[GameGPU]in Game Summary performance

[Keysplayr]

So we shouldn't test them? You see tests for Titan from places but I'm sure that is even more niche than SLI/CFX.

I would say its to a point these days that if you asked if people preferred single gpu over multi-gpu, almost none would disagree due to the chronic handicaps associated with mgpu setups
So, no. Not really any point in testing mgpu except for sh...s and giggles or if you're bored.
And.... Titan is a single gpu. you can try to strawman that into the conversation, but it really doesnt work.
Folks want a single card over multicard.
You can solicit all your multi gpu friends to now post saying "I lo e my multi gpu setup. Works perfect for me always for yearzorz!"
Well, we'd know their real plight though.

Also, Titan may be more niche (dont actually know if it is or not. Pretty hot seller obviously. ) but only due to its cost. Not because of inherent mgpu issues.

 

[Headfoot]

The only way to know over time if there are still handicaps associated with mGPU set ups is to test mGPU set ups...

 

[Bacon1]

I would say its to a point these days that if you asked if people preferred single gpu over multi-gpu,

Why are you going so off topic on a complete rant?

I was pointing out that MGPU is more common than Titan yet I don't see you complaining when people test Titans.

Also there is no way to know how well MGPU can perform (poor or good) if its not tested.

I never once said people should buy MGPU setups so point your ranting somewhere else.

 

[Keysplayr]

I would say its to a point these days that if you asked if people preferred single gpu over multi-gpu, almost none would disagree due to the chronic handicaps associated with mgpu setups
So, no. Not really any point in testing mgpu except for sh...s and giggles or if you're bored.

Why are you going so off topic on a complete rant?

I was pointing out that MGPU is more common than Titan yet I don't see you complaining when people test Titans.

Also there is no way to know how well MGPU can perform (poor or good) if its not tested.

I never once said people should buy MGPU setups so point your ranting somewhere else.

Ummmmmmmmmmm...... You asked me? And you asked me what about Titan? Then you blow this gasket...
I'm going with the flow of the conversation. That is until you don't like what I say. Then I'm off topic. K.

 

[Bacon1]

Ummmmmmmmmmm...... You asked me? And you asked me what about Titan? Then you blow this gasket...
I'm going with the flow of the conversation. That is until you don't like what I say. Then I'm off topic. K.

Go re-read your post, its a rant not a discussion.

I asked a simple question. Do you think that Titan XP is more common than all 700-1000 and 200-400 series combined SLI/CFX?

 

[Keysplayr]

Go re-read your post, its a rant not a discussion.

I asked a simple question. Do you think that Titan XP is more common than all 700-1000 and 200-400 series combined SLI/CFX?

What is your damage, Bacon? Did I not say that Titan may be more niche? But not for the same reasons that mgpu is niche?
Listen up. If your goal is to make someone believe your aren't really paying attention, you're succeeding.
No rant here whatsoever. My "rant" is 100% entirely fabricated in your mind. we are just talking. well. i am.

 

[ElFenix]

Go re-read your post, its a rant not a discussion.

I asked a simple question. Do you think that Titan XP is more common than all 700-1000 and 200-400 series combined SLI/CFX?

What is your damage, Bacon? Did I not say that Titan may be more niche? But not for the same reasons that mgpu is niche?
Listen up. If your goal is to make someone believe your aren't really paying attention, you're succeeding.
No rant here whatsoever. My "rant" is 100% entirely fabricated in your mind. we are just talking. well. i am.

This stops now. Also, I note your selective quote there, keys.

AT Moderator ElFenix

 

[Keysplayr]

This stops now. Also, I note your selective quote there, keys.

AT Moderator ElFenix

All due respect, ElFenix, I did not realize I cutoff the last word "SLI/CFX." in that quote. I actually had to go back and search for what you were talking about. I fixed it. BUT! No reason to "note" it as it had no affect on the meaning of this conversation as it was implied mgpu was being discussed. Thanks.

As explained to you by PM, this was not your selective quote. You were given the opportunity to fix your error, and you rejected doing so. Further, this constitutes arguing with the moderation staff - if you have a problem with moderation, take it to MD. Enjoy your time off.

AT Moderator ElFenix

 

[PontiacGTX]

I find it interesting that the 470 basically beats a 970 across the board. The 470 brought 970 levels of power to the sub $180 range.

I also find it interesting that the 480 beats a 290x at 1080p. Hawaii had a resurgence Vs Maxwell but is falling off compared to Pascal or Polaris.

Polaris Have more available optimization in drivers. also it was improved for culling primitives/triangles, improved DX11 CPU performance around 5-10% and has twice vram

There is a very big difference, and you're mistaken about the method TPU uses. While the actual error in using GameGPU's method relative to the correct way may be small, the GameGPU adding up FPS method doesn't work - it's not correct, no matter how close it comes to the correct answer.

I'll give you a relevant example. Let's compare two hypothetical scores (made up)

In Game 1 (sort of like DOOM Vulkan): RX 480 gets 110 FPS, while GTX 1060 gets 100. RX 480 ahead by 10%.
In Game 2 (sort of like Watch Dogs 2): RX 480 gets 30 FPS, while GTX 1060 gets 33. GTX 1060 ahead by 10%.

Using these two games, which card is faster on average? Neither - they're equal, of course. But, what if we use GameGPU's method?
.

that onyl means if it were the case most of games have optimization for AMD were tested and calculated an average, if there were more neutral games the result would be even. TPU using NMS put AMD GPUs below Nvidia GPUs because of OpenGL performance

 

[IllogicalGlory]

that onyl means if it were the case most of games have optimization for AMD were tested and calculated an average, if there were more neutral games the result would be even. TPU using NMS put AMD GPUs below Nvidia GPUs because of OpenGL performance

I don't see what it has to do with optimization for one side or another; the same rule applies to any measurement of performance. If one method gives back higher numbers overall than another, the same relative advantage in that method appears greater simply because it gives back higher numbers and vice versa.

It'd be like adding together Cinebench and CPU-Z benchmark scores; CPU-Z just gives bigger (meaningless, except relative to another of the same kind) numbers, so the CPU that does better in that score will have its overall score boosted disproportionately, and being behind in Cinebench would appear less important.

Whether it balances out or not with a lot of games is beside the point; you only use approximate methods like that when the true result is too difficult to obtain, which is definitely not the case in this situation. It may get "close enough" sometimes, but it doesn't in the situation I described. Furthermore I presented a method that works all the time because it gives equal weight no matter the raw numbers and because it gives the same result no matter which card is chosen as reference, and that can be easily done with basic use of MS Excel.