• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

NVIDIA GeForce GTX 1060 Reference Board Design and Clocks Confirmed

People who make graphs that don't start from 0 (zero) should be shot in the face or deported to some gulag at the very edge of coldest Siberia. With graphs starting somewhere in the middle, graphs entirely lose all meaning. You may just as well just show a spreadsheet and no graph at all because in the end you still have to look just at numbers because graph makes it look like it's a 3x difference where in reality it's 0.1x...
Is called 'Marketing' and is actually a special course that teach you those kind of tricks to enhance perception over reality.
Its all official dude, no big conspiracy happening there
 
Agree.
Which just proves that AMD's speciality is getting things wrong, particularly marketing wise; if your competitor is technically superior (and NVIDIA certainly is so) you should keep expectations low and try to surprise. AMD does otherwise: they create hype, fail to deliver and the green team just has to release sooner and/or lower prices.


AMD should start their graphs at -100 and end them at 1,000,000!! lol

Is called 'Marketing' and is actually a special course that teach you those kind of tricks to enhance perception over reality.
Its all official dude, no big conspiracy happening there

ooooOOOoooo...... sounds sciency!!! It's actually called bullshitting where I come from.
 
Then exactly which part of the process did Nvidia rush? It had to be sitting there ready to launch since they are able to lanuch it in the next two weeks.</facepalm>
You simply can't rush a silicon chip that way.
Chips are made much further in advance than their release. However @rtwjunkie is right, this is rushed as no matter how you look at it. Nvidia does not normally release the mid range cards this soon as they like to let people first have the new high end and second tier cards and remove old stock. This shows they are worried about losing part of this market areaas they completely changed their dates by a significant amount. This just means they will be in shorter supply than nvidia likes which is why this is considered rushed.

If nothing else, remember that when the GTX 1070 and GTX 1080 were coming soon, they stopped production of their old counterparts and announced that. They have not even done that yet for the GTX 960 (or at least said they have).

AMD also did a similar thing with moving the release of Vega up.

Regardless, we only have to wait a little longer to see how this card does.
 
Chips are made much further in advance than their release. However @rtwjunkie is right, this is rushed as no matter how you look at it.
<cut>
AMD also did a similar thing with moving the release of Vega up
This is BS, then for goodness sake please explain how they manage to rush it? When a new chip is taped out it has pretty much locked the release window, starting ~10 months after tape out. Of course the maker can move the release date within the target window, but it can't be pushed forward several months, that is 100% nonsense. And BTW, the rumors of Vega being released in October is just coming from some random guy at the 3Dcenter forums, and it's even confirmed by AMD that Vega is not coming anytime soon.
 
There's no way a card with 192GB of bandwidth can beat a card with 256GB of bandwidth. These performance charts are definitely fake.
 
GTX 1080/1070 was not rushed, it was exactly as planned. The supply is somewhat limited though, not because of yields (which are excellent), but because too much of the production capacity is tied up with GP100.

I agree the 1080/1070 Paper Launch announcement was not rushed, just the actual production of these cards.
 
There's no way a card with 192GB of bandwidth can beat a card with 256GB of bandwidth. These performance charts are definitely fake.
Then how can GTX 1070 beat RX 480 by over 50%?, given:
-GTX 1070: 5783 GFlop/s, 256 GB/s memory, 150W TDP
-RX 480: 5161 GFlop/s, 256 GB/s memory, 150W TDP

I agree the 1080/1070 Paper Launch announcement was not rushed, just the actual production of these cards.
Still you fail to answer and your reply makes absolutely no sense. What part of the process was rushed? The pending announcement is evidence that it's not rushed.</facepalm>
 
This is BS, then for goodness sake please explain how they manage to rush it? When a new chip is taped out it has pretty much locked the release window, starting ~10 months after tape out. Of course the maker can move the release date within the target window, but it can't be pushed forward several months, that is 100% nonsense. And BTW, the rumors of Vega being released in October is just coming from some random guy at the 3Dcenter forums, and it's even confirmed by AMD that Vega is not coming anytime soon.
Uhh no your the one full of it. Its rushed because they don't have the production they want not to mention its poor timing with their own launch schedule because it hurts sales of the higher value cards. All you have to do is look at history to know this is rushed because its not how they do things...Done with this since your just screaming at others at this point...

There's no way a card with 192GB of bandwidth can beat a card with 256GB of bandwidth. These performance charts are definitely fake.
Memory speed only matters if the card can actually use it...
 
Uhh no your the one full of it. Its rushed because they don't have the production they want
100% BS once again. GTX 1080/1070 has sold more than GTX 970/980 in the first months. The demand this time around is even higher so Nvidia can't deliver enough, it has nothing what whatsoever being rushed or not.
 
There's no way a card with 192GB of bandwidth can beat a card with 256GB of bandwidth. These performance charts are definitely fake.
Then please explain how the 390x, with 384 GB/s is roughly on par at stock with a 980 with 224gb/s? In fact, why is the 980ti so much more powerful with only 336gb/s?

Much like bits, bandwidth ain't everything. Maxwell needed significantly less bandwidth then Kepler, and Pascal needs significantly less then Maxwell.
 
I think that little nugget is debatable, and will have to wait until actual prices are released. NVIDIA has always priced higher than the competition, and if the first two Pascal's are any indication, we're likely not in for a cheap 1060.
+1.. Any takers at $299 on this if it is at least 10% faster?
 
+1.. Any takers at $299 on this if it is at least 10% faster?

I don't remember where, but yeah, I also guessed earlier that we will be lucky if the 6GB version is less than $299. So yeah, I'm in with you, out on that limb!
 
+1.. Any takers at $299 on this if it is at least 10% faster?
I am sure it will do fine at that price point :P

Either way it will be curious, I am interested in where they price the 3gb version more than anything just because I wonder how low it will be.
 
Its not looking good for AMD,
 
The GTX 1060 is claimed to be faster than the GeForce GTX 980

Really? Hmmm interesting. I wonder what 980 owners feel like

I have a pair of them (Gigabyte GTX-980 Windforce cards) running in SLI in an i5-6600K system.

It plays the hell out of all of the games that I love.

I feel good.

Most forget that partners cards will most likely be more expensive than reference hence closing the price between the rx480 and the 1060....lets see how competitive the 1060's pricing is. fun times ahead.

Chanting: Price War! over and over.
 
Last edited:
Most forget that partners cards will most likely be more expensive than reference hence closing the price between the rx480 and the 1060....lets see how competitive the 1060's pricing is. fun times ahead.
 
I am sure it will do fine at that price point :p

Either way it will be curious, I am interested in where they price the 3gb version more than anything just because I wonder how low it will be.

Yeah, and next year GTX 1150 at $300 for the win! yay!

By 2020 entry level vcards would be $400 woooooooot
 
Yeah, and next year GTX 1150 at $300 for the win! yay!

By 2020 entry level vcards would be $400 woooooooot
That is why I vote with my wallet, I worry about pricing as well and wait hence why I skip a generation as my new goal.
 
Most certainly RX480 forced NVIDIA to do something, fast. And even if GTX 1060 is not the same segment, NVIDIA had to release it because they don't even have GTX 1050 ready... They'll have to justify higher price with also higher performance and somehow convince people who otherwise spend up to 200 to spend some more. Quite some more...
 
So supply is short of GTX 1080 (2560 cores) and GTX 1070 (1920 cores) cards and now they're preparing to launch a GTX 1060 (1280 cores). Is NVIDIA having severe yield problems? Why else would they be compelled to cut the chip down this far, this soon? That's 50% of the Pascal chip disabled! :eek: By comparison GTX 960 has 37.5% of the cores disabled compared to GTX 980.
 
Last edited:
AMD RX 480 PCB & VRM Design Beefier Than GTX 1080 Founder’s Edition – PCIe Power Issue Detailed

Read more: http://wccftech.com/amd-rx-480-pcie-power-issue-detailed-overclocking-investigated/#ixzz4DDxltUVA

gsdfgsd.png
05-Asus-GTX-960-Strix-75-Watts-Limit_w_600.png
15-Gaming-3D-PEG-Overwiew_w_600.png
asdfasd.png
pcie-power-rail-2-0.png
 
It appears AMD learned a thing or three from Nano.
 
Last year GTX 960 and GTX 970 had practically the same price/performance ratio. If Nvidia were to target that this year as well, then $299 would be too high and $249 would be too low.

This is what I expect this generation to look like (performance values averaged at 1440p):

Graphics Card - Specifications - Performance - Price - Ratio
RX 470 4GB - 1792@1.2, 256@6 - 400 - $149 - 2.68
GTX 1050 4GB - 960@1.6, 128@8 - 450 - $199 - 2.26
RX 480 4GB - 2304@1.3, 256@7 - 510 - $199 - 2.56
RX 480 8GB - 2304@1.3, 256@8 - 540 - $229 - 2.36
GTX 1060 6GB - 1280@1.7, 192@8 - 600 - $269 - 2.23
GTX 1070 8GB - 1920@1.7, 256@8 - 830 - $379 - 2.19
GTX 1080 8GB - 2560@1.7, 256@10 - 1000 - $599 -1.67



Though if Nvidia really wants to hurt AMD, they need to price 1060/1050 at $249/$179 respectively.
 
Back
Top