check this out... hope its true...
http://www.fudzilla.com/index.php?option=com_content&task=view&id=12486&Itemid=65
http://www.fudzilla.com/index.php?option=com_content&task=view&id=12486&Itemid=65
cyber_jockey
Distinguished
- Jul 6, 2008
- 397
- 0
- 18,780
Meh gtx 285 is almost 30 percent faster then a 4870. I doubt the 4890 can beat the gtx 285 in raw performance but just like the 4870 gtx 280 battle the people might choose the 4890 over a gtx 285. I sense another price performance war.
spathotan
Distinguished
- Nov 16, 2007
- 2,390
- 0
- 19,780
Who cares? I would hope AMD does if they plan on staying in business. It dosent have to beat it to stay in business, but it needs to just be a legit improvement over the 4870, which it will be. Claiming it might beat a GTX 285 like that fud link does is stupidity.
spathotan
Distinguished
- Nov 16, 2007
- 2,390
- 0
- 19,780
L1qu1d :
They should be^
It would be easier if Sli and Crossfire were 2nd to Single GPus
Either make some decent drivers (pointed to both companies) or stop making quad sli or tri sli
It would be easier if Sli and Crossfire were 2nd to Single GPus
Either make some decent drivers (pointed to both companies) or stop making quad sli or tri sli
My personal experience with Crossfire also has me leaning towards this. The support is just not there.
spathotan
Distinguished
- Nov 16, 2007
- 2,390
- 0
- 19,780
Dosent matter, the 4870 and GTX 260 are competitors, not the GTX 280 or 285.
Its nothing to debate, the debate isnt even there. The 4890 will just overtake the GTX 260 all together. Still too far away from the GTX 280/5, as it should since it remains competitive with the GTX 260.
Its nothing to debate, the debate isnt even there. The 4890 will just overtake the GTX 260 all together. Still too far away from the GTX 280/5, as it should since it remains competitive with the GTX 260.
cyber_jockey
Distinguished
- Jul 6, 2008
- 397
- 0
- 18,780
If its 5 percent slow then the gtx 280 and 30 percent cheaper I dont know bout you guys but thats a sweet deal .
cyber_jockey
Distinguished
- Jul 6, 2008
- 397
- 0
- 18,780
If the 4890 is 5 percent slower then the gtx280 and 50 percent cheaper I dont know bout you guys but thats a sweet deal to me.
mactronix
Judicious
- Sep 17, 2006
- 7,928
- 1
- 32,115
What matters is its performance in its price segment and that will be the best.
Also dont forget that AMD were all set to roll with 40nm chips but are releasing this on a 55nm leaving a die shrink revision for later with improvements to come.
Mactronix
Also dont forget that AMD were all set to roll with 40nm chips but are releasing this on a 55nm leaving a die shrink revision for later with improvements to come.
Mactronix
- Apr 29, 2006
- 29,241
- 437
- 93,590
% matters, a lot.
The link isn't really worth reading. The only interesting thing in it is that it claims that the notebook versions are shipping. This does put the chip a lot closer to being finished. We still have no idea what internal changes have happened that might make it C4C faster then the 4970. We still have no idea if there are more SP or not, or high much higher it might be clocked. The only thing that seems set in stone about this chip is that it will be on the 55nm node. Dropping to 40nm probably won't do much, other then make it cheaper for them to produce. (unless there are even more tweaks they can make.)
The link isn't really worth reading. The only interesting thing in it is that it claims that the notebook versions are shipping. This does put the chip a lot closer to being finished. We still have no idea what internal changes have happened that might make it C4C faster then the 4970. We still have no idea if there are more SP or not, or high much higher it might be clocked. The only thing that seems set in stone about this chip is that it will be on the 55nm node. Dropping to 40nm probably won't do much, other then make it cheaper for them to produce. (unless there are even more tweaks they can make.)
cyber_jockey
Distinguished
- Jul 6, 2008
- 397
- 0
- 18,780
The very same reason why many people choose the 4870x2 over a gtx295. Performance goes and in hand with price.
- Apr 29, 2006
- 29,241
- 437
- 93,590
Stranger, we might be talking about two different things here. Some people say "big deal, you got 30FPS and I got 20FPS, so I'm only 10FPS slower." The problem is this is absolute values, and totally ignores percentage. His "not bad" 20FPS is actually 33% slower, which sounds a lot worse, because at these low FPS it IS worse. Same is true if you are talking about 40FPS instead of 50FPS. Might be only 10 frames, but its also 20% slower.
If the 4970 is going to be 30% faster then the 4870, then we know it will come close to the GTX280. You can't just say it will add 10 FPS to your games, as that lacks any real detail. (what res is the first that comes to mind.) By saying 10% slower then the GTX280, but 30% cheaper, we have some idea about its performance.
If the 4970 is going to be 30% faster then the 4870, then we know it will come close to the GTX280. You can't just say it will add 10 FPS to your games, as that lacks any real detail. (what res is the first that comes to mind.) By saying 10% slower then the GTX280, but 30% cheaper, we have some idea about its performance.
mactronix
Judicious
- Sep 17, 2006
- 7,928
- 1
- 32,115
@ 4745454bWell to be fair no one said FPS was a better way of doing it either, it isnt. You get the same issue, ie what system ? what settings ?
People who know a little about computers will understand that these percentages and FPS being thrown about are just ball park. People who dont know so much and come for advice could get misslead and take some of these figures as gospel.
Everyone is of course entitled to thier own opinion and to express it how they like.
My opinion on the subject is that S/S has a point here.
Mactronix
- Apr 29, 2006
- 29,241
- 437
- 93,590
So your saying "paper numbers" have no relation to actual performance? Who is "taking a piss" now? Sorry, but "paper numbers" mean things whether you choose to accept that or not. I didn't say context doesn't matter, but % does matter as well.
JAYDEEJOHN
Champion
- Aug 23, 2006
- 19,840
- 0
- 50,860
mactronix
Judicious
- Sep 17, 2006
- 7,928
- 1
- 32,115
Im assuming that paper numbers are the same thing as a paper launch ? no?
Which means nothing without the actual hardware.
Synthetics are fine as long as you are using the same hardware every where else in the rig and are just comparing the performance of one component.
This is the whole problem with benchmarks, a lot of people take then as fact and they just cant be, there are to many differentials to take into account.
Every single piece of hardware will work differently, even if its only very slightly. Drivers give different results on different systems, again even if its not a lot, its still there. Even the graphics chips themselves all offer various levels of performance.
The same is true for gaming benchmarks. You could build two differant systems with seemingly identical components and you wouldnt get the exact same performance.
Mactronix
Which means nothing without the actual hardware.
Synthetics are fine as long as you are using the same hardware every where else in the rig and are just comparing the performance of one component.
This is the whole problem with benchmarks, a lot of people take then as fact and they just cant be, there are to many differentials to take into account.
Every single piece of hardware will work differently, even if its only very slightly. Drivers give different results on different systems, again even if its not a lot, its still there. Even the graphics chips themselves all offer various levels of performance.
The same is true for gaming benchmarks. You could build two differant systems with seemingly identical components and you wouldnt get the exact same performance.
Mactronix
Facebook
Twitter
Instagram