If this question is dumb, I apologize. I am a somewhat advanced novice in all things hardware, and still learning.

I have an AMD A8-7600. It has 10 cores (4 CPU + 6 GPU). I needed multiple (4) monitors, so I added two graphics cards.

I disabled the integrated graphics. So does the processor use those 6 GPU cores for anything? I found this, but the comments there are more for intel chip. And they are arguing over whether the CPU cores perform better after adding a dedicated graphics card, not really addressing what happens to the GPUs. The consensus seems to be a faster CPU without the heat generated from GPU and better performance without sharing RAM with GPU. This implies the GPUs are sitting idle, but I am not sure if that's the case, and whether it holds for both Intel and AMD.

My question is more about what those 6 cores DO now that graphics are taken care of. Do the GPU cores sit idle? Or do they get tasked? Is there a way to test this?

(What I would like to hear is that these cores are available. Part of the reason I built this rig was to be able to run several million statistical simulations. That would be helped considerably by being able to run 6 or 8 cores at a time in parallel, instead of just 2 or 3.)

  • 1
    If disabled, then no. – Moab Aug 12 '16 at 17:45
  • Part of the reason I built this rig was to be able to run several million statistical simulations. That would be helped considerably by being able to run 6 or 8 cores at a time in parallel, instead of just 2 or 3. Do the simulations run on cpu? If so more graphics cores will do nothing. – TheKB Aug 12 '16 at 17:56
  • The answer to this questions depends on how your doing your computations and what brand your GPU cards are. Crossfire would not let you, but DX12 would, it all depends how the simulations are calculated – Ramhound Aug 27 '16 at 13:12

I didn't read the link you have, but I disagree with those saying the CPU cores performance is affected whether you have a integrated or dedicated GPU. The only case in where integrated graphics may affect the CPU is when you have single channel system memory, because it would create a bottleneck. An integrated graphics card is either built onto the motherboard, or is built into the die of the CPU. Although they may be housed on the same die, they each have their own separate processing capabilities, because the graphics controller is separate from the CPU.

The biggest difference between a dedicated GPU and an integrated GPU is the memory it allocates. An integrated GPU uses some of your system memory (RAM), rather than having it's own dedicated memory. A discrete graphics card has a GPU and it's own set of VRAM, rather than using the system RAM. The RAM on your graphics card is also much faster than system RAM.

Your question about using the integrated GPU cores as CPU cores, the answer would be no. These cores are designed differently than CPU cores. It could be possible that some programs can use GPU cores, like mining for bitcoins. But those programs are specifically made to run like that.

If they're disabled, then–by definition–they're disabled and doing nothing. As @DrZoo mentioned, the only performance improvement that might be seen is additional RAM being freed.

If, however, you do want to take advantage of these GPU cores within a gaming workload, AMD Dual Graphics allows you to use the GPU cores on your APU in Crossfire with certain AMD cards. According to the product page, AMD recommends pairing with an R7 240 if this is a path you want to go down.

Other workloads such as GPU based render may have other varying support for utilizing the GPU cores on your APU.

I am more knowledgeable of NVidia than AMD, but I believe this answer would apply to AMD as well. Assuming you bought a card from the past couple years, and that it cost more than $25 then you should have massively more GPGPU power available to you than you do in a combo ADM CPU/GPU like the AMD A8 (having owned an A8 & A10 I can safely tell you that).

As DrZoo points out regarding the memory use and architecture of the A8 compared to the discreet card, when it comes to actually coding the simulations you want to do, you will learn how the GPU itself is comprised of memory levels similar to the CPU (L1,L2, L3) and that this memory can be used wisely to be shared between GPU Blocks and Threads, or between Host & Device (CPU & GPU) and also static memory for unchanging data like a value used in your statistical models that must be used in all your calculations. In this case the static memory will only cost you one memory read, and even if your calculation needs that value millions of times it still only needs to go into memory once to retrieve it.

Once you get into the architecture of today's video cards (again, only knowledgeable of NVidia) you will discover that you have access to supercomputing type powers of massively parallel processing that you could never do in an AMD A8 chip. You will realize thoughts of whether the GPU cores are sitting idle of disabled no longer matter, as your $100 video card (or $700 NVidia GTX 1080 in my case) is so substantially more powerful it is meaningless to worry about those cores.

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.