New AMD Polaris based GPU

[ub4ty]

Deciding where to move the eye next after a visual stimulus on its own takes 100-140ms - what do you expect to achieve by seeing an image 5ms earlier than someone else? In a fast-paced game things like aim and skill are much more important than if you are able to see your opponent a few milliseconds before he can see you. So, no you don't win if you can see your opponent 5ms earlier due to your monitor.

He is a perfect example of how effective marketing is. Scores of people are buying overpriced hardware w/o any knowledge, intelligence, or critical thinking about the human body and its limitations... They actually think they're going to beat someone with true skills and APM capability pushing 240 (THE PROS) because they have a GPU that costs 2-3x as much and a goofy $1000 gaming monitor. Anyone remember the 'Killer' NICs? LOL

It's like the scrub who can't get picked up at the neighborhood basketball court thinking because he has $500 Nikes on that he's Jordan.

I'm never amazed at how confidently people reiterate marketing lines in arguments and other such gimmicks. He literally thinks that 5ms is going to make the difference between him and a skilled player... He reasserts it multiple times as if we have a brain disfunction...

You can't make this up
...

 

[Phynaz]

Since we are quoting Black Friday prices,

$399 for Vega 64 + 3 games that is 15-20% faster than GTX 1070

or

$190 for RX 580 8GB + 2 games that is as fast as GTX 1060 6GB.

I wonder why you had to only quote NVIDIA black Friday prices in a RX 590 topic

Evidence that's a Black Friday price?

You have none.

 

[Phynaz]

Since we are quoting Black Friday prices,

$399 for Vega 64 + 3 games that is 15-20% faster than GTX 1070

A Vega 64 has roughly the same performance as a GTX 1070 for $100 more and 150W more heat.

Doesn't exactly stand up to a greater performance increase, for less money, while having lower power consumption.

You're defense of AMD is digging yourself a hole again.

 

[PhonakV30]

PCLab Undervoltaing RX590:

PCLab was able to get 1580Mhz at 261w while Rx580@1450mhz with 276W ( both undervolt )

also Brent on HardOCP forum did Undervolting RX590

RX480@1266Mhz (stock voltage 1.168v) , 269w
RX590@1266Mhz (simulating voltage around 1.168v) , 256w

Both cards are XFX fatboy

 

[coercitiv]

A Vega 64 has roughly the same performance as a GTX 1070 for $100 more and 150W more heat.

A Vega 64 is roughly GTX 1080 performance. Let's keep this within the bounds of reality.

 

[linkgoron]

A Vega 64 has roughly the same performance as a GTX 1070 for $100 more and 150W more heat.

Doesn't exactly stand up to a greater performance increase, for less money, while having lower power consumption.

You're defense of AMD is digging yourself a hole again.

According to TPU the Vega64 is around 20% faster at 4k, 16% faster at 1440p and 12.5% at 1080p. It is actually as fast or faster than the 1070ti in all resolutions.

Vega 64 is much closer to the 1080 than the 1070... The 1080 is faster by 5% at 1080p, 3% at 1440p and 0.7 percent at 4k.

The numbers were taken from the latest 590 review.
https://www.techpowerup.com/reviews/Sapphire/Radeon_RX_590_Nitro_Plus/32.html

 

[AtenRa]

Evidence that's a Black Friday price?

You have none.

It clearly says "Sale ends in 3 days (fri)"

As of today, this is clearly a better product than RX 590 , as is the RX 580 at $190. But what happens next Monday when prices will return to 350+ for the GTX 1070 and 240+ for the RX 580 ???

 

[Qwertilot]

It isn't clear they'll go all the way back - those cards are well into stock clearance mode at this point.

Obviously not remotely fair to expect a refreshed card to offer better value at this point in the cycle though!

 

[lifeblood]

As prices fluctuate so does the value proposition. If I’m a 1080p gamer, do I need anything faster than a RX 580/GTX 1060? But if I think my next monitor might be a 2K, then suddenly a RX 590/1070/Vega 56 become more interesting. And if I’m thinking 4k, then GTX 1080/Vega 64 comes into play. RX 590 is cheaper but generates more heat. If I live in cold climates or have cheap electricity then the cheaper product is (probably) better. If I live in a hot climate or with expensive electricity, I should (probably) should go with the more expensive but less “hot” product.

Too many people are looking at their own specific situation or experience and act like that applies to everybody. When regular people do it on forums, it leads to silly arguments. When politicians or leaders do it, it leads to wars. We really should stop thinking that everybody should think like us.

 

[SpaceBeer]

So it is ok to pay $600+ for 250W in your PC case (nVidia x80 (Ti)), but if you are buying card for $250 (RX 590), then 250W is an issue, it is too loud, and no PC case and cooling system can handle it? And in both cases perf/price is same.

 

[PeterScott]

So it is ok to pay $600+ for 250W in your PC case (nVidia x80 (Ti)), but if you are buying card for $250 (RX 590), then 250W is an issue, it is too loud, and no PC case and cooling system can handle it? And in both cases perf/price is same.

It's never that simple. But if someone is buying a Top end GPU like a x80 Ti, it's a LOT more likely that they will have a big PSU, to handle any extra power demands.

Also the RX590 is $280 minimum on newegg. The minimum GTX 1070 is $310 (after rebate), which is $30 more for significantly higher performance and significantly lower power draw.

The RX590 just doesn't look that great overall at current pricing.

 

[Dribble]

So it is ok to pay $600+ for 250W in your PC case (nVidia x80 (Ti)), but if you are buying card for $250 (RX 590), then 250W is an issue, it is too loud, and no PC case and cooling system can handle it? And in both cases perf/price is same.

The primary questions will be is does it fit in my case, can my psu handle it, is my case cooling good enough. Obviously those with high end gpu's will tend to have high end cases, psu's and cooling. By the time you get to 580 performance levels that's not so guaranteed, so you loose a chunk of your market there who simply can't fit the card in the case, cool it properly or power it properly without spending more $$$ on other components.

Other then that it's just a reason not to buy it over the competition - taken alone perhaps not that important, but they all add up. Other reasons - it's old tech, I can buy the same performance second hand for a fraction of the price, I can buy new cards at better price/perf, "it's not Nvidia" (fair or not it's fact).

They all add up, and probably mean most of the world will keep buying 1060's (just like they bought 970's not 290's), and the value conscious (much more likely to buy AMD) will buy a 580 or buy second hand.

 

[SpaceBeer]

If there's better perf/price offer, that's fine argument, whether it's 1070 or 580/1060. But these power arguments are complete nonsense. At least for 95% of gamers. I mean, full system power consumption with this card (and Ryzen or Core ix CPU) is around 350W. How many people willing to spend ~$800 for PC, will have trouble to spend $50-60 for decent 80+ bronze ~500W PSU (e.g. SeaSonic S12II 520), and would rather save some money, buy some for $30-40 and risk complete system failure. Even with GTX 1060/1070? Cause PSU is more than just max power output.

Maybe some i3/Athlon + R7/GTX 750 owner with cheap PSU looking for GPU upgrade should care about power consumption. But even in that case, when you can't afford new PSU, putting $250-300 card on that old one is not so smart move. Regardless of power consumption

So maybe the only reason to consider low TDP cards is if you are building some ITX/HTPC system

 

[kawi6rr]

Completely agree! I've had a 700W PSU in my system since 2012 they're cheap so why not go big on that front.

 

[PhonakV30]

AMD set too much voltage for RX590 to ensure that the card works without any issue. a reviewer tested and saw 46w for reducing only 50mv (1.150 V to 1.100 V ) at 1580Mhz.

 

[frozentundra123456]

So it is ok to pay $600+ for 250W in your PC case (nVidia x80 (Ti)), but if you are buying card for $250 (RX 590), then 250W is an issue, it is too loud, and no PC case and cooling system can handle it? And in both cases perf/price is same.

An apples to oranges comparison at best. The 1080 (or Ti) has no match in performance that is more efficient. There are alternatives to the 590 however that give much better performance for the same or even less power consumption, and also alternatives that give very similar performance at half the power consumption.

And it still never ceases to amaze me how 250 watts for a gpu from company X is completely negligible, while 150 to 200 watts for a cpu from company Y is portrayed as something that will make you PC explode, short out the electrical grid, and cause a nuclear power plant meltdown.

 

[killster1]

You can get a GTX 1060 6Gb for $199 after MIR with Monster Hunter World now. I'm happy that AMD decided to release this card, even if the only thing its good for is to drive prices down of cards that have been long overdue for a price drop. I think we will continue to see price drops of all mid-ranged cards, because they have been overpriced for some time now.

https://store.zotac.com/refurbished 120$ for 1060 6gb refurb zotac is decent brand to me too.

 

[PeterScott]

If there's better perf/price offer, that's fine argument, whether it's 1070 or 580/1060. But these power arguments are complete nonsense. At least for 95% of gamers. I mean, full system power consumption with this card (and Ryzen or Core ix CPU) is around 350W. How many people willing to spend ~$800 for PC, will have trouble to spend $50-60 for decent 80+ bronze ~500W PSU (e.g. SeaSonic S12II 520), and would rather save some money, buy some for $30-40 and risk complete system failure. Even with GTX 1060/1070? Cause PSU is more than just max power output.

Maybe some i3/Athlon + R7/GTX 750 owner with cheap PSU looking for GPU upgrade should care about power consumption. But even in that case, when you can't afford new PSU, putting $250-300 card on that old one is not so smart move. Regardless of power consumption

So maybe the only reason to consider low TDP cards is if you are building some ITX/HTPC system

It's a non issue if building a new PC.

It's a real issue if you are looking to upgrade an older PC without 8-Pin power connectors and you don't want to upgrade your PSU just so you can choose this card over a GTX 1060.

It makes the price performance of RX 5x0 that much worse than GTX 1060 if you also have to factor in buying a new PSU to go with it, not to mention time and hassle of swapping the PSU.

 

[Justinbaileyman]

I dont see why everyone is hating the rx 590's?
$279 for 3 games and a GPU, which less then 8 months ago would have gone for $600+ for the GPU alone.
Any ways my Red Devil rx 590 arrived today and I have been playing with it all day.
So far I am really digging this card.
Not only is it faster then my gtx 1060 6gb sc in all the games I have tried out so far, but its like 3 times the size and weight.
I was expecting a 2 slot card but its closer to 3 slots if anything.
Very chunky and feels like quality compared to my tiny little gtx 1060.
Its also very very pretty looking in person!!
Should hold me over nicely till AMD launches its next gen GPU's.
So now im just wondering if thats gonna be Polaris or Vega??

 

[PeterScott]

I dont see why everyone is hating the rx 590's?
$279 for 3 games and a GPU, which less then 8 months ago would have gone for $600+ for the GPU alone.
Any ways my Red Devil rx 590 arrived today and I have been playing with it all day.
So far I am really digging this card.
Not only is it faster then my gtx 1060 6gb sc in all the games I have tried out so far, but its like 3 times the size and weight.
I was expecting a 2 slot card but its closer to 3 slots if anything.
Very chunky and feels like quality compared to my tiny little gtx 1060.
Its also very very pretty looking in person!!
Should hold me over nicely till AMD launches its next gen GPU's.
So now im just wondering if thats gonna be Polaris or Vega??

It's not hate worthy, it's "meh" worthy.

It's poor value. It's only about 5% more performance that rx580 AIB cards and it's a lot more than 5% more expensive.

3 times the size and weight is not an advantage. It's out for SFF cases.

 

[Justinbaileyman]

Yeah there is definitely no way its gonna fit in a SFF case unless you buy a custom water block!!
But its working great for me right now and thats all that matters to me!!
Really the only thing I can complain about is the card takes a 8 pin and a 6 pin power cable which forced me to pull off the back case panel to add the second cable.
Other then that I am a pretty happy camper today!!

Edit : Also to let everyone know expect to be mad about having to wait till January to be able to play the free games cause we're not getting them till them !!

 

[wilds]

A Vega 64 has roughly the same performance as a GTX 1070 for $100 more and 150W more heat.

Doesn't exactly stand up to a greater performance increase, for less money, while having lower power consumption.

You're defense of AMD is digging yourself a hole again.

V64 decimates my 1070 in every game I play. Overclocked 1070 vs stock V64...
I thought the same as you until I got to try V64. Fantastic card!

 

[beginner99]

and any advantage of 5ms in seeing the image won't matter since 5ms is well within the error margin of reaction times

Ok. So I propose a game (like already replied to ub4ty). 2 players. They both get a visual stimuli. The one that reacts faster gets $100 from the other player. of course you play 1000th of rounds. The catch is the other player may or may not be the same one every round (but you are) and you get to see the stimulus exactly 5ms later than the other player every single time.

Will you play this game? I mean according to you the delay has 0 effect on your chance to win. At what delay will you stop playing? 10ms? 20ms? And what logic defines your threshold?

They actually think they're going to beat someone with true skills and APM capability pushing 240 (THE PROS) because they have a GPU that costs 2-3x as much and a goofy $1000 gaming monitor. Anyone remember the 'Killer' NICs? LOL

Yeah sure. Except you never play against pros. I could make the same argument for biking. I get the expensive bike that weights 12 lbs less over your average one. He have similar stamina (eg what i have been repeatign about 100 times now: all other things being equal). Do you think I do not have an advantage when climibing a hill? Of course I have one. against you. but I for sure won't beat Lance Armstrong and even less so a actual current pro.

Your real point is, why should I pay more to beat you for essentially no actual gain especially not monetary. Fair enough. But that is pure subjective. Mabye it's worth it for me or maybe not.

And besides that I actual bought the 144 ht display because a) i didn't want to go higher than 1080p for not needing to spend $1000 on GPUs and b) to still have a benefit in old games that could easily run at >1080p. (plus I had a lower resolution one before). In fact I'm far closer to your point about not spending huge amounts on electronics. But not spending it doesn't make it useless of without benefit.

 

[wilds]

I play a lot of R6 Siege, and playing below 144hz is a serious disadvantage. 60hz players are simply destroyed.

1080p144hz is getting seriously cheap now, and without the GSYNC tax, AMD is looking quite good. I also bring up R6 Siege because of how surprisingly well it runs on AMD hardware. I say this because the NVIDIA splashscreen appears on starting up R6.

The most popular card for 144hz on Siege appears to be the GTX 970. Most 144hz competitive players do not run max settings.

 

[tamz_msc]

Ok. So I propose a game (like already replied to ub4ty). 2 players. They both get a visual stimuli. The one that reacts faster gets $100 from the other player. of course you play 1000th of rounds. The catch is the other player may or may not be the same one every round (but you are) and you get to see the stimulus exactly 5ms later than the other player every single time.

Will you play this game? I mean according to you the delay has 0 effect on your chance to win. At what delay will you stop playing? 10ms? 20ms? And what logic defines your threshold?

Accepting your challenge I went ahead and tested how fast I can double-click. This is what I measured in five trials, in seconds:
0.137, 0.138, 0.169, 0.152, 0.155
The average value is 0.1502s. Now for a second set of double clicks:
0.158, 0.162, 0.137, 0.165, 0.156
The average value is 0.1556s. Now if I assume that for the first set the image appeared with a 5ms delay every time then adding it to the average value results in a total reaction time of 0.1552s, which is basically identical with the average of the second set up to three significant figures. If you let this experiment run for 1000 trials then it is pretty obvious that the averages will converge(when you carry out this experiment among different people), regardless of the set to which you add the delay because when you calculate it the standard deviations of the two data sets is already larger than 5ms. In other words, the delay is well within the margin of error of reaction times, and thus it will have no measurable difference whatsoever.