• 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.

GTX 670 and 2560x1440

Did you run out of VRAM?

0001ts8g.gif
 
must be great feedback... because a Titan has more VRAM, more transistors and the same basic bandwidth so by this logic the OP should go with Titan...

I think he'd be better off with a low profile 5450 personally Andy.

Why post that? It is an honest question and one that is quite important for future buyers. If he ran out of VRAM at 1440P, then that makes a big difference to recommendations.

He said it runs better because of the extra bandwidth. I have no reason to doubt what he said. He said the 670 runs 1440p fine but the 7970 runs it better. I see no problem here. I only posted good feedback as he's owned both cards unlike me so i can't comment on the 670 performance. Take the issue up with him plox.

EDIT

To clarify, im not saying or thinking that he's saying it runs faster because of 1gb of extra vram. More the fact that 256 bit bus vs 384 bit bus.
 
Last edited:
My MSI GTX670 ran fine with that res but I changed to a HD7970 & it run's better thanx to the bandwith & extra Vram !

He said it runs better because of the extra bandwidth. I have no reason to doubt what he said. He said the 670 runs 1440p fine but the 7970 runs it better. I see no problem here. I only posted good feedback as he's owned both cards unlike me.

That is cool Matt but the part that is bold and underlined has my attention. If he ran out of VRAM, this could be something of interest to all of us who recommend GPU's. Mine was a genuine question and your image was a bit pointless and not sure why you added it?
 
That is cool Matt but the part that is bold and underlined has my attention. If he ran out of VRAM, this could be something of interest to all of us who recommend GPU's. Mine was a genuine question and your image was a bit pointless and not sure why you added it?

Because the other day you said to someone and i quote "The vram debate has been done to death, im not doing it again". Something along those lines. ;)
 
Because the other day you said to someone, and I quote "The vram debate has been done to death, i'm not doing it again". Something along those lines. ;)

Ermm :confused:

If someone has run out of VRAM, that has some relevance to the question I asked. Surely you have the same interest for future reference?
 
Ermm :confused:

If someone has run out of VRAM, that has some relevance to the question I asked. Surely you have the same interest for future reference?

Ok Greg lets get back on topic.

There is a possibility i suppose he was using all or most of his 2gb limit at 1440p though its still enough for the majority of titles id imagine. I expect he was talking more about the memory bandwidth as its quite a bit larger on the 7970. Either way lets wait for him to clarify.
 
Ok Greg lets get back on topic.

There is a possibility i suppose he was using all or most of his 2gb limit at 1440p though its still enough for the majority of titles id imagine. I expect he was talking more about the memory bandwidth as its quite a bit larger on the 7970. Either way lets wait for him to clarify.

how would he know if the memory bandwidth was giving better results?
he hasn't specified what overclocks he was running on either card and he hasn't specified what games he was playing, what settings, and what FPS results he got on either

he has basically just said that he likes the 7970 more than the 670 and then appended some reasons which bear no relation to what he actually said... memory usage and bandwidth are only relevant if you can quantify why they are giving you an advantage... the 7970 is the "top" single card from AMD's stable, and the 670 is the 2nd or 3rd in Nvidia's depending on your point of view... I would naturally expect it to be as fast or faster based on the specs of the GPU itself... to make claims about it's other attributes you would need to show that you had done some testing that specifically called out the items mentioned and shown that they have an effect

such as, running the 2 cards at 1920x1080 and then again at 2560x1440... showing a bigger disparity at 2560x1440 you could argue that the memory speed, bandwidth or even amount is having an effect (though as we all know, memory amount being an issue usually leads to quite dramatic fall offs in framerate)

just saying "i personally prefer this card of the 2" and then appending some made up reasons for that preference is not good feedback

using the same logic to recommend a 7970 would lead you to recommend a Titan even more

and seeing as the OP has stated that his choice would either be Titan or another 670, turning this thread in to yet another AMD vs. Nvidia war is a massive waste of time
 
how would he know if the memory bandwidth was giving better results?
he hasn't specified what overclocks he was running on either card and he hasn't specified what games he was playing, what settings, and what FPS results he got on either

he has basically just said that he likes the 7970 more than the 670 and then appended some reasons which bear no relation to what he actually said... memory usage and bandwidth are only relevant if you can quantify why they are giving you an advantage... the 7970 is the "top" single card from AMD's stable, and the 670 is the 2nd or 3rd in Nvidia's depending on your point of view... I would naturally expect it to be as fast or faster based on the specs of the GPU itself... to make claims about it's other attributes you would need to show that you had done some testing that specifically called out the items mentioned and shown that they have an effect

such as, running the 2 cards at 1920x1080 and then again at 2560x1440... showing a bigger disparity at 2560x1440 you could argue that the memory speed, bandwidth or even amount is having an effect (though as we all know, memory amount being an issue usually leads to quite dramatic fall offs in framerate)

just saying "i personally prefer this card of the 2" and then appending some made up reasons for that preference is not good feedback

using the same logic to recommend a 7970 would lead you to recommend a Titan even more

and seeing as the OP has stated that his choice would either be Titan or another 670, turning this thread in to yet another AMD vs. Nvidia war is a massive waste of time

Spot on Andy.
 
how would he know if the memory bandwidth was giving better results?

Improvement in fps maybe?

You'd need to ask him not me.

and seeing as the OP has stated that his choice would either be Titan or another 670, turning this thread in to yet another AMD vs. Nvidia war is a massive waste of time

I only said good feedback as he was giving an opinion having owned and used both cards at 1440p. Something which neither you, me or Greg has done. If that was starting an AMD vs Nvidia war then please excuse me.
 
I only said good feedback as he was giving an opinion having owned and used both cards at 1440p. Something which neither you, me or Greg has done. If that was starting an AMD vs Nvidia war then please excuse me.

I have owned both 670(s) and a 7950 and tried them at 2560x1440 and 1920x1080

if anything I would say the 7950 when fully overclocked was marginally slower than a single 670, but then my 670 was on water so it's not entirely a fair comparison

the 7950 has the benefit of both the wider bus and the larger amount and yet this didn't cause it to be the clear winner, even at 1440p

from what I've seen of comparisons (including my own) of 7950 / 670 / 7970 / 680, they are all pretty much of a muchness, there are some small differences game to game, but not anything you would be able to see without an FPS counter running

that is why I question when someone says any one of the four is clearly better than any other one of the four, and why I question when someone says "bigger bus and more VRAM must be better" when my own testing clearly shows that this isn't the case
 
Last edited:
I have owned both 670(s) and a 7950 and tried them at 2560x1440 and 1920x1080

if anything I would say the 7950 when fully overclocked was marginally slower than a single 670, but then my 670 was on water so it's not entirely a fair comparison

the 7950 has the benefit of both the wider bus and the larger amount and yet this didn't cause it to be the clear winner, even at 1440p

from what I've seen of comparisons (including my own) of 7950 / 670 / 7970 / 680, they are all pretty much of a muchness, there are some small differences game to game, but not anything you would be able to see without an FPS counter running

that is why I question when someone says any one of the four is clearly better than any other one of the four

I don't think he said that. I think he said the 670 ran 1440p fine but the 7970 ran it better. From what i see in benchmarks that sounds about right. Rusty even confirmed as much. I don't think anyone is saying the 670 is a bad card or can't run 1440p well at all. Just that the 7970 provides better performance which every benchmark ive looked at confirms.

How long ago did you own a 7950? Was it before the performance increases?
 
I don't think he said that. I think he said the 670 ran 1440p fine but the 7970 ran it better. From what i see in benchmarks that sounds about right. Rusty even confirmed as much. I don't think anyone is saying the 670 is a bad card.

How long ago did you own a 7950? Was it before the performance increases?

Up until a couple of weeks ago
when my Titan's arrived I put the 670's in my secondary machine instead

he gave specific reasons why the 7970 ran faster though, without any info to back up that claim... the 7970 is a much bigger chip than the 670, so you would expect it to be a little bit faster at least, but without specific testing on the effects of memory bandwidth (like over and under clocking the memory for example), it's impossible to say what the cause of the extra FPS is

memory bandwidth is very easy to calculate based on the bus size and memory speed, you could overclock the RAM on the 670 and underclock the ram on the 7970 so that they match, and then compare FPS, that would tell you the effect that memory bandwidth has in a specific game at a specific resolution

I don't really get why you are being so obtuse... you argue at much greater depth in many threads and now you are being deliberately simple to try and claim ignorance of a statement you previously supported
 
Last edited:
Up until a couple of weeks ago
when my Titan's arrived I put the 670's in my secondary machine instead

he gave specific reasons why the 7970 ran faster though, without any info to back up that claim... the 7970 is a much bigger chip than the 670, so you would expect it to be a little bit faster at least, but without specific testing on the effects of memory bandwidth (like over and under clocking the memory for example), it's impossible to say what the cause of the extra FPS is

memory bandwidth is very easy to calculate based on the bus size and memory speed, you could overclock the RAM on the 670 and underclock the ram on the 7970 so that they match, and then compare FPS, that would tell you the effect that memory bandwidth has in a specific game at a specific resolution

The 7970ghz is definitely faster than a 670 Andy. Recent reviews of current titles show it to be faster than a 680 as well.

Pulled from a titan review - 7970Ghz anywhere from 5-45% faster than a 680 at 1440p in the latest listed titles. It should note than AMD released performance improvements for farcry3 and hitman after this review went live. There's some gaming evolved titles missing from that list as well Tomb Raider etc.


Pvvb3Sr.jpg



You really need to be taking this up with him not me. I only made a comment on his opinion having owned and used both. If you think he's lying and the 670 is the faster of the cards at 1440p take it up with him. It goes against every recent benchmark ive seen of the current titles though. I agree that a person might not notice the performance difference though without a fps counter.
 
Last edited:
You really need to be taking this up with him not me. I only made a comment on his opinion having owned and used both. If you think he's lying and the 670 is the faster of the cards at 1440p take it up with him. It goes against every recent benchmark ive seen of the current titles though. I agree that a person might not notice the performance difference though without a fps counter.

I didn't say he was lying in saying that a 7970 was faster than a 670, stop putting words in my mouth.

What I said was that a) you probably wouldn't notice without an FPS counter, which you agree with, and that b ) memory bandwidth and amount have little to no relevance to that performance difference

you told him this was good feedback, and you are the one arguing with me so you are the one I'm responding too... if you have no input on the subject then stop throwing unrelated charts at me :D

even on your chart I am seeing differences of 1%-15%, which is inline with my experience, the only outlier is Dirt showdown, which I've never played to can't comment, but great, go ahead, pick up that one title and run with it as outright proof that a 7970 is equal in performance to a Titan if that is what you wish

again, from my experience of having owned a 670 and a 680, they are both within single digit FPS of each other, in other words, not noticeable
 
Last edited:
I didn't say he was lying in saying that a 7970 was faster than a 670, stop putting words in my mouth.

What I said was that a) you probably wouldn't notice without an FPS counter, which you agree with, and that b ) memory bandwidth and amount have little to no relevance to that performance difference

you told him this was good feedback, and you are the one arguing with me so you are the one I'm responding too... if you have no input on the subject then stop throwing unrelated charts at me :D

I said 'good feedback' based on the fact that he found the 7970 to provide better performance than a 670 at 1440p. I said i assumed he was thinking it could be because of the wider memory bus but that's just a guess on my part, as before you need to take it up with him . I was just commenting on his opinion having used both cards. Nothing more, nothing less. I even replied to Greg explaining my good feedback was nothing to do with his mention of vram. User opinions are just that opinions but its still interesting to hear them.

May i suggest you drop him a trust titled "Memory bandwidth and amount have little to no relevance to that performance difference". If you want to debate it further send him a trust not me. Thanks Andy.
 
Last edited:
I said 'good feedback' based on the fact that he found the 7970 to provide better performance than a 670 at 1440p. I said i assumed he was thinking it could be because of the wider memory bus but that's just a guess on my part, as before you need to take it up with him . I was just commenting on his opinion having used both cards. Nothing more, nothing less. I even replied to Greg explaining my good feedback was nothing to do with his mention of vram. User opinions are just that opinions but its still interesting to hear them.

May i suggest you drop him a trust titled "Memory bandwidth and amount have little to no relevance to that performance difference". If you want to debate it further send him a trust not me. Thanks Andy.

you said "good feedback" because it agreed with your pre-conceived idea that a 7970 is significantly faster than a 670

it isn't good feedback when it contains no quantifiable information upon which this opinion has been created

here is mine;

having owned a 680, 670 and 7950, and based on 7970 user reviews on here, I find them all to be broadly equal when fully overclocked, memory bandwidth and amount seems to have very little effect on performance at 2560x1440

is that also good feedback? or does it offend your AMD sensibilities?
 
Last edited:
Back
Top Bottom