causing a singularity, red and green together...

[Erenhardt]

Why not test the freaking apu with high end gpu (and 2012 highend gpu) for crying out loud!

Much dissapoint anandtech.

 

[tential]

^^ This. Currently its just a glorified AFR mixed SLI/CF mode.

Cool first step, but its not moving any bars.

So being able to use different gpu vendors together isn't moving any bars? Ok....

 

[stuff_me_good]

Looking good for team red. Just wondering when team green is realising some "magic" drivers? :whiste: But going team red is money well spent no matter how you try to spin it.

 

[NTMBK]

Why not test the freaking apu with high end gpu (and 2012 highend gpu) for crying out loud!

Much dissapoint anandtech.

Because it's AFR. Mismatched performance + AFR = stuttering.

 

[tential]

Why not test the freaking apu with high end gpu (and 2012 highend gpu) for crying out loud!

Much dissapoint anandtech.

Because the number of people with apus and high end gpus is a joke?

Edit:ntbmk may have a better reason above.

 

[ShintaiDK]

So being able to use different gpu vendors together isn't moving any bars? Ok....

I know what you mean. But in the grand scheme of things no.

However it is alpha, it is early stage and what not. So I hope we see more than a crossvendor AFR in the future.

 

[monstercameron]

Because the number of people with apus and high end gpus is a joke?

Edit:ntbmk may have a better reason above.

i7 6700k? nearly every highend gpu is paired with a decent "apu" or "igp."

 

[tential]

i7 6700k? nearly every highend gpu is paired with a decent "apu" or "igp."

You know exactly what the post I'm responding to is talking about. Please don't nitpick it derails the conversation. Erenhardt asked for apus to be tested with high end gpus. He clearly did not mean the tests already done using the 6700k....

 

[tential]

I know what you mean. But in the grand scheme of things no.

However it is alpha, it is early stage and what not. So I hope we see more than a crossvendor AFR in the future.

I agree but at least we are having less and less vendor locks.
Ya it'd be infinitely better if we could mix and max gpus of any performance caliber though and have smooth gaming I agree I just don't think we'll get it sadly.

 

[BigDaveX]

Why not test the freaking apu with high end gpu (and 2012 highend gpu) for crying out loud!

Much dissapoint anandtech.

Because the amount of CPU power (especially on a Kaveri) it'd take to balance out the load between the 512 shaders of the iGPU and the 4096 shaders of a Fury X would probably cancel out any performance gain you'd get from using the iGPU?

Not to mention that AMD themselves didn't extend support for iGPU+dGPU Crossfire to anything above the R7 250 on the graphics card side, so there's no reason to think that this solution would work any better.

 

[ShintaiDK]

IGP could have been used for post processing. But the application would have to support that mode.

 

[Erenhardt]

Because the amount of CPU power (especially on a Kaveri) it'd take to balance out the load between the 512 shaders of the iGPU and the 4096 shaders of a Fury X would probably cancel out any performance gain you'd get from using the iGPU?

Not to mention that AMD themselves didn't extend support for iGPU+dGPU Crossfire to anything above the R7 250 on the graphics card side, so there's no reason to think that this solution would work any better.

I somehow thought it was running SFR, not AFR. Nevermind then.

 

[frozentundra123456]

Why not test the freaking apu with high end gpu (and 2012 highend gpu) for crying out loud!

Much dissapoint anandtech.

FWIW, direct quote from Ryan in the comments of that article:

Ryan Smith - Monday, October 26, 2015 - link
Since this early release is limited to basic AFR, there's little sense in testing an iGPU. It may be able to contribute in the future with another rendering mode, but right now it's not nearly fast enough to be used effectively.

Not sure that really clarifies things, but that is his comment.

 

[dogen1]

FWIW, direct quote from Ryan in the comments of that article:

Ryan Smith - Monday, October 26, 2015 - link
Since this early release is limited to basic AFR, there's little sense in testing an iGPU. It may be able to contribute in the future with another rendering mode, but right now it's not nearly fast enough to be used effectively.

Not sure that really clarifies things, but that is his comment.

The iGPU would be taking way longer to render each alternate frame so you'd either get really uneven frame delivery or you'd have to slow the dGPU down to the pace of the iGPU.

Unless you're really interested in seeing an iris 6200 with a 7750 or something.

 

[ViRGE]

So I guess this puts us one step closer to the Bitchin' Fast! 3D 2000?

 

[Silverforce11]

This is really cool, but I think at least one - probably nvidia - and likely both of them will try to squash it. It IS part of the DX12 spec though, but we've already seen neither of them feel obligated to support all facets of the API.

It's reasonable that we could possibly see using a halo card from each vendor giving overall better performance than SLI or CF just using one vendor's cards. What gives gamers the best experience is not going to matter to them though if they feel it is losing them money.

They have no reason to squash it, if they do, they are actually NOT DX12 compatible and cannot put DX12 logos on their products, period.

The real interesting mode is SFR with iGPU helping dGPU rendering. A 10-20% performance uplift from the normally useless iGPU is a major win for PC gamers since we're mostly all rocking a CPU with an iGPU that's just being a waste of space & money in DX11.

 

[ShintaiDK]

They have no reason to squash it, if they do, they are actually NOT DX12 compatible and cannot put DX12 logos on their products, period.

The real interesting mode is SFR with iGPU helping dGPU rendering. A 10-20% performance uplift from the normally useless iGPU is a major win for PC gamers since we're mostly all rocking a CPU with an iGPU that's just being a waste of space & money in DX11.

I dont think we gonna see SFR that way unless its in the low end. Just post processing.

And I fear any combo with the IGP involved=ekstra lag. Not a problem most games, but can be a huge problem in those needing the extra FPS.

 

[SPBHM]

7970 and 680 was the most interesting part because the 680 was a good bit slower and would kind of make sense as displaying a situation where you get a newer card that is faster as an upgrade but not all that much but with this you can use them both combined for added performance... but realistically, since it's game dependent it's kind of useless, because most of the time you would run in single GPU while having all these combinations, maybe in 3 years it will be viable or something.

BUT for Intel IGPs it makes sense because almost everyone have one anyway, so if a single game supports it it's already a win.

 

[tential]

I dont think we gonna see SFR that way unless its in the low end. Just post processing.

And I fear any combo with the IGP involved=ekstra lag. Not a problem most games, but can be a huge problem in those needing the extra FPS.

Isn't the igp doing post processing (or something else) still in the most simplest of terms sfr? Since it's 2 gpus working together to complete the frame vs alternate frames?

 

[bystander36]

My question is just how much is the IGP helping with post processing. After all, at the point you do post processing, the frame is done, and I'd think the GPU would be faster at it, unless the GPU starts the next frame while the IGP does the post process stuff. At which point you gain a few FPS but also add latency.

I'd think for best results, the IGP might be best used with physics stuff and multi-GPU stuff is best off with similarly fast GPU's.

 

[Silverforce11]

The latency for communications with other GPU or iGPU over the PCIe lanes was mentioned as incurring a 2ms latency penalty.

This means if your FPS is already very high, offloading would be a disadvantage. As long as the performance gains for shifting work to the iGPU is more than 2ms worth, then you get a net advantage.

This is obviously more advantageous the lower the FPS is, if its 30 fps for instance, 2ms is 2/32 of the frame.

That's for real-time rendering, I'm sure there are tasks which can run asynchronously without impacting the rendering performance. If used as such, then it's pure performance gains.

I hope all the major AAA game engines & devs move to support offloading to the iGPU in the DX12 era, it would be exciting for gamers to leverage more of their hardware. It would entice Intel & AMD to include stronger iGPUs, maybe even HBM2 powered iGPUs in the near future with much higher performance ceilings due to no system-ram bandwidth limitations.

 

[lopri]

The visual is impressive but I am not very optimistic the final game will be anywhere near epic. Prove me wrong please!

 

[Erenhardt]

FWIW, direct quote from Ryan in the comments of that article:

Ryan Smith - Monday, October 26, 2015 - link
Since this early release is limited to basic AFR, there's little sense in testing an iGPU. It may be able to contribute in the future with another rendering mode, but right now it's not nearly fast enough to be used effectively.

Not sure that really clarifies things, but that is his comment.

Thanks, makes sense.

They have no reason to squash it, if they do, they are actually NOT DX12 compatible and cannot put DX12 logos on their products, period.

They could still slap their own dx12.3 sticker onto it and plow the sales.

About the extra lag from using igp. AMD did a great job with kaveri +r7 gpu.
http://www.pcper.com/reviews/Graphi...0-Dual-Graphics-Testing-Pacing-Fixed/Battlefi

 

[dogen1]

About the extra lag from using igp. AMD did a great job with kaveri +r7 gpu.
[/QUOTE] Right, but any AFR setu...s further ahead of when they'll be displayed.

 

[ShintaiDK]

Isn't the igp doing post processing (or something else) still in the most simplest of terms sfr? Since it's 2 gpus working together to complete the frame vs alternate frames?

But that's not SFR is it?