AMD A10-5800K "Trinity" APU On Linux

[ownagefool]

He said nowadays, no game is CPU bound. I prove him wrong. Now, if you want to talk about something else, you're welcome to do it.

If thats true then yu're taking the literal meaning of his statment as opposed to the implied meaning to deflect from your original point. Very good, though I didn't actually see where he said "no game is CPU bound". Let me take your through the discssion from my point (paraphrasing)

• You said nobody would crossfire with such a puny CPU.
• He suggests games were generally gpu bound with the APU, as proven by adding a stronger card (thus making crossfire seem like a sensible approach)
• You pointed out a system which had a significantly stronger GPU (which costs more than the entire APU) doing questionable benchmarks to prove that not all games on all systems are GPU bound.

That doesn't even make sense.

[bug77]

If thats true then yu're taking the literal meaning of his statment as opposed to the implied meaning to deflect from your original point. Very good, though I didn't actually see where he said "no game is CPU bound". Let me take your through the discssion from my point (paraphrasing)

• You said nobody would crossfire with such a puny CPU.
• He suggests games were generally gpu bound with the APU, as proven by adding a stronger card (thus making crossfire seem like a sensible approach)
• You pointed out a system which had a significantly stronger GPU (which costs more than the entire APU) doing questionable benchmarks to prove that not all games on all systems are GPU bound.

That doesn't even make sense.

Point #2: if there's more horespower than the embedded GPU can handle, you need a discrete GPU, not a crossfire setup.
Point #3: what questionable about running several games on several CPUs, using the same GPU? It just shows games can benefit from a more powerful CPU.

[HokTar]

Point #3: what questionable about running several games on several CPUs, using the same GPU? It just shows games can benefit from a more powerful CPU.

Sure, at 1024x768 low detail everything is CPU bound. But does anybody still play at that level?
Turn it to 1920x1200 with medium/max detail and check again. Surprise-surprise, you don't benefit from the stronger CPU now and this is exactly what he's been trying to explain you.
Hence spending more money on a stronger GPU and buying a cheaper APU is likely better than buying a stronger CPU and a weaker GPU for the same sum.

[bug77]

Sure, at 1024x768 low detail everything is CPU bound. But does anybody still play at that level?
Turn it to 1920x1200 with medium/max detail and check again. Surprise-surprise, you don't benefit from the stronger CPU now and this is exactly what he's been trying to explain you.
Hence spending more money on a stronger GPU and buying a cheaper APU is likely better than buying a stronger CPU and a weaker GPU for the same sum.

Two of the test I mentioned were using 1680x1050. Crysis was set at Mainstream details, Civ V at High Quality. Yet the difference was still there. I wish there were more in-depth tests out there. In the meantime, I'll take cold hard numbers over wishful thinking.

[Ansla]

Point #2: if there's more horespower than the embedded GPU can handle, you need a discrete GPU, not a crossfire setup.

Where are the cold hard numbers o back this up? What if the hybrid crossover brings better performance then a more expensive dedicated card alone can? Here's an example where at full HD resolusion it provides a significant improvement compared to the same dedicated card alone: http://www.tomshardware.com/reviews/...0k,3224-4.html. Unfortunalety I couldn't find more in depth benchmarks for hybrid crossfire nor any benchmarks of hybrid crossfire on Linux (if it works at all).

Point #3: what questionable about running several games on several CPUs, using the same GPU? It just shows games can benefit from a more powerful CPU.

It's questionable because it's not a realistic scenario, most desktops used for gaming today have displays with at least full HD native resolution while the benchmark was only performed at lower resolutions.

[tuke81]

Sure, at 1024x768 low detail everything is CPU bound. But does anybody still play at that level?
Turn it to 1920x1200 with medium/max detail and check again. Surprise-surprise, you don't benefit from the stronger CPU now and this is exactly what he's been trying to explain you.
Hence spending more money on a stronger GPU and buying a cheaper APU is likely better than buying a stronger CPU and a weaker GPU for the same sum.

It's not that simple. It usually depending on the engine is the game cpu or gpu pounded(physics,ai,etc.). But granted in 1200p resolution it's usually gpu pounded. And I don't think that these apus can run modern games with max settings at 1200p, maybe 720p which is okay resolution for casual gaming in living room.

[ownagefool]

Point #2: if there's more horespower than the embedded GPU can handle, you need a discrete GPU, not a crossfire setup.
Point #3: what questionable about running several games on several CPUs, using the same GPU? It just shows games can benefit from a more powerful CPU.

#2: Not when having the discrete system you're suggesting costs more. You compare by total cost of ownership or you're comparing about a metric that nobody really cares about.

#3: Running the games on lower than normal settings so they'll use more CPU power than GPU power is questionable. Nobody would really do that unless they're trying to make a point whilst benchmarking. Those same games will be running at 1000 FPS in the future when we have CPUs and GPUs which far exceed todays. So yes, they can benefit from a more powerful CPU, but that point is completely and utterly stupid, just like yours is. You talk about performance per $, or you're not talking the same language as anyone else.

Two of the test I mentioned were using 1680x1050. Crysis was set at Mainstream details, Civ V at High Quality. Yet the difference was still there. I wish there were more in-depth tests out there. In the meantime, I'll take cold hard numbers over wishful thinking.

Here are some cold hard numbers:-

1, 2, 3, 4, 5, 6, 7, 8, 9, 0.

They're about as relevant as yours.

[bug77]

Here are some cold hard numbers:-

1, 2, 3, 4, 5, 6, 7, 8, 9, 0.

They're about as relevant as yours.

They're not my numbers, they're Anand's. The only benchmark I could find that tests the impact of the CPU on gaming. I don't think it's irrelevant just because it doesn't support your point of view. Until you can show me benchmarks that say otherwise, you could at least have the courtesy not to call me names.

[ownagefool]

They're not my numbers, they're Anand's. The only benchmark I could find that tests the impact of the CPU on gaming. I don't think it's irrelevant just because it doesn't support your point of view. Until you can show me benchmarks that say otherwise, you could at least have the courtesy not to call me names.

The problem with those numbers is they don't support your point of view.

[bug77]

The problem with those numbers is they don't support your point of view.

Then why are you so quick to dismiss them?
And how come they don't support my point of view? If Trinity would have enough horsepower as to not bottleneck games, then more powerful CPUs wouldn't push more FPS. And again, two tests were run at 1680x1050 with pretty high settings.