) where read and write data are referenced to.

    Being more precise, the GDDR5 SGRAM uses two write clocks, each of them assigned to two bytes.

    The WCK runs at twice the CK frequency.

    Taking a GDDR5 with 5 Gbps data rate per pin as an example, the CK clock runs with 1.25 GHz and WCK with 2.5 GHz.

    The CK and WCK clocks will be aligned during the initialization and training sequence.

    This alignment allows read and write access with minimum latency.

    See here for details:

    And here to start exploring further on the subject:

    Hope this helps

    • Osama Javaid
      September 16, 2012 at 6:33 pm

      Too much over-techyy answer, can't understand

      • Paul Girardin
        September 16, 2012 at 9:15 pm

        Simple answer: Less dedicated video memory on a better GDDR5 type card is not a compromise, it's an improvement!

        You need less memory to do the job twice as fast than GDDR3!

  3. ha14
    September 15, 2012 at 7:14 am

    GB GDDR5 GTX should be better than 4GB DDR3 GT, sometime can be little difference, but then testing will be needed.

    GT 650M GDDR5 vs DDR3 and 2GB vs 1GB memory capacity

  4. Mike
    September 15, 2012 at 6:20 am

    In general anything between 1-2GB is fine. The only case where you'd benefit from more dedicated memory is probably some GPU based 3D rendering (3DSM, Blender) or something like that.

    However, you will have to be more specific about the models if you want a better answer. Without knowing the cards and making an actual comparison I wouldn't recommend a switch from GT to GTX.

Ads by Google