-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Is 1GB VRAM enough?
- Thread starter Cleeecooo
- Start date
+1
So you just ignore the fact that your 1GHz 560 1GB SLI got pwn'ed by my stock 6950 2GB CF in Metro?
People have no problem overclocking 6950s to 900MHz and some can manage 1GHz. I don't see where your claim of 6950s unable to oc comes from. When OC'ed, they even beat 580 SLI in Shogun 2.
Shogun 2 is AMD optimised and runs worse on all Nvidia hardware. This is not a Vram issue, as it has been addressed on the Nvidia forums.
Also metro is a terribly coded game, and a horrible example for comparing graphics card performance.
Double post.
^^ I see "Run 1 (Frontline)" above. How many times did you attempt to produce a plot with less lag spikes?
You should work harder and eliminate large gaps like between second 48 and 50. Also ignore more spikes like second 3, 35, 38, 41, 42, 43, 43.5, 44, 44.5, 47.5, 48, 48.5, 50 etc.
When you start to realize that you need a statistical measure to do the comparison, you should understand that ideally you'll need to run the benchmark several times on each setup, and calculate the VARIANCE of each case, which defines the smoothness of the gameplay. Strictly, get a program to count "how many gaps are there larger than, say, 33ms (or 50ms, for mid ranged cards like 560 Ti SLI)". Don't just pick blue arrows by eye. This is just a rough illustration for people who can understand the idea.
Also, I hate to repeat, but Metro 2033 benchmark does NOT rotate the camera quickly or switch cutscene, and hence there are a lot more lag spikes in the real-game play than in this small benchmark. I pick this benchmark merely because it makes plots based on frametimes, not fps numbers simply averaged over 1-second time intervals smoothing out the lag spikes.
Don't report to me that the max vram you used was 1000/1024. You should let Jen-Hsun Huang know it. Obviously he's wasting money using more than 1GB on some cards, and he should cut the cost on Kepler cards.
So let's see who's gonna trust posts made by a famous ترول from السعودية on the internet? Probably those who'd never used more than 1GB and like to believe? Or is this joke tosh thread just for entertainment purposes?
Yes at 1200p and lower.
Irrefutable proof that VRAM over 1 Gb desnt affect performance at 1920x1200 (according to harmony's methods of proof finding).
For more performance, you need faster GPUs, not more Vram.
^^ I see "Run 1 (Frontline)" above. How many times did you attempt to produce a plot with less lag spikes?
You should work harder and eliminate large gaps like between second 48 and 50. Also ignore more spikes like second 3, 35, 38, 41, 42, 43, 43.5, 44, 44.5, 47.5, 48, 48.5, 50 etc.When you start to realize that you need a statistical measure to do the comparison, you should understand that ideally you'll need to run the benchmark several times on each setup, and calculate the VARIANCE of each case, which defines the smoothness of the gameplay. Strictly, get a program to count "how many gaps are there larger than, say, 33ms (or 50ms, for mid ranged cards like 560 Ti SLI)". Don't just pick blue arrows by eye. This is just a rough illustration for people who can understand the idea.
Also, I hate to repeat, but Metro 2033 benchmark does NOT rotate the camera quickly or switch cutscene, and hence there are a lot more lag spikes in the real-game play than in this small benchmark. I pick this benchmark merely because it makes plots based on frametimes, not fps numbers simply averaged over 1-second time intervals smoothing out the lag spikes.
Don't report to me that the max vram you used was 1000/1024. You should let Jen-Hsun Huang know it. Obviously he's wasting money using more than 1GB on some cards, and he should cut the cost on Kepler cards.
So let's see who's gonna trust posts made by a famous ترول from السعودية on the internet? Probably those who'd never used more than 1GB and like to believe? Or is this joke tosh thread just for entertainment purposes?
Last edited:
If you're so bothered about Nvidia's cards and lack of Vram, why dont you try trolling on this forum instead and see how people react to what you say?
http://forums.nvidia.com/index.php?
Nope. Metro 2033 is better optimised to run on your GPUs than mine. The performance difference is NOT due to Vram.
So you just ignore the fact that there is absolutely no difference between a 1 Gb and 2 Gb card with the SAME GPU on the SAME PC in any game currently out there at 1200p resolution?
Put a pair of 1 Gb 6950s into your PC and run Metro Benchmark at 1200p. There will be absolutely no difference at all, just like there isnt between the 1 Gb and 2 Gb GTX 560s.
Nope. Metro 2033 is an nVidia's game. Proof: a pair of GTX580 1.5GB gets better results than 6950 2GB CF.
I thought about buying a pair of 6950 1GB for the test, then return under DSR, but I think that is too lame so I gave up that idea. If you can fund me I'm more than happy to do that.
GTX 580 beating a 6950 isnt a proof of optimization, the GTX 580 is a more powerful card.
Hence 6950s beating a GTX 580 (as in Shogun 2) is proof of optimization.
I really would hope that a pair of 6950s beat a pair of GTX 560s, since they actually cost quite a bit more.
Also you ignored the actualy topic again, the speed differences are due to GPU, not Vram.
Why the fudge would I need to fund you? You're a professional super rich GPU reviewer who simultaneously owns two 6950s and two GTX 580s. Surely you can fund yourself a pair of 1 Gb 6950s if constantly blabbering on about Vram is so important to you. If you cant even make a valid and fair comparison yourself, then all of your charts, graphs and comparisons are lies, and you are a faker.
(And you should also be funding everysingle person that uses this forum a pair of 3 Gb GTX 580s, because according to you nothing else is good enough to play all the latest games fully maxed out).
Also if GTX 560s are so much worse than your 6950s, then why do you actually recommend other people to buy them? You dont even have the courage to recommend your own cards, nor 2 Gb ones so surely you are a liar.
Right here you said you would actually buy the 1 Gb GTX 560, so obviously its better than your 6950s since you have no courage to recommend others to buy what you have:
http://forums.overclockers.co.uk/showthread.php?t=18303903
(Last couple of lines are using Harmony logic and reasoning, not my own).
Hence 6950s beating a GTX 580 (as in Shogun 2) is proof of optimization.
I really would hope that a pair of 6950s beat a pair of GTX 560s, since they actually cost quite a bit more.
Also you ignored the actualy topic again, the speed differences are due to GPU, not Vram.
Why the fudge would I need to fund you? You're a professional super rich GPU reviewer who simultaneously owns two 6950s and two GTX 580s. Surely you can fund yourself a pair of 1 Gb 6950s if constantly blabbering on about Vram is so important to you. If you cant even make a valid and fair comparison yourself, then all of your charts, graphs and comparisons are lies, and you are a faker.
(And you should also be funding everysingle person that uses this forum a pair of 3 Gb GTX 580s, because according to you nothing else is good enough to play all the latest games fully maxed out).
Also if GTX 560s are so much worse than your 6950s, then why do you actually recommend other people to buy them? You dont even have the courage to recommend your own cards, nor 2 Gb ones so surely you are a liar.
Right here you said you would actually buy the 1 Gb GTX 560, so obviously its better than your 6950s since you have no courage to recommend others to buy what you have:
http://forums.overclockers.co.uk/showthread.php?t=18303903
(Last couple of lines are using Harmony logic and reasoning, not my own).
Last edited:
I kept recommending 6950 2GBs all the time until I had been exhausted by you
Just don't want to get more trolls from people like you lolSeriously, for single setup, it is fine to go for a GTX560 Ti 1GB in most cases. What I discourage is to put 1GB cards into SLI/CF.
Last edited:
Those charts are a whole generation out of date. Also hotlinking isnt allowed.
Even though two GTX 560s are far superior to one GTX 580 for the same price.
I've been using 1 Gb cards in SLI/CF for the last few generations and havnt had a problem with then (No lag, hitching, microstutter or whatever other fancy myths that people keep on inventing without any proof).
Last edited:
But posting results from hardware produced AFTER the game was developed is irrelevant, because the game code cannot be optimized for hardware designed AFTER the game. Only the more recent things can optimize for previous things. This is simple logic. Unless you mean the game developers could witness the future?
I was looking for current generation comparisons, and your examples do not prove that Vram makes any difference, it proves that faster GPUs are faster, and slower GPUs are slower.
Why do you keep on posting so many charts and graphs just to prove the obvious?
Why do you keep on posting so many charts and graphs just to prove the obvious?
Last edited:
That was just to prove Metro 2033 was optimized for nVidia and that's all.
I have already posted comparison of different vram capacity on current gen many times and I don't want to do that again. It is extremely difficult to beg GTX560 Ti users of different vram size to run games eating more than 1GB vram with the same benchmark scenes, using Fraps or whatever to make a record of the Frametimes and upload. Even if I manage that, the results would be from different machines, with different CPU/clock etc.
I have already posted comparison of different vram capacity on current gen many times and I don't want to do that again. It is extremely difficult to beg GTX560 Ti users of different vram size to run games eating more than 1GB vram with the same benchmark scenes, using Fraps or whatever to make a record of the Frametimes and upload. Even if I manage that, the results would be from different machines, with different CPU/clock etc.
Then why dont you make the comparisons yourself before constantly whining that 1 Gb Vram isnt enough? Why do you make baseless comments when you cant even write your own review with unbiased results?
Every professional review you can read on this topic concludes that more than 1 Gb Vram does not improve performance at 1200p or lower. So what exactly makes you think that you are right, and all those reviews are wrong?
Get some 1 Gb and 2 Gb cards of the same GPU and same clock speeds, compare them, and demonstrate to everyone how much better 2 Gb is over 1 Gb at 1200p before making things up when you have no personal experience with using the latest 1 Gb cards.
Again, you only tested in a limited set of games, and you have no proof of smoothness in a statistical measure, i.e. the number of lag spikes with time gap >33ms, especially when you quickly rotate the camera or when the game switches cutscenes.
Because I have yet to see any 1GB cards in SLI/CF being able to achieve very few lag spikes in the metro 2033 benchmark. 560 Ti 2GB SLI got less lag spikes than your heavily overclocked 560 Ti 1GB SLI was already a good proof. You should counter me with better selected pictures, but it won't hide the very first picture you posted.
then why not prove your point by getting some 6950 1gb's?