Announcement

Collapse
No announcement yet.

AMD Radeon HD 8670D: Gallium3D vs. Catalyst

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • smitty3268
    replied
    Originally posted by Kivada View Post
    We are discussion, all mid to high end desktop parts, power consumption isn't a big deal. Furthermore modern dedicated GPUs have very good idle power consumption.
    The Intel parts being discussed are mobile only, so claiming power usage doesn't matter is just stupid.

    Now, even if we are looking at mobile parts, the AMD A10-5745m and A10-4655m are both 25w parts. AMD always overstates their TDP requirements as they go for theoretical maximum.
    And those parts are both slower. Like i said, it's a trade off based on what is most important to you.

    Your point is moot.
    What point? That different people value different features?

    Leave a comment:


  • curaga
    replied
    Originally posted by Kivada View Post
    So now the kernel is holding back a setting you change in the BIOS? We are talking an APU here, you can control it's memory settings at time of boot, the same is not possible with a dedicated GPU.
    You cannot tell whether the speed is still in effect. The kernel may have changed it after the BIOS boot.

    Leave a comment:


  • liam
    replied
    Originally posted by Kivada View Post
    Well I'm always an asshole and it's a point I like to make quite often.




    The point I was making there is that the AMD APUs are bandwidth limited. I used the HD5670 as the closest analogue to HD8670D in terms of GPU core performance, comparing it's availible bandwidth to that of the HD8670D shows that while the GPU core of the HD8670D should in theory be capable of spitting out frames far faster then an HD5670 it's held back by being starved for bandwidth even with quite fast ram.

    Intel tries to squeeze by with a very expensive eDram chip on the GPU, allowing very low latency. If AMD had done the same it's fairly obvious that the 5200 would not be able to keep up based souly on the GPU core tech.

    I stand by my statement that for the cost of the 4950HQ/Iris Pro 5200 is going to be $657 for just that chip and nothing else. At least in the US you can build a nice non iGPU system with a GPU that blows the doors off of the 5200 for that much cash.
    This was a referenced link on the AT site: http://www.inpai.com.cn/doc/hard/182755.htm
    According to that you see increasingly diminishing returns as the memory clocks increase to the point where a 12% clock increase results in a 3% to 6% fps change for the A10-5800K, whereas going from ddr-1333 to 1600 followed bandwidth almost exactly.
    Assuming a similar core design, I would expect similar diminishing returns.
    As for cost, I'm not going to argue over much other than to say I'd expect the 4850HQ (http://www.anandtech.com/show/6993/i...50hq-tested/19) to perform very nearly the same as the $200 more 4950HQ since the gpu clocks are identical and the cpu clocks only differ by 100MHz. That's a much better proposition, but, as you say, you very well might be able to find a cheaper system but it's openCL seems quite a bit better than even the nvidia 640 (http://www.anandtech.com/show/6993/i...50hq-tested/17).

    EDIT:
    Looks like you already found the cheaper iris pro, and someone also mentioned the power issue that discrete cards can't ever match.
    Last edited by liam; 08 July 2013, 09:01 PM.

    Leave a comment:


  • Kivada
    replied
    Originally posted by curaga View Post
    The issue you were talking about was the reporting of memory speeds, which as Michael correctly pointed out is not possible without kernel changes. That's a level hugely above what any review site does, so hardly a strawman. No review site goes in and codes something that deep to make their review.

    Your complaints I agree with, but we were discussing the memory part, so I don't get your jump to everything else suddenly.
    So now the kernel is holding back a setting you change in the BIOS? We are talking an APU here, you can control it's memory settings at time of boot, the same is not possible with a dedicated GPU.

    Leave a comment:


  • Kivada
    replied
    Originally posted by smitty3268 View Post
    If you are willing to trade increased power usage, yes. That's a valid tradeoff that some people might want to make and others won't. I'm not sure i agree that only an idiot would buy the product, which i think is your assertion.
    We are discussion, all mid to high end desktop parts, power consumption isn't a big deal. Furthermore modern dedicated GPUs have very good idle power consumption.

    Now, even if we are looking at mobile parts, the AMD A10-5745m and A10-4655m are both 25w parts. AMD always overstates their TDP requirements as they go for theoretical maximum.

    Your point is moot.
    Last edited by Kivada; 08 July 2013, 04:52 PM.

    Leave a comment:


  • duby229
    replied
    Originally posted by curaga View Post
    The issue you were talking about was the reporting of memory speeds, which as Michael correctly pointed out is not possible without kernel changes. That's a level hugely above what any review site does, so hardly a strawman. No review site goes in and codes something that deep to make their review.

    Your complaints I agree with, but we were discussing the memory part, so I don't get your jump to everything else suddenly.
    I agree but this issue does bring to light an issue that needs resolved. The more people that know about it the more likely it is that someone with the knowledge and will to write such a code will step up and do it.

    Leave a comment:


  • curaga
    replied
    Originally posted by Kivada View Post
    Strawman much? They also don't undermine the OS they are running by intentionally using outdated versions of the drivers or OS unless they are being paid to quite laughably try to show a bias toward competing hardware in a comparison review.

    ..snip generic complaints...
    The issue you were talking about was the reporting of memory speeds, which as Michael correctly pointed out is not possible without kernel changes. That's a level hugely above what any review site does, so hardly a strawman. No review site goes in and codes something that deep to make their review.

    Your complaints I agree with, but we were discussing the memory part, so I don't get your jump to everything else suddenly.

    Leave a comment:


  • smitty3268
    replied
    Originally posted by Kivada View Post
    Clarification: 4850HQ @ $468, comments on purchasing power of that amount still stand. In a $468 budget for CPU and GPU you can get quite powerful hardware that makes the Iris Pro 5200 an extremely niche product priced far outside of its market space.
    If you are willing to trade increased power usage, yes. That's a valid tradeoff that some people might want to make and others won't. I'm not sure i agree that only an idiot would buy the product, which i think is your assertion.

    Leave a comment:


  • Kivada
    replied
    Originally posted by Kivada View Post
    I stand by my statement that for the cost of the 4950HQ/Iris Pro 5200 is going to be $657 for just that chip and nothing else. At least in the US you can build a nice non iGPU system with a GPU that blows the doors off of the 5200 for that much cash.
    Clarification: 4850HQ @ $468, comments on purchasing power of that amount still stand. In a $468 budget for CPU and GPU you can get quite powerful hardware that makes the Iris Pro 5200 an extremely niche product priced far outside of its market space.

    Leave a comment:


  • Kivada
    replied
    Originally posted by liam View Post
    I thought you were being an asshole with some of your previous posts (and maybe you were), but if the above is what you were trying to communicate earlier then I'm forced to change my mind and agree with you.
    Well I'm always an asshole and it's a point I like to make quite often.



    Originally posted by liam View Post
    I was only responding to the link you provided in the AT forums (IIRC), and I stand by it (though I may be misunderstanding your point).
    The point I was making there is that the AMD APUs are bandwidth limited. I used the HD5670 as the closest analogue to HD8670D in terms of GPU core performance, comparing it's availible bandwidth to that of the HD8670D shows that while the GPU core of the HD8670D should in theory be capable of spitting out frames far faster then an HD5670 it's held back by being starved for bandwidth even with quite fast ram.

    Intel tries to squeeze by with a very expensive eDram chip on the GPU, allowing very low latency. If AMD had done the same it's fairly obvious that the 5200 would not be able to keep up based souly on the GPU core tech.

    I stand by my statement that for the cost of the 4950HQ/Iris Pro 5200 is going to be $657 for just that chip and nothing else. At least in the US you can build a nice non iGPU system with a GPU that blows the doors off of the 5200 for that much cash.

    Leave a comment:

Working...
X