When did the format of reviews change so the conclusion is on the front page? 9 times out of 10 I skip the content and just read the conclusion, so was caught out today :p
Originally Posted by Dreaming When did the format of reviews change so the conclusion is on the front page? 9 times out of 10 I skip the content and just read the conclusion, so was caught out today :p
They did it the wrong way round with the NAS box reviews too.
Still looking forward to seeing some overclocked 6 series benchmarks to see if either card when overclocked is as good as an overclocked 460 1GB.
Originally Posted by Dreaming When did the format of reviews change so the conclusion is on the front page? 9 times out of 10 I skip the content and just read the conclusion, so was caught out today :p
I think this new format benefits those of us who rarely read the full review. All we need is right there on the first page. I approve.
However one thing I don't particularly agree with (and this is the standard format for almost every review site so i'm not pointing the finger at Bit-Tech here) is the way that is is apparently necessary to rerun every single benchmark for an aftermarket card. What I want to know about in an aftermarket card is how good the cooler is, how quit it runs, and how well does it manually overclock. I really feel that unless there is a significant factory overclock, it wastes a lot of time runnign all the benchmarks and you can always guess that the overclock will make just a few FPS difference. I'd rather see you guys push the card to the limit yourselves and then see how this increases performance from the stock card in maybe one benchmark.
Originally Posted by Snips Will you be doing the same benchtest as the GTX460 got in CustomPC isue 085? they reviewed all manufacturers and came up with the perfect card.
I agree with this, I'd rather they didn't so extensively test each single aftermarket card and instead just listed heat, noise and overclocking (see above) and an ultimate conclusion as to which are the ones to get.
So the last component I need to get for my new toy is the GPU. I have a budget of £150 and effectively this gives me either a 460 or a 6850.
I keep looking at benchmarks and the truth is there is bugger all difference between these two cards in any meaningful way.
Looks like the buying decision will be down to whoever has some sort of offer or value added pack or has a pretty finish on the fan casing. Not really the best way of deciding.
Originally Posted by ADJB So the last component I need to get for my new toy is the GPU. I have a budget of £150 and effectively this gives me either a 460 or a 6850.
I keep looking at benchmarks and the truth is there is bugger all difference between these two cards in any meaningful way.
Looks like the buying decision will be down to whoever has some sort of offer or value added pack or has a pretty finish on the fan casing. Not really the best way of deciding.
Your right, it's a very tough call at the moment at that price point.
However if I had to make the choice I'd go for the GTX 460 as it overclocks much better, comparing thesetwo pages the GTX 460 gained over 20% performance from overclocking whereas the 6850 gained just over 10%. They also said in the review that that particular GTX 460 overclocked lower than other models (note that i'm not necessarily recommending those versions of the card, they're just very close to the reference and are a good indicator of overclocking performance).
Also SLI seems to perform marginally better than Crossfire at the moment, if you want to get another in the future, however it doesn't look like that'll last much longer as Crossfire continues to improve.
Originally Posted by ADJB So the last component I need to get for my new toy is the GPU. I have a budget of £150 and effectively this gives me either a 460 or a 6850.
Save up just a little more money and get the HD 6870. But if I had to choose between the GTX 460 1gb and the HD 6750 then I would get the GTX 460 because of Physx and a slight edge in performance when Tessellation is applied in games such as Metro 2033 which needs to be on all DX11 Video Card reviews.
the conclusion of the review is that the 6850 is on a par with the GTX 460 1GB. That's just not true, the only way that 6850 beats the GTX 460 1GB is on just cause 2 which is obviously an AMD biased game (the 5770 beats a 460 786mb!). Everywhere else the 460 1gb beats it, also imagine a DX11 game like metro with 8xAA on, the 6850 would be dreadful. Also the 460 1gb isn't even an overclocked one, it gains about 25% performance compared to the 6850's 10% so would be even better, the 6850 isn't on a par, it's definately a step down from the 460.
I think the really interesting information to be had here is in comparing the performance of the 6870 to that of the 5770, the former being some 50%+ of the later.....the 6870 is the top of the Barts chip range of the 6000 series and 5770 is the top of the Juniper chip range of the 5000 series....both the Juniper and Bart chips are second from top in performance in their given series.
If that performance ratio carries forward when comparing the Cayman and Cypress chips and 6970 ends up with 50%+ more performance than 5870 then it will be one hell of a card.....don't think Nvidias GTX580 will be enough to answer such a performance increase.
you know to be honest I think starting with the conclusion isnt a bad idea really, I tend to jump to the end, get an idea of what they think of the card and check out the ratings and then go through and look at the test results to which ever test I wanted to look at. This could work.
We’ve not seen many Radeon HD 6850 1GB cards use the reference cooler, probably because it was loud and didn’t cool very well.
I just skimmed through the original HD6850 review again and couldn't find any mentioning of the stock cooler's loudness. Actually, the closest thing I could find was this:
Quote:
Despite the lower power consumption though, ATI has obviously made the decision to run the HD 6850 1GB hotter rather than noisier.
To me that doesn't sound like the stock HD6850 cooler is noisy.
I'd really like for bit-tech to focus a bit more on graphics card noise (or the lack thereof) as this becomes increasingly more important to many users.
What is the point in running a stock clocked card through all of the tests when you already have the results from a stock clocked (reference) card included in said review?
More over why bother finidng out how far a HD6850 can clock and then NOT including any overclocking results in the tests at all? To me this just seems a waste of time.
The same thing happened with the Asus HD5870 VT. Bit reviewed the card at reference clocks where suprise suprise it scored identical to the reference card. Only one result was shown for its overclock when in reality the card should have been put through the tests overclocked and the reference card should have acted as a proxy for stock clocks.
Sorry to sound like I am moaning but really this 'review' doesn't realy tell us anything we don't already know.
Been thinking of buying one of these for a new build however, the GTX 460 still seems like the way to go if im right in thinking the benchmarked 1GB GTX 460 is stock? If so, surely the msi or Gigabyte 460 would just completely embarrass the 6850?
+1 for keeping the conclusion at the start of the review.
Unless I'm actively looking for new hardware I don't usually bother looking through all the graphs, just a quick overview to keep up.
Comments 1 to 25 of 29
ReplyThey did it the wrong way round with the NAS box reviews too.
Still looking forward to seeing some overclocked 6 series benchmarks to see if either card when overclocked is as good as an overclocked 460 1GB.
Which is why they did it.....
Very funny BT.....
Correct on one of those! :)
XFX has some sense, and put in normal sized connectors.
I think this new format benefits those of us who rarely read the full review. All we need is right there on the first page. I approve.
However one thing I don't particularly agree with (and this is the standard format for almost every review site so i'm not pointing the finger at Bit-Tech here) is the way that is is apparently necessary to rerun every single benchmark for an aftermarket card. What I want to know about in an aftermarket card is how good the cooler is, how quit it runs, and how well does it manually overclock. I really feel that unless there is a significant factory overclock, it wastes a lot of time runnign all the benchmarks and you can always guess that the overclock will make just a few FPS difference. I'd rather see you guys push the card to the limit yourselves and then see how this increases performance from the stock card in maybe one benchmark.
I agree with this, I'd rather they didn't so extensively test each single aftermarket card and instead just listed heat, noise and overclocking (see above) and an ultimate conclusion as to which are the ones to get.
O_o
But then I saw that there were more pages on the review... Nice prank BT :P
Will be interesting to see some Vapor-x versions of the 6800 series to see how they perform.
I keep looking at benchmarks and the truth is there is bugger all difference between these two cards in any meaningful way.
Looks like the buying decision will be down to whoever has some sort of offer or value added pack or has a pretty finish on the fan casing. Not really the best way of deciding.
Your right, it's a very tough call at the moment at that price point.
However if I had to make the choice I'd go for the GTX 460 as it overclocks much better, comparing these two pages the GTX 460 gained over 20% performance from overclocking whereas the 6850 gained just over 10%. They also said in the review that that particular GTX 460 overclocked lower than other models (note that i'm not necessarily recommending those versions of the card, they're just very close to the reference and are a good indicator of overclocking performance).
Also SLI seems to perform marginally better than Crossfire at the moment, if you want to get another in the future, however it doesn't look like that'll last much longer as Crossfire continues to improve.
Save up just a little more money and get the HD 6870. But if I had to choose between the GTX 460 1gb and the HD 6750 then I would get the GTX 460 because of Physx and a slight edge in performance when Tessellation is applied in games such as Metro 2033 which needs to be on all DX11 Video Card reviews.
I did the same thing, looks like now I can log in and get the rating on the front page.
If that performance ratio carries forward when comparing the Cayman and Cypress chips and 6970 ends up with 50%+ more performance than 5870 then it will be one hell of a card.....don't think Nvidias GTX580 will be enough to answer such a performance increase.
I'd really like for bit-tech to focus a bit more on graphics card noise (or the lack thereof) as this becomes increasingly more important to many users.
More over why bother finidng out how far a HD6850 can clock and then NOT including any overclocking results in the tests at all? To me this just seems a waste of time.
The same thing happened with the Asus HD5870 VT. Bit reviewed the card at reference clocks where suprise suprise it scored identical to the reference card. Only one result was shown for its overclock when in reality the card should have been put through the tests overclocked and the reference card should have acted as a proxy for stock clocks.
Sorry to sound like I am moaning but really this 'review' doesn't realy tell us anything we don't already know.
Unless I'm actively looking for new hardware I don't usually bother looking through all the graphs, just a quick overview to keep up.
-
« Previous
-
1
-
2
-
Next »
Discuss in the forums