Tuesday, November 13th 2018
NVIDIA Deploys GP104 GPU for GDDR5X version of GeForce 1060
NVIDIA has just shown us one of the most ingenious ways of creating new custom, competitive SKUs for the midrange market without spending any additional amounts of money on R&D, wiring, or memory controller work: just reuse the chips that already have that work done. This is the case for NVIDIA's new GTX 1060 GDDR5X graphics card, which the company has "designed" to further fill in the gaps on its midrange offerings against a revamped Radeon RX 590.
Instead of reworking a purpose-built memory controller solution compatible with the GP106 GPU, the company has gone and carved the SKU from its existing GP104 silicon - which already supports the GDDR5X memory subsystem due to its implementation on the GTX 1080 (GP104) and 1080 Ti (GP102) graphics cards. A smart usage of GP104 inventories - which have been superseded by NVIDIA's new RTX 20-series in the high end - or of very defective dies (remember the GTX 1060 has half the shaders, at 1280, compared to the GTX 1080's 2560). This decision by NVIDIA could also go some way in explaining dwindling inventories and increasing pricing of GTX 1080 graphics cards, as chips that could have been used for that SKU are (possibly) being used for the new GTX 1060.
The discovery came courtesy of a teardown on iGame's GTX 1060 U-TOP V2, which features a triple-fan cooling solution, 2x 8-pin power connectors, and an 8+2 phase power delivery design, via Taobao. Apparently, even the SLI fingers remain on the card, which if you'll remember, never where supported on NVIDIA's GTX 1060 - a result of iGame's decision to simply reuse their PCB design for the usually much more powerful, GP104-based GTX 1080.
Sources:
Taobao, via Videocardz
65 Comments on NVIDIA Deploys GP104 GPU for GDDR5X version of GeForce 1060
They never wanted the 1080 to exist for too long. Full die, GDDR5X. That is why we got a 1070ti. Much easier, much cheaper, more leeway.
Today we have RTX and they desperately need to push it because it won't sell on its own, its a shitty deal no matter how you twist it. The 2070 is what they want to sell in place of the 1080's slot. In the meantime they can make super crappy 104's and still use them, they can get the worst bins and still use them. What's not to like? They still have a place for everything they've produced for Pascal, and they push the RTX market share on the high end. Win win.
Works very well in #cluelessbuygreen market.
If it runs like a 1060 and performs like a 1060 then it's a 1060 no matter what die they use.... if the batch failed to be a 1080 then it fails to be a 1070 ti scrap it right? No you tone it down and sell it as the next best thing, it's not greed it's how the industry is supposed to run, it's all based on strategy, plus even if the dies were supposed to be full fledged 1080's it makes 0 sense to make a product that would cannibalize the sales of a new piece of hardware you're actively pouring money into, it's not a new practice, besides if you want 1080 performance the 2070 at the same price bracket fills the slot, don't know about you but I rather have new hardware rather than last gen's part for the same price.....
What are they supposed to do with Chips that dont meet the grade for 1080s. or 1070s?? Just throw it in the trash?
::EDIT::
This 'trickle down' is good....It means that consumers might get an even more capable 1060 compared to the rest of the current 1060s available. Wouldnt you like it if you managed to pick up a GPU that overclocked really well?
By all means THIS is a good thing. though I really think they should rename the card GTX1065 instead as it will confuse the hell out of people. Some folks might only want the 5X GP105 variant rather than the old one. Not only does this mess with consumers. It mess with the people who reviewed the original 1060s but the difference is. Instead of downgrading the memory and the memory bus and not telling anyone about it. they beefed it up instead.
#stickersmakeadifference even though it's the same old die with the same old arch in it each year?
People buy what fits their needs and i'm sure calling them clueless to reinforce your narrow point of view won't help your case.
But who know maybe users may have the opportunity to unlock them themselves. I got back all my TMUs on my Fury to get it back to full “X” So while I may be pie in the skying here If that was a possibility they would fly off the shelves too n
By handicapping/platforming the 1060's performance, and because the 1070 is hard to find at a reasonable price, the jump to an RTX 2070 is much easier, they kind of force it this way.
I seriously can't think of anything more plausible than this being Nvidia readjusting their stack and trying to normalize higher price levels with smart positioning of products. It falls right in line with their BS of 'lots of Pascal cards in stock', as if they didn't see it coming. Of course they did, and they also knew RTX was going to be a hard sell.
I think you need to brush up on their own naming conventions and where they fall.
I'll say it again , them using "high end" silicon in a "lower end" product is proof of what I claim. Don't believe one second that Nivida would actually ship silicon that's truly reserved for high end parts like this. Realize for once that this is a company in the business of manufacturing ICs where transistor count and die space matters above all else and most aspects of their products are dictated by those two features.
But you can continue to believe anything you want, namely that die space is unrelated to how a product stack is built. It that were true, the industry would collapse in an instant. Thank God that's not the case.
580% is 5% faster than 480, 590% is 10% faster than 580.
Where the heck does the confusion come from?
Should you think first, post second, perhaps?
So nVidia and AMD can name their products whatever they choose to see fit, they don't have to please you, nor me nor anyone else. It's only creating confusion if you're grasping at straws, which in this case you are. As for the 960's success you can base that on 2 things, nVidia's PR which is still in a far better place that AMD ever was and the fact that most of the contracts went towards nVidia instead of AMD from suppliers like Dell, HP etc etc for prebuilt systems, if you want to find fault at why the 380 didn't do a better job in the sales graph start pointing fingers to AMD's camp.