AMD vs. Nvidia

[UaVaj]

not surprise by the poll either. nowadays amd is an awesome VALUE when it comes to single gpu. only if amd would just get cross broke fixed once and for all. currently running 680 tri sli. would not mind upgrading to 290x tri fire.

lastly how do you beat the performance of a 7950 for $190 shipped AMIR.
http://www.newegg.com/Product/Produc...FSdp7AodH3YAPA

 

[amenx]

not surprise by the poll either. nowadays amd is an awesome VALUE when it comes to single gpu.

With the massive price cuts AMD had applied to their mid to top cards, I'm not surprised either. But ultimately I dont think they're too happy about the horrible margins they may have had to incur in the process.

 

[rgallant]

With the massive price cuts AMD had applied to their mid to top cards, I'm not surprised either. But ultimately I dont think they're too happy about the horrible margins they may have had to incur in the process.

this - amd needs to make a fair return on investment maybe the new cards will bring a fair mind share on the net.

 

[Rvenger]

nvm

 

[lavaheadache]

sure you can. which is your primary gaming machine now (today)? what gpu(s) is in that one machine. not your wife's machine. not your kid's machine. not your friend's machine. not what you had last month, last year.

direction is clear as day. no need to make this hard than it is.

You do know that you can run cards from both vendors at the same time right? No need to assume that I just try to make things difficult. In fact I think I try to facilitate things when I post.

 

[UaVaj]

You do know that you can run cards from both vendors at the same time right? No need to assume that I just try to make things difficult. In fact I think I try to facilitate things when I post.

yes you can run both amd and nVidia in the same machine. obviously you cannot sli-fire so which one are you gaming with? that is the one that count toward the poll.

exception is you are talented enough to multi-task on 2+ different games simultaneously on multi-monitor with mulit-gpu. :thumbsup::thumbsup: that would require 2+ keyboard and 2+ mouse.

 

[chimaxi83]

yes you can run both amd and nVidia in the same machine. obviously you cannot sli-fire so which one are you gaming with? that is the one that count toward the poll.

exception is you are talented enough to multi-task on 2+ different games simultaneously on multi-monitor with mulit-gpu. :thumbsup::thumbsup: that would require 2+ keyboard and 2+ mouse.

Um. Hybrid PhysX?

 

[UaVaj]

Um. Hybrid PhysX?

another semi exception. nVidia PhysX would not run by itself. as a slave gpu, it would need a master GPU, the GPU that does the bulk of the rendering/gaming would be the one counted toward the poll.

Reasonable?

 

[Black Octagon]

I currently use the card that I considered to be the best buy for what I could afford when building my rig back in Jan 2012.

It just so happens that that card, at the time, had an AMD chip. Next time I chose to purchase a card it may be that (again, at that point in time) an NVIDIA offering meets my performance and price needs

 

[Gikaseixas]

I didn't get my 7970's only because of price, it did had a major impact in the decision making but more importantly was the performance, the higher memory, the cooler and it's noise characteristics. Those who are downplaying AMD's lead here are Nvidia users with no first hand experience and rely on internet buzz to make decisions lol.

 

[natto fire]

I voted AMD because of the 7950 in my desktop gaming rig, but lately I have had to do a lot of gaming on my laptop which has a GT630M, so more Nvidia atm.

 

[cmdrdredd]

NVs frame pacing does not work in every title either.

Cause they don't use frame pacing, they use Frame Metering which works globally. The problem is lots of game engines stutter by default. Skyrim, Farcry 3, Bioshock Infinite to name three prime examples.

 

[Black Octagon]

FC3 being the worst of your the examples by far. The other 2 are acceptable on a single gpu, after a bit of tweaking.

 

[Final8ty]

Cause they don't use frame pacing, they use Frame Metering which works globally. The problem is lots of game engines stutter by default. Skyrim, Farcry 3, Bioshock Infinite to name three prime examples.

Stop with the semantics and the point is it does not work in every game so stop applying different rules and absolutes, multi GPU scaling does not work 100% in every game either but your still using it.

 

[Deders]

Stop with the semantics and the point is it does not work in every game so stop applying different rules and absolutes, multi GPU scaling does not work 100% in every game either but your still using it.

Cause they don't use frame pacing, they use Frame Metering which works globally.

^^^

 

[3DVagabond]

Stop with the semantics and the point is it does not work in every game so stop applying different rules and absolutes, multi GPU scaling does not work 100% in every game either but your still using it.

Agreed. When a game stutters on AMD, it's their crappy drivers. When a game stutters on nVidia, it's the crappy game.

 

[bystander36]

Agreed. When a game stutters on AMD, it's their crappy drivers. When a game stutters on nVidia, it's the crappy game.

That might be because the majority of the time it stutters on Nvidia, it also stutters on AMD, but the opposite could not be said without the latest frame pacing drivers in conditions it works on.

Of course I don't know how much involvement is needed on Nvidia's part to get frame pacing to work on games. It may work be default on all games, or it may require tinkering.

 

[blackened23]

To be fair any and all "stutter" issues were fixed and resolved long ago. If you look at the PCPer review of the 280X, it consistently meets or beats the GTX 770 in terms of frametimes, in every single game tested. These issues are a non issue aside from the clipping issue with CF-eyefinity. Yet, there are indeed some games with horrible engines that are stutter prone, and no hardware can fix it. Fallout 3 and New Vegas come to mind.

Here is PCPer's review. http://www.pcper.com/reviews/Graphics-Cards/AMD-Radeon-R9-280X-R9-270X-and-R7-260X-Review

280X seems to beat the 770 in most of those tests in frametimes. That said, AMD still needs to fix the clipping issue with CF eyefinity. AMD's software isn't quite on NV's level yet, I would agree on that. I would also agree that AMD still has a lot of work to do in the software area, and they need to do it post haste. But this smoothness and stutter talk is nonsense for the time being, long ago fixed for single screen crossfire resolutions.

 

[Deders]

Agreed. When a game stutters on AMD, it's their crappy drivers. When a game stutters on nVidia, it's the crappy game.

I actually found in my case when I had 2x 9800GTX+'s in SLI, upgrading from an EVGA 590 motherboard with PCIe 1.1x slots at 16x to an Asrock P55 board with PCIe 2.0 8x caused a lot of stuttering. I returned that board and replaced i with my current MSI, also PCIe 2.0 8x and found things were smoothed out again.

This is not the first Asrock I found had PCIe issues with stuttering, even with a single card.

 

[Enigmoid]

At the moment there is simply no comparison for mobile. Nvidia >>>>>>> AMD.

 

[3DVagabond]

That might be because the majority of the time it stutters on Nvidia, it also stutters on AMD, but the opposite could not be said without the latest frame pacing drivers in conditions it works on.

Of course I don't know how much involvement is needed on Nvidia's part to get frame pacing to work on games. It may work be default on all games, or it may require tinkering.

You mean like this?

Or this?

Not as bad, but still better on the 280X

To be clear, I'm not trying to start some big FUD campaign against nVidia. I'm just saying that the same results would invariably be shown as Crappy, or cheating, AMD drivers. It goes completely unnoticed when it's reversed though. :\

 

[bystander36]

You mean like this?

Or this?

Not as bad, but still better on the 280X

To be clear, I'm not trying to start some big FUD campaign against nVidia. I'm just saying that the same results would invariably be shown as Crappy, or cheating, AMD drivers. It goes completely unnoticed when it's reversed though. :\

To be clear, I was talking about microstutter, not the occasional high latency frame. None of those types of spikes have ever been the topic of any review I've seen.

 

[blackened23]

To be clear, I was talking about microstutter, not the occasional high latency frame. None of those types of spikes have ever been the topic of any review I've seen.

Have you not paid attention to any review in the past 6 months or so? Frametimes measure EXACTLY that. Microstutter. Generally using nvidia's FCAT software which is designed explicitly to objectively measure "smoothness", PCPer has their own in house method which captures footage. PCPer was also one of the first sites to objectively measure "smoothness" in this way and were one of the websites to bring attention to the CF issues that AMD had at the time. Apparently, the 280X is on par or better than the GTX 770 in terms of smoothness, so AMD has fixed the issue despite what you said earlier, aside from eyefinity CF.

 

[bystander36]

Like, are you serious or what? Frametimes measure EXACTLY that. Microstutter.

Those are occasional spikes. Even the AMD ones had occasional spikes in those.

This is what the Microstutter issues looked like:

You showed a chart with a few bad frames, but nothing like the microstutter issue we've been talking about the last few months.

 

[blackened23]

Oh. You just linked old frame time charts from many months ago prior to AMD's driver fix.