Perspective of silicon area budget.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
- Status
- Not open for further replies.
Perspective of silicon area budget.
DavidGraham
Veteran
Power consumption, heat dissipation, die area, and cost are real practical limits that will prevent any APU from surpassing a high end dGPU + dCPU, choosing to ignore these limits shifts you to the realm of fantasy.Why should the APU be any less capable? Because you choose to impose artificial limits to the design?
You lost your scientific credibility right when you used ALWAYS.Tightly integrated designs ALWAYS result in lower power usage and higher performance in electrical systems.
GPUs were created on the premise of spreading out processors to increase performance. When we reach enough satisfactory visual quality they could come back inside, seeing that we are so much far away from a good enough visual experience, the current status quo stands to remain.I'd challenge you to find a single example where spreading out processors increased performance.
And yes, the current PC structure gives the best performance possible, it's how super computers are built anyway. As opposed to phones or tablets, where performance (even compared to low end PCs) is non existent really.
Magnum_Force
Newcomer
I think people are muddling the argument here by sticking to currently established form factors.
Given an equal die area, memory system, power budget and silicon process, I can't see how an APU would fair worse than a separate GPU and CPU, who's power budget and die area you would need to combine for a fair comparisons.
Otherwise, why not make the argument "which would perform better, a separate GPU and CPU, or 2 x APUs on the same motherboard"
Given an equal die area, memory system, power budget and silicon process, I can't see how an APU would fair worse than a separate GPU and CPU, who's power budget and die area you would need to combine for a fair comparisons.
Otherwise, why not make the argument "which would perform better, a separate GPU and CPU, or 2 x APUs on the same motherboard"
But thats the point some people are trying to make dgpu's are faster because they dont have to have the same die area, power budget or memory systemGiven an equal die area, memory system, power budget
No one is claiming dgpu's are more efficient or better value for money, just that they are faster, and no one is claiming the comparison is fair
We have apu's and we have discreet cpu's (ok they have an igp sometimes) and discreet gpu's one of these two options is faster and its not the apu
Magnum_Force
Newcomer
Then the comparison is meaningless.
You can't talk about non-specific discrete GPUs and CPUs in comparison to non-specific APUs and claim one is faster.
For example, tell me what would perform better, a nvidia GT 710 and a Celeron G1620 or a A10 7890k
You can't talk about non-specific discrete GPUs and CPUs in comparison to non-specific APUs and claim one is faster.
For example, tell me what would perform better, a nvidia GT 710 and a Celeron G1620 or a A10 7890k
we are not, we are talking the fastest in each bracket, the fastest apu vs the fastest cpu+gpuYou can't talk about non-specific discrete GPUs and CPUs in comparison to non-specific APUs and claim one is faster.
The original premise for the spin-off was APU vs discrete in terms of delivering VR performance, but it should have been obvious what the answer would be instead of >100 posts later.
Instead we have caveats upon caveats to prove one's own POV is the correct one.
Enough is enough.
...from a certain point of view.
Instead we have caveats upon caveats to prove one's own POV is the correct one.
Enough is enough.
...from a certain point of view.
- Status
- Not open for further replies.