It's R9 380 vs GTX 960 all over again.
And AMD keeps winning.
>>55429775
Until you pair it with anything weaker than an overclocked haswell i5
Pretty low clock there for pascal desu senpai.
more 1060 benchmarks when
It looks like its going to be as fast as a GTX980 stock clocks.
It all depends on how efficient memory compression is on Pascal compared to Maxwell to make up for that bandwidth.
Someone do the math.
>>55431132
Why would memory bandwidth matter do much? 1060 still be the new 1080p card.
>>55431362
The faster the memory the better performance you get.
how low do they have to price this to be able to compete with the 480?
"1060 is 8-10% faster than a 480"
>>55431512
It will probably be $270.
There's no way Nvidia will undercut the GTX970 when theres already plenty of them on shelves and the 480 is competing with the GTX970 not this.
>>55431132
GTX970 looks to have better specs.
>>55431362
A GXT 980 or AMD equivalent (390/390x) can comfortably play games at 2560x1440.
1060 would too, if it wasn't memory gimped
>>55431546
It all depends, Pascal has better caches and improved math/geometry pipeline plus better memory compression.
a GTX980 has a bandwidth of 224 GB/s, if Pascal memory compression is good as Nvidia claims, 192.2GB/s could easily match the 224 GB/s from the GTX980.
>>55431444
Because HBM really helped the fury x for 1080p
>>55431561
That's true, but nvidia doesn't care. Until 1440p is mainstream they'll reserve 1440p cards for enthusiasts like the 1070 and 1080.
Still, 192 bit with compression and good memory frequency should be enough for decent 1440p performance.
>>55431605
>Because HBM really helped the fury x for 1080p
It did though, the Fury X problem was being limited to 4GB and serious ROP bottlenecks.
>>55429775
Well, 11.225 is only 3% more than RX 480's 10.890, so if Nvidia wants to match AMD's price/performance ratio they must price the 1060 at around $260 or otherwise they're gonna lose a lot of the mainstream market share.
That's lower than my Oced 970 with 13500 graphics score and 10700 total (which is bottlenecked by 7500 score of my 3570)
>claims top vr performance
>cant even connect to vr devices
hahahahahahahaha cucked so many times nvidiots
>http://www.tomshardware.com/news/nvidia-vive-displayport-incompatible,32204.html
>https://forums.geforce.com/default/topic/939426/geforce-1000-series/displayport-does-not-work-with-the-htc-vive-on-gtx-1080-/
>https://www.reddit.com/r/Vive/comments/4mgpqv/warning_upgrading_to_gtx_1080_breaks_mini_dp_port/
>>55431699
http://www.3dmark.com/fs/9115640
>2 years old
>Still beats 480 and 1060
>No 3.5 vram issue
If it wasn't for the 3.5 meme 970 would be best card since 4850.
>>55431765
It is.
>>55431765
3.5gb isn't an issue though
>https://www.youtube.com/watch?v=rhjC_8ai7QA
>When even professional AMD shills eat their words and admit 480 sucks.
>>55431984
I know I've been posting Lords of the Fallen webms using 4 gigs of vram for a while.
WHY DO THEY CALL IT A GTX 1080
BECAUSE IT ONLY DOES 1080p
>>55431132
>how efficient memory compression is on Pascal
https://www.youtube.com/watch?v=yblANXXMuds
Seems good enough.
>>55431664
>if Nvidia wants to match AMD's price/performance ratio they must price the 1060 at around $260 or otherwise they're gonna lose a lot of the mainstream market share.
I'm an AMD fanboy, but I think if I had ~$250 for a video card, I'd pony up the extra $20-30 for a 1060, even if it had the same performance as a 480
nvidia, as of Maxwell, does efficiency better. a 480 is close to 1080/1070 power draws but has half the performance. the 1060 will probably use a 6pin and still have room to spare, power-wise
this is coming from someone who's had a 290 tri-x for a few years
>>55432399
The thing is that, knowing Nvidia, I'm pretty sure they're gonna price it at least $50 more than an RX 480. I don't know how far everyone's willing to go for a 3% increase in performance.
>>55432399
I really wonder why so many people care about power consumption of GPUs in stationary Gaming/work PCs. As long as you're not running a server or a ton of these things in work PCs the only thing i care about is that the card/system is silent and performs well. Who gives a shit about a few cents more for power?
>>55432038
how much VRAM you use isn't really that important with the 970s segmented interface, especially since Nvidia got the hang of making it work. The card runs into problems when you have important assets the chip needs split up in the two memory pools. In the worst case it looks like pic related, but that didn't really happen all that often and in almost all cases got fixed quickly.
I wonder what will happen once special driver support for the 970 stops when they got the full pascal line up out. Considering they dropped GK110 like a hot potato i wouldn't be so sure about continous support with drivers that fix the problems that might come out of the whole segmented VRAM bogaloo.
>>55432527
again, I'd be willing to pay more for the same performance if I make gains somewhere else
>>55432546
>who gives a shit about a few cents more for power?
I don't care at all about the extra few cents, but I do care about the added heat and noise that the card produces. the noise isn't that bad if your AIB cooler is good, but the extra heat is fucking shit, especially during summer
>more 3.5 cards for cucks
>>55432592
Enjoy your 4GB "4k" flagship
>>55432573
yeah, i care about my card being silent too, but with all the AIB models out there that really is more related to which card exactly you choose than to how much power the chip burns through.
I doubt that the added energy from the card is any problem. I come from a cold country, but most space heaters start at like 2kW for a reason. Heck, your own body puts out like 100+W if you're not asleep.
>>55432621
that isn't OK either. Both cards suck, the only cool thing about Fiji is the impressive tech behind it. In some year's i'll buy a broken used one for a few bucks, desolder the package and cast it in an epoxy block.
>>55432633
I live in Canada, and in the winter I don't mind the 290's extra heat, but in the summer, well, like I said
an OC'd 290 draws ~225W under load, which is about the same as an extra two humans in my room, whereas a 970 uses about 80W less
>>55432546
i don't care about the increase in my electrical bill. electric is piss cheap in 'merica. at least where i'm located. what i do care about is the extra heat.
yes the card may run cool because the cooler does a great job at cooling the card. but it does a great job at cooling the card because it gets the heat OFF the card. where does the heat go? sadly it doesn't magically teleport into another dimension. no, instead, it pumps out into the room.
with it being 46c outside, anything that helps keep my room cooler with my struggling ac is most welcomed. if the 1060 is 3% faster but uses 10% less power, i will gladly pay the extra $30 for it.
>>55432718
>>55432744
well, i have no empirical evidence whatsoever for this, but is this really noticable? 100 or so W difference sound like nothing as long as your room isn't ridiculously small. I mean the electric furnace drops like 1kWh into my kitchen for a pizza and i hardly notice a difference in room temperature.
>>55429775
>inb4, 5.5 GB
>>55432828
>I mean the electric furnace drops like 1kWh into my kitchen for a pizza and i hardly notice a difference in room temperature.
different circumstances
a furnace/oven is designed to keep heat in, whereas a computer case is meant to take cold air in and output hot air. I'm sure if you left your pizza oven open, you'd quickly realize how much heat it's putting out
>>55432857
Did you not read the thread?
>>55432868
all the energy still ends up in the room in a relatively short amount of time. How long does it need the oven to cool back down? A bit less than an hour i'd guess. So that's a bit more than 1 kWh dumped into my room in a bit more than an hour, or to say it in other words, about 1kW of energy draw dumped into my room for about an hour.
It's interesting, i'll close my kitchen in the next days an measure how much it really heats up when i bake something. But except for the direct heat you get from the oven when you open it or sit directly in front of the glas panel i never noticed a difference in room temperature.
>>55432950
What card is that
>>55429775
>>55431664
>>55432527
Historically, AMD's cards do better in synthetic benchmarks but worse in actual games. So if the synthetic benchmark is better, expect real world performance to be significantly better.
