1 of 1 people found this helpful
Q4: Where do I plug my monitor in to get the best performance when using AMD Radeon™ Dual Graphics technology?
A4: For best performance, AMD recommends plugging the display into your fastest GPU. This will ensure that even applications that do not take advantage of AMD Radeon™ Dual Graphics will still be able to run on the faster graphics card in your system. We recommend that you plug your monitor into the AMD Radeon™ GPU with the higher model number. If that is your APU, plug into the motherboard, if it is your discrete graphics card, plug into that.
Well what it says in your reply is something which did not happen, in my system scenario.
The PCI-e card which I use (MSI R7-250 OC 2GB) is a bit of lower specs than the R7-igd graphics that comes in AMD A10-7850k APU.
IGD R7 Graphics: 2GB, 512 Shader Cores, Mem-25GB/s
R7-250 OC 2GB : 2GB, 384 Shader Cores, Mem-15GB/s (PCI-e x16)
That is my IGD is more powerful. But when I use IGD for display, hell goes the performance. 5-7 FPS on NFS-2016 Deluxe Edition, Crysis3, COD-AW, Battlefield-4, Battlefield-1 BETA..
While if I use PCI-e as display port, then above games work at 25-30-35 FPS
// Battlefield-1 BETA worked at 19-26 fps(Don't know what c'd have happened with it on igd dual graphics!!!)
2 of 2 people found this helpful
Your main GPU which is the integrated one, will downgrade itself to match the other GPU if you enable dual graphics with it.
The reason why the Discrete GPU performs better is because by not using the integrated GPU, your APU can focus on more CPU oriented work and offloads the GPU work to an actual GPU component, that gets it's own wattage and such.
You won't see decently performing APU configurations until the AM4 and Zen APUs comes out, which you can pair with the RX 460 in that case.
Heck, by just buying a RX 460 and using your APU as a CPU, you'd get a ton of performance back!
Btw, even if a APU has more shaders than a specific GPU, the Discrete GPU has a larger 'silicone' to use, over a APU that is split between CPU and GPU modules inside one large 'silicone'.
For gaming, memory bandwidth is generally for texture intensive games that likes to load in and store many many objects at once.
Overall, the core clock is better valued for it's overall rendering speeds for effects, physics, and such.
(Where as memory clock is for 'storing' texture files into the VRAM, the core clock affects how fast the display can render and update your screen).
Gotcha bro.. i have read about the HSA architecture in this APU.. It shuts of GPU work where possible to use the cores for CPU work and they do not sit idle.. Thats something cool though..!!!