Lorem ipsum dolor sit amet

[I/O]

How are the settings skewed? The 6GB card allowed them to play at higher settings. Which is the point of the benchmark to prove whether or not 3GB of memory is a bottleneck in certain scenarios. 

Then why did the 290 perform within margin of error of the 6GB card ?

[BiG StroOnZ]

Then why did the 290 perform within margin of error of the 6GB card ?

 

Both the 290 and the 6GB 780 have over 3GB of available memory. Whereas the regular 780 has only 3GB to use.

[I/O]

Both the 290 and the 6GB 780 have over 3GB of available memory. Whereas the regular 780 has only 3GB to use.

The 290 used higher setting . + 1GB Vram with the settings used @ 1600P 2xMSAA did not make the card perform better .... Unless that is if you SKEW the numbers.

[BiG StroOnZ]

The 290 used higher setting . + 1GB Vram with the settings used @ 1600P 2xMSAA did not make the card perform better .... Unless that is if you SKEW the numbers.

 

There's only one benchmark where they used 2x MSAA and that's in Far Cry 3:

 

 

Both the 4GB 290 and the 6GB 780 performed better than the standard 3GB 780. 

[I/O]

There's only one benchmark where they used 2x MSAA and that's in Far Cry 3:

 

 

Both the 4GB 290 and the 6GB 780 performed better than the standard 3GB 780. 

Yes the 290 and the 770 WITH OC are faster than reference clocked 780 LOL

[BiG StroOnZ]

Yes the 290 and the 770 WITH OC are faster than reference clocked 780 LOL

 

Firstly I don't see a 770?

 

Secondly, we went over this already:

 

 

The differences in clock speed aren't enough to translate for the extra fps. 

 

You must have forgotten when you made this claim, and then you changed your tune after I showed you this card configuration 

[I/O]

Firstly I don't see a 770?

 

Secondly, we went over this already:

 

 

The differences in clock speed aren't enough to translate for the extra fps. 

Says who .... ? That's why the Intel in non conclusive.

[BiG StroOnZ]

Says who .... ? That's why the Intel in non conclusive.

 

Just looking at before and after overclocking results of a 780 you can see in order to get those kind of increases in framerates due to pure power increases you need much higher clocks (and even then sometimes it doesn't make that much of a difference):

 

 

 

 

So as you can see to get those type of results because of an overclock you are looking at 75MHz on the core as opposed to the 40MHz difference between the 3GB 780 and 6GB Strix used in the review. Which as I said before isn't that much of a difference in core clocks to make such a substantial difference in performance. To at least chalk up the performance increases as purely related to clock speeds.

[I/O]

Just looking at before and after overclocking results of a 780 you can see in order to get those kind of increases in framerates due to pure power increases you need much higher clocks (and even then sometimes it doesn't make that much of a difference):

OCing does not always mean great or even better performance. That said you are still not quantified because the games and setting have much more an impact on performance than does OCing. Don't know if trolling or just unknowing !

[Faa]

TH proves that the Vram capacity does not increases performance. Contrary to your limited view ...

More vram doesnt improve performance unless the GPU is using the storage drive as extra vram, having more vram just removes the vram bottleneck therefore allowing the gpu to hit 99% load because it wouldn't need to use the storage drive as swapfile anymore. You're right though.

[I/O]

More vram doesnt improve performance unless the GPU is using the storage drive as extra vram, having more vram just removes the vram bottleneck therefore allowing the gpu to hit 99% load because it wouldn't need to use the storage drive as swapfile anymore. You're right though.

Well correct about the page file swapping. I only experienced this anomaly in Skyrim "heavily modded" @ 1080P on a 1GB card. Vram does not make a GPU faster ever though rather it just allows it to work properly ala the page swapping anomaly.

[Faa]

Well correct about the page file swapping. I only experienced this anomaly in Skyrim "heavily modded" @ 1080P on a 1GB card. Vram does not make a GPU faster ever though rather it just allows it to work properly ala the page swapping anomaly.

Yea exactly. You either have enough or you'll be measuring your performance in frames per hour.

[I/O]

Yea exactly. You either have enough or you'll be measuring your performance in frames per hour.

On a 1GB HD 7850 in heavily modded skyrim it would take upwards of 5min or more to load into some towns because of the page file swapping. When you would finally get it load the game would run at like 1fps in said towns. On the 2GB HD 7850 the game ran just fine with same settings and mods.

[BiG StroOnZ]

OCing does not always mean great or even better performance.

 

Then why exclaim that the performance increases in the benchmark are because of the OC then...?

 

Here you go changing your tune again after you get proven wrong, as expected.

[I/O]

Then why exclaim that the performance increases in the benchmark are because of the OC then...?

 

Here you go changing your tune again after you get proven wrong, as expected.

Nope it's just the nature of OCing .. Some games really favor an OC on the core and memory while others just want more GPU cores not fewer running faster. This is why on some games even the OCed card you will see that the min framrate stays the same or does not change at all sometime even lower.

[JDS]

Nope it's just the nature of OCing .. Some games really favor an OC on the core and memory while others just want more GPU cores not fewer running faster. This is why on some games even the OCed card you will see that the min framrate stays the same or does not change at all sometime even lower.

I guess I'll be the one to tell you to just shut the hell up. Learn to accept a loss, and find a new forum to ruin.

[I/O]

I guess I'll be the one to tell you to just shut the hell up. Learn to accept a loss, and find a new forum to ruin.

 

Note in some games do not really benefit from an OC 2fps is a very small gain. Some games are even less.

 

[BiG StroOnZ]

Note in some games do not really benefit from an OC 2fps is a very small gain. Some games are even less.

Okay you seem to be losing track of your claims. You claimed the performance increase was because of the factory overclock of 40MHz that the Strix had over the standard 780. Now you are claiming overclocks dont make a difference in certain games. So which is it, it cant be both. LOL

[I/O]

Okay you seem to be losing track of your claims. You claimed the performance increase was because of the factory overclock of 40MHz that the Strix had over the standard 780. Now you are claiming overclocks dont make a difference in certain games. So which is it, it cant be both. LOL

Okay some games like OCs but others like Crysis 3 as shown above do not really gain any real time perceivable performance. Judging by the metric you speak of in that as you say vram increases performance then that should mean that the Titan 6GB should blow away the 3GB 780 when in fact the small gain in performance at leased in Crysis 3 as shown above is due to the increase in Cuda cores not Vram or clock speed. Heck in the graph posted above the 7990 which is a 3GB card effectively outperforms all because it has more stream processors and is not Vram limited in said game. Crank up the AA in Crysis 3 and at 2560x1600 we might see the Titan 6GB pull away from the 7990 and the 3GB 780 but I doubt it as at that resolution 3GB should be enough Vram.

[BiG StroOnZ]

Okay some games like OCs but others like Crysis 3 as shown above do not really gain any real time perceivable performance. Judging by the metric you speak of in that as you say vram increases performance then that should mean that the Titan 6GB should blow away the 3GB 780 when in fact the small gain in performance at leased in Crysis 3 as shown above is due to the increase in Cuda cores not Vram or clock speed. Heck in the graph posted above the 7990 which is a 3GB card effectively outperforms all because it has more stream processors and is not Vram limited in said game. Crank up the AA in Crysis 3 and at 2560x1600 we might see the Titan 6GB pull away from the 7990 and the 3GB 780 but I doubt it as at that resolution 3GB should be enough Vram.

 

Unfortunately for you they tested Crysis 3 in the Strix 6GB 780 vs the standard 3GB 780 review:

 

 

Now according to you this shouldn't make such a drastic difference in performance since you are claiming that in Crysis 3 you do not gain any real time perceivable performance after an overclock. Well I would say nearly 7 fps is real time perceivable performance. On top of that, in the benchmark you posted to use as a comparison the 780 is clocked at 1,146MHz and is boosting 1,228MHz. With that OC they only gained 2 fps, whereas before you were claiming that the only reason why the 6GB 780 Strix had an advantage was from its measly 40MHz increase. Well now by posting that information, you just further proved that you are wrong and it is not the 40MHz increase of core clocks that the Strix has over the standard 780 that resulted in it performing better. Which then brings you back to only reason. It's the extra VRAM that the Strix has over the vanilla 780. Since we now know indefinitely it's not the core clocks that resulted in the performance increase, thanks to you for posting that graph... 

 

Firstly, the Titan does perform better than 780, as a matter a fact it even performs better than the 780 Ti which is knowingly faster than the standard Titan (not Titan Black). The only reason why the 780 stood a chance against the Titan was because it was boosting to 1228MHz, which is a 33 percent higher clock speed than the standard GTX 780's. This is common knowledge, that a nicely overclocked 780 can come within Titan performance after overclocked. But again, it has nothing to do with the topic at all and is irrelevant. We're trying to see if VRAM can become a bottleneck and this is not what the benchmark you posted is trying to do.

 

Lastly, in this terrible benchmark they didn't even have AA turned on, they disabled it. Which defeats the point of testing out whether or not lesser amounts of VRAM can become bottlenecks. Which is the point of the 780 6GB vs the 3GB 780 review. To see if VRAM becomes a bottleneck. Not like the benchmark you posted which is simply reviewing a non-reference 780. 

[I/O]

Unfortunately for you they tested Crysis 3 in the Strix 6GB 780 vs the standard 3GB 780 review:

 

Now according to you this shouldn't make such a drastic difference in performance since you are claiming that in Crysis 3 you do not gain any real time perceivable performance after an overclock. Well I would say nearly 7 fps is real time perceivable performance.

4fps on the minimum is not much at all. In fact it's really not percivable unless you are looking at the fps counter. None of those three cards would offer a noticeably better of worse gaming experience when actually playing the game. Also that graph you posted is inconclusive as we do not know what the Vram count was at so we cannot tell if the 3GB 780 was Vram bottlenecked.

[BiG StroOnZ]

4fps on the minimum is not much at all. In fact it's really not percivable unless you are looking at the fps counter. None of those three cards would offer a noticeably better of worse gaming experience when actually playing the game.

 

Well if you are going to nitpick about the minimums (even though it's entirely irrelevant). The max FPS is exactly 10 fps higher, and the average framerate is about 7 fps higher. I would say that is more than perceivable. Hate to break it to you but you are running out of bullcrap to spew. 

[SlaughterSmurf]

Unless you are planning on SLI then 780 ti.

[I/O]

Well if you are going to nitpick about the minimums (even though it's entirely irrelevant). The max FPS is exactly 10 fps higher, and the average framerate is about 7 fps higher. I would say that is more than perceivable. Hate to break it to you but you are running out of bullcrap to spew. 

You are still not quantified. We need to know how much vram the game was using in order to verify if the game was running into a Vram bottleneck on the 3GB 780. If not then the extra Vram on the 6GB 780 was all but useless in that game as we already know that Vram count does not make a GPU faster.

http://www.guru3d.com/articles_pages/crysis_3_graphics_performance_review_benchmark,8.html

[BiG StroOnZ]

You are still not quantified. We need to know how much vram the game was using in order to verify if the game was running into a Vram bottleneck on the 3GB 780. If not then the extra Vram on the 6GB 780 was all but useless in that game as we already know that Vram count does not make a GPU faster.

http://www.guru3d.com/articles_pages/crysis_3_graphics_performance_review_benchmark,8.html

 

Firstly, the only one who said anything about extra VRAM making a GPU faster is you. 

 

Secondly, taken directly from the same review:

 

 

 

Please keep in mind though that memory utilization will vary a bit per level, it's not a very precise thing to measure. 

 

So according to your logic, "to quantify the situation" we would need to know exactly what levels were played that resulted in those numbers. Then we would also need to see if different levels used different amounts of VRAM. Then we would need to see all the VRAM results of each level. That  would "quantify" your information as valid. Unfortunately, we don't have these results. So we can basically chalk up the above results as useless. Since they themselves did not "quantify" them 

 

But what we do have is Hard OCP, who is completely unbiased when it comes to testing out graphics cards with extra VRAM. When they themselves did a review on a 2GB 670 vs 4GB 670 not to long ago and concluded the following: 

 

 

 

We've given the ASUS GeForce GTX 670 DirectCU II 4GB video cards plenty of opportunities to prove to us why we would need 4GB of VRAM on these. Unfortunately, these have come short of proving this

 

http://www.hardocp.com/article/2013/01/14/asus_geforce_gtx_670_directcu_ii_4gb_sli_review/9

 

That was nearly a year ago. Unfortunately times have changed with their 3GB 780 vs 6GB 780 results. They aren't going to skew results. They just don't do that. The benchmark doesn't lie, this is coming from people who tried once to give a card a chance with extra VRAM. Today however, different story, the results from the 3GB 780 to 6GB 780 are vastly different from their 2GB 670 vs 4GB 670 results. They found a benefit in the extra VRAM, and concluded so. 

 

If you want to keep grinning  your teeth and ignore the data. That's your prerogative, but you have been proven wrong too many times to count. Every single time you get proven wrong, you change your tune to the next topic. Dropping your previous argument that you made, and starting up a new one...