NVIDIA GeForce GTX 1060 Founders Edition PCB Pictured

[btarunr]

Here's one of the first pictures of the NVIDIA GeForce GTX 1060 Founders Edition (reference) PCB. The PCB is about 2/3rds the length of the actual card, and despite that, it's pretty barren. Power is drawn from a 6-pin PCIe power connector, however, this connector isn't on the PCB, but is on a receptacle towards the end of the cooler. NVIDIA designed this in response to complaints that on cards with PCB shorter than the cooler, the power connector would be in the middle of the card. It would also block the illuminated GeForce GTX logo along the top.

The 6-pin PCIe power receptacle connects to the card at big solder points. This approach has one downside. If you want to change the cooler (to, say, an aftermarket air cooler), you will have to deal with that ugly cabling. The card uses a simple 3+1 phase VRM to power the GPU, with its TDP rated at just 120W. The GP106 GPU is neighbored by six 8 Gbps GDDR5 memory chips populating its 192-bit memory bus. There's no SLI support. Display outputs include three DisplayPort 1.4, and one each of HDMI 2.0b and DVI.



View at TechPowerUp Main Site

 

[Fluffmeister]

This will make a superb HTPC card, and with apparent GTX 980 level performance it will pack a mighty punch too.

 

[GhostRyder]

Its funny to see a card without a 6 or 8 pin that needs it. I applaud the thinking done to get the connector into a spot that makes more sense on the card as its makes the card much more attractive.

 

[Casecutter]

I'm looking at that little connector and thinking can that effectively transfer 75W (6.25A) might be an issue to effectively OC such F-E. What is the wire gauge their using from the 6-pin then over to the littler 4-pin? Can anyone point to another connector in the computer industry (or any) that are that delicate and can pass the amount of power?

 

[Air]

I'm looking at that little connector and thinking can that effectively transfer 75W (6.25A) might be an issue to effectively OC such F-E. What is the wire gauge their using from the 6-pin then over to the littler 4-pin? Can anyone point to another connector in the computer industry (or any) that are that delicate and can pass the amount of power?

The 4-pin is the pwm fan conector. The 6 pci-e power connector is soldered (image from videocardz):

 

[cdawall]

I give it 3 months and they release a 256 bit version with SLi and call it a 1060Ti

Also anyone curious about expected power consumption the pictures out kind of scream that it will be substantially lower than the 480

Notice how the RX480 is a 6 phase card and the pictures of the 1060 put it as a 3 phase card...

 

[PP Mguire]

Wonder how quickly people will solder on an 8pin. That solution is terrible IMO.

 

[ppn]

I give it 3 months and they release a 256 bit version with SLi and call it a 1060Ti

GDDR5X too.

Any estimates on the dies size?

 

[cdawall]

GDDR5X too.

Any estimates on the dies size?

No clue, but for anyone doubting (now it may not have SLi that's a guess), but there will be an 256B/8GB version of this card, the PCB literally tells us that.

 

[INSTG8R]

I give it 3 months and they release a 256 bit version with SLi and call it a 1060Ti

Also anyone curious about expected power consumption the pictures out kind of scream that it will be substantially lower than the 480

Notice how the RX480 is a 6 phase card and the pictures of the 1060 put it as a 3 phase card...

It does look really cheap compared to the 480...

 

[Fluffmeister]

It does look really cheap compared to the 480...

That's the whole point! And Polaris does seem like it drinks more juice anyway.

 

[cdawall]

It does look really cheap compared to the 480...

It will be quite a bit cheaper to produce than the 480. That is what happens when you produce something that is 2x (this is a made up number) as efficient as the AMD equal, but don't worry nvidia wont pass these savings onto you so expect this card to always be more expensive.

 

[FordGT90Concept]

RX 480 has better voltage control on it than the GTX 1080. There was an article on WCCFtech a while ago that showed RX 480 voltage spikes occur maybe once every 20 seconds (less than 200w) where a GTX 960 spiked every second with massive spikes (exceeding 350w) about every 20 seconds. RX 480 is built like a tank; GTX 1060 is not. The design philosophy is completely different between the two companies.

 

[cdawall]

RX 480 has better voltage control on it than the GTX 1080. There was an article on WCCFtech a while ago that showed RX 480 voltage spikes occur maybe once every 20 seconds (less than 200w) where a GTX 960 spiked every second with massive spikes (exceeding 350w) about every 20 seconds. RX 480 is built like a tank; GTX 1060 is not. The design philosophy is completely different between the two companies.

It has to be it consumes about the same amount of power as the 1080 and has worse power coming in...The 1060 is a cheap low wattage card and is build as such.

 

[chief-gunney]

According to AnandTech article, the $299 FE GTX 1060 will only be available for purchase from Nvidia website.

 

[FordGT90Concept]

http://www.guru3d.com/articles-pages/amd-radeon-r9-rx-480-8gb-review,5.html
184w versus 166w yet the 166w can handle 50% more amps by design.

 

[cdawall]

http://www.guru3d.com/articles-pages/amd-radeon-r9-rx-480-8gb-review,5.html
184w versus 166w yet the 166w can handle 50% more amps by design.

Correct that is peak consumption figures during normal gaming loads it will run about the same. From nvidia's standpoint why spend extra money when the card doesn't need it? There are not power delivery issues with the 1080 with half the PWM...I personally would have much rather seen AMD use a single 8 pin, a cheaper phase setup and cut the card prices another $30-50

 

[Casecutter]

The 6 pci-e power connector is soldered (image from videocardz):

Whew, that's more like it!

 

[ppn]

At 193 mm² this can only be 192 bit. I think full 256 bit GP104 is pin compatible with this PCB. Maybe 1536 enabled shaders instead of 2560,

 

[cdawall]

At 193 mm² this can only be 192 bit. I think full 256 bit GP104 is pin compatible with this PCB. Maybe 1536 enabled shaders instead of 2560,

Correct and I assume that will be the GPU that the 1060Ti will be based off of, although NV could really mess with AMD and release a 1070LE using the PCB and just insult AMD's midrange purely out of naming scheme

 

[hojnikb]

No clue, but for anyone doubting (now it may not have SLi that's a guess), but there will be an 256B/8GB version of this card, the PCB literally tells us that.

i highly doubt gp106 has 256bit interface. This must be some weird quirk or something. Maybe they reused the gp104 pcb and simply cut the vrm section.

 

[cdawall]

i highly doubt gp106 has 256bit interface. This must be some weird quirk or something. Maybe they reused the gp104 pcb and simply cut the vrm section.

Nope the PCB is setup to handle more, but isn't using it, like was said its probably pin compatible with a GP104 which is likely a later design idea for a wipe the floor with something AMD sells plan.

 

[BiggieShady]

i highly doubt gp106 has 256bit interface.

Well gp106 is a new chip and has half clusters as gp104. It could be setup like this:

gp104:

gp106:


This way both would have 256 bit memory interface.

 

[ppn]

Gp106 is wider but shorter (rotated right 90°). So they took 1 64 bit channel and moved it to the wider side. Just theory, can't be proven. Apparently that's not how it works.

 

[ShurikN]

The 1060 is a cheap low wattage card and is build as such.

At $300 it's really not that cheap. For that amount of money it looks rather poor. But hey it's NV, no surprises there...