Seems like AMD is targeting the mid-range segment for now, which is good, at least from the way I see it that is. Hopefully, reviews of the RX-480 come in next week so we can finally know just how good (or bad) of a card it really is.
The current new cards AMD is putting out are not any better than their past cards. They are more efficient and the like, but so far look to be rehashes of the last line with some minimal improvements. You won't see cards to compete with the 1070/1080 just yet from AMD and this has been stated multiple times multiple places...
Actually its other way around, 1070 and 1080 are just pretty much Maxwell on 16nm node wih higher clocks while AMD has new architecture.
1 guy made some effort to figure out how much he paid for NVidias x70 cards he bought:
Wow. Wasn't the 770 more expensive than the 970 at launch?
Launch price for GTX 770 was 399$ Launch price for GTX 970 was 329$ Launch price for GTX 1070 is 449$
Notice the Order placed -date on the picture Malabooga posted. GTX 770 was ordered from Amazon 1 year 3 months after its launch date.
It's a perfect picture for Malabooga, he has some good info but he's AMD fan who goes to great lengths at choosing information that makes NVidia look as bad as possible.
My friend's selling me his 970 for a rather cheap price and I'm thinking of getting get it. But, with the 1070 and RX-480 releasing in a few weeks, I was wondering if I might benefit more from waiting for either of the two to release, or am not losing much if I got the 970 for, let's say, around $170?
My monitor's resolution is 1080/60hz, by the way, but I might upgrade to something like a 1080/144hz (if one ever does get released) in a few months or a year.
It looks like it is about 2 970 the 1070. At 1080p there isn't anything I know of you cant play. . It really depends on what games you want to play. There isn't much out there that can even use a 980. Most games are built after the garbage ps4 stats, and most people have shitty computers. So that card is gonna cost like 500 bucks. If getting 110 fps in games you can only see 60 fps is worth 300 bucks then do it. I would go with the 970 for 170 that is a seriously good deal, and will play everything. I even played ark with good fps, and that is not even optimized. So really unless you are gonna dump 600 bucks in some vr stuff, or 600 bucks on a good 4k tv or monitor, I don't see the point in waiting to spend 300+ more on something you won't even really be able to tell the difference.
I mean unless you are one of those people who just like to see numbers like fps numbers for no other reason, then you are getting better fps then some one else.
1 guy made some effort to figure out how much he paid for NVidias x70 cards he bought:
Wow. Wasn't the 770 more expensive than the 970 at launch?
Launch price for GTX 770 was 399$ Launch price for GTX 970 was 329$ Launch price for GTX 1070 is 449$
Notice the Order placed -date on the picture Malabooga posted. GTX 770 was ordered from Amazon 1 year 3 months after its launch date.
It's a perfect picture for Malabooga, he has some good info but he's AMD fan who goes to great lengths at choosing information that makes NVidia look as bad as possible.
1. im not AMD fan, in fact i currently have GTX970 and have owned many cards from both camps (thats why noone can BS about drivers as im aware how both camps have fared first hand)
2. Guy posted what he paid for each (he also bought 1080 mind you) You can even disregard 770. Because if you take into account youll see how bad 1070 is - 970 brought same relative performance as 1070 at much lower price, while 1070 is pretty much historic high for NVidia
It looks like it is about 2 970 the 1070. At 1080p there isn't anything I
know of you cant play. It
really depends on what games you want to play. There isn't much out
there that can even use a 980. Most games are built after the garbage
ps4 stats, and most people have shitty computers. So that card is gonna
cost like 500 bucks. If getting 110 fps in games you can only see 60 fps
is worth 300 bucks then do it. I would go with the 970 for 170 that is a
seriously good deal, and will play everything. I even played ark with
good fps, and that is not even optimized. So really unless you are
gonna dump 600 bucks in some vr stuff, or 600 bucks on a good 4k tv or
monitor, I don't see the point in waiting to spend 300+ more on
something you won't even really be able to tell the difference.
I
mean unless you are one of those people who just like to see numbers
like fps numbers for no other reason, then you are getting better fps
then some one else.
The current new cards AMD is putting out are not any better than their past cards. They are more efficient and the like, but so far look to be rehashes of the last line with some minimal improvements. You won't see cards to compete with the 1070/1080 just yet from AMD and this has been stated multiple times multiple places...
If a level of performance that used to cost you $400 and 250 W now costs you $200 and 120 W, do you not see that as improvement?
Furthermore, if you have a 438 mm^2, 250 W die, you really can't scale it up to something much faster. If you can now get that performance in 230 mm^2 and 150 W, you can readily scale up the architecture to give something much faster than before. And AMD more or less will, up to whatever architecture differences there are between Polaris and Vega.
^ True. It really seems that it's the other way around now, with Pascals looking like improved rebrands and Polaris looking like entirely new architectures. Hopefully, AMD keeps this up. God knows how much we need some competition to get the damn prices to drop. Lol
Both Pascal and Polaris get a lot of mileage out of being die shrinks. I'm not expecting either to be wildly new architectures. Probably the most notably new architecture is the top end Pascal, but even that can mostly be described as a mix of GCN and Maxwell. And from the specs Nvidia has announced, it's likely not going to be a worthwhile gaming card, anyway.
Well, I'm curious if Nvidia will pull a "big" pascal like they did with Maxwell (980ti/titan black) and build a pascal chip that's around 600mm2 and all the double precision bits removed with HBM2 and 16gb of RAM.
I'm sure they will, there would be no reason not to, eventually of course, to take the performance crown if needed.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
Well, I'm curious if Nvidia will pull a "big" pascal like they did with Maxwell (980ti/titan black) and build a pascal chip that's around 600mm2 and all the double precision bits removed with HBM2 and 16gb of RAM.
I'm sure they will, there would be no reason not to, eventually of course, to take the performance crown if needed.
Judging by the looks of things right now, there's just no reason to do that yet. Hopefully they do, though. It'd be nice to see what a card like that would be capable of.
Well, I'm curious if Nvidia will pull a "big" pascal like they did with Maxwell (980ti/titan black) and build a pascal chip that's around 600mm2 and all the double precision bits removed with HBM2 and 16gb of RAM.
I'm sure they will, there would be no reason not to, eventually of course, to take the performance crown if needed.
When Nvidia has built a big compute chip in the past, that's always been the top end for graphics, too. It costs a ton of money to make another huge chip--more so than to simply make more copies of another huge chip that you already made and simply disable 1/10 of the chip or whatever. Of course, from the specs Nvidia has announced, GP100 sure sounds broken in ways that GP104 isn't. That wasn't the case for Fermi (where everything in the 400 series was broken, and the 500 series was the fixed version) or Kepler.
Whether or not GP100 or perhaps a respin of it will be Nvidia's top end GeForce GPU on 16 nm has presumably already been decided. These things take a lot of time, and you can't just decide to build a new chip and do a hard launch 3 months later. If Nvidia decided today to start on a brand new 16 nm chip, there's no guarantee that it beats 10 nm chips to market. And it has to beat 10 nm chips to market by a lot in order to have any point, unless 10 nm ends up being a disaster to rival 20 nm.
Comments
Launch price for GTX 970 was 329$
Launch price for GTX 1070 is 449$
Notice the Order placed -date on the picture Malabooga posted. GTX 770 was ordered from Amazon 1 year 3 months after its launch date.
It's a perfect picture for Malabooga, he has some good info but he's AMD fan who goes to great lengths at choosing information that makes NVidia look as bad as possible.
Sources:
http://www.anandtech.com/show/6994/nvidia-geforce-gtx-770-review
http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/7
http://www.anandtech.com/show/10304/nvidia-announces-the-geforce-gtx-1080-1070/2
I mean unless you are one of those people who just like to see numbers like fps numbers for no other reason, then you are getting better fps then some one else.
2. Guy posted what he paid for each (he also bought 1080 mind you) You can even disregard 770. Because if you take into account youll see how bad 1070 is - 970 brought same relative performance as 1070 at much lower price, while 1070 is pretty much historic high for NVidia
used 970 for 170$ is not a very good buy, sorry.
Furthermore, if you have a 438 mm^2, 250 W die, you really can't scale it up to something much faster. If you can now get that performance in 230 mm^2 and 150 W, you can readily scale up the architecture to give something much faster than before. And AMD more or less will, up to whatever architecture differences there are between Polaris and Vega.
Well, I'm curious if Nvidia will pull a "big" pascal like they did with Maxwell (980ti/titan black) and build a pascal chip that's around 600mm2 and all the double precision bits removed with HBM2 and 16gb of RAM.
I'm sure they will, there would be no reason not to, eventually of course, to take the performance crown if needed.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
Whether or not GP100 or perhaps a respin of it will be Nvidia's top end GeForce GPU on 16 nm has presumably already been decided. These things take a lot of time, and you can't just decide to build a new chip and do a hard launch 3 months later. If Nvidia decided today to start on a brand new 16 nm chip, there's no guarantee that it beats 10 nm chips to market. And it has to beat 10 nm chips to market by a lot in order to have any point, unless 10 nm ends up being a disaster to rival 20 nm.