Fatalmasterpiece Posted May 7, 2016 Share Posted May 7, 2016 http://www.geforce.com/hardware/10series/geforce-gtx-1080 The 1080 is coming sooner than expected :D Personally I will never be able to afford one but at least it means that it will drive down the price on older cards. Link to comment Share on other sites More sharing options...
Guest deleted2630050 Posted May 8, 2016 Share Posted May 8, 2016 Just read about it on a PCWorld article. I'm surprised a GTX 1080 would cost as much as a 980Ti does at the moment (around $600) And as long as it doesn't force me to use Windows 10 like Gen6 Intel CPUs will, then I would get one (If I had the money that is) Link to comment Share on other sites More sharing options...
FMod Posted May 8, 2016 Share Posted May 8, 2016 It's just a website. The 1080 is coming out when expected, late may. Too bad the price isn't right. A 980Ti can be bought for as "little" as $530 today, while $600 has been their steady price. For $600, and actual prices will likely stay above that figure at first, they're essentially pricing it at the same level as the 980Ti. Meanwhile, the card doesn't seem to be quite that impressive. High clock rates are a bad thing; high clocks mean they're sacrificing some efficiency for smaller cheaper silicon - offsetting much of the savings from a smaller node. On the memory side, while a relatively narrow bus was expected, the GDDR5X on it is run relatively slow, leaving the memory slower overall than the 980 Ti's. The card should edge out the 980 Ti overall (take NV's claims with a grain of salt, they have a right to puffery, legally protected exaggeration), but whether it will be by much remains to be seen. The 1070 won't even have the new-ish memory, just clocking the old GDDR5 even higher, which means more power draw and less room for overclocking. 8GB vs 6GB is a distinction without a difference, people who buy new high-end cards don't keep them around long enough for requirements to creep up. The only point is SLI for multiple display combinations. The above might be indirectly confirmed by the 1080 and 1070 having higher TDPs than the 980 and 970 respectively. The 30W difference between the two concerns me - given the 1070's power-hungry memory, I wonder how much of the chip they had to cut to achieve that. The performance gap between Nvidia's top chip and its cut-down version has been slowly widening ever since the 400 series, and it might be even wider now. But not necessarily, they could've cut the chip minimally and dropped the clock rate hard to make up for the hungry RAM instead. Of course, the new generation is still going to be worthwhile. Given the news, one bit of advice is clear - don't buy the 980 - the 1070 will be better in every way, and cheaper. The R9 Nano and Fury are also better, but they're not cheaper. On the other hand, the 980Ti may become the new value champion, not just among Nvidia cards, but all. With prices down to the $550 range, it's a wide chip with VRAM headroom and likely "almost 1080" performance. In other words, if you want a bit more power or memory for the future than the R9 Nano/Fury, it might still be the card to get. The third conclusion is that, if you're waiting for the 1070, might as well waiting for AMD's Polaris reveal to compare. The Polaris is expected to target the same $350-ish segment, so it's going to target that level of performance - who knows, it may end up faster than the 1070, with new memory and new features to boot. Or not. That's it with the practical side of things.On the less-practical side, Nvidia's strategy of rationing new technology is disappointing. Pascal was expected to be a mini-revolution, instead they delayed the big card, dropped HBM2, raised clocks to save on silicon, then gutted even GDDR5X from all but the top SKU. Rather than use the new tech to cover for the past years' stagnation (price/performance has only increased by 50% since 2012), they're releasing it in a trickle with each successive product calculated to be just good enough to sell. They better not be following in Intel's footsteps, which used its near-monopoly to freeze consumer CPU at 4 cores for the last 10 years with single digit performance gains over the last 5. Link to comment Share on other sites More sharing options...
Daiyus Posted May 9, 2016 Share Posted May 9, 2016 I'll be waiting for the GTX 1080 Ti and AMD Vega next year before upgrading. I want to go 4K on my single GPU mITX gaming HTPC. My gut just tells me that getting a GTX 1070/80 right now would be a mistake. I'm looking forward to seeing them put through their paces just to see the minimum of what I will be getting into next year, but like hell am I wasting the money to find out first hand at this point in time. My HD 7870 and 720p screen will last another 12 months yet. Not to mention I just don't have £1000+ sat around for a screen/GPU combo right now. Link to comment Share on other sites More sharing options...
Dark0ne Posted May 9, 2016 Share Posted May 9, 2016 Very disappointed with the lack of HBM2. I think I'll be fine with my 2x980Tis in SLI for now!!! Link to comment Share on other sites More sharing options...
obobski Posted May 10, 2016 Share Posted May 10, 2016 I agree with Dark0ne (although I have AMD GPUs, but same idea) - nothing that's really screaming "time to upgrade" and very much yet another card in the long line of "slow trickle" improvements that FMod mentioned. That said, I kind of like the new heatsink's aesthetic, if that's really its shape and not a sticker. :ermm: Something worth reading/mentioning when nVidia (and AMD) start talking about "TDPs" these days:http://www.tomshardware.com/reviews/nvidia-geforce-gtx-980-970-maxwell,3941-11.html This is not meant to be some sort of "scandalous revelation" or anything - modern GPUs do some very clever stuff to achieve great power efficiency (and it looks like Pascal is continuing that trend), but it appears that nVidia (and AMD) are providing "average" numbers for their power draw these days, as opposed to maximum values (which is what they used to do, with nVidia often over-stating the numbers to try and ensure compatibility). Will be interesting to see more in-depth power measurements as boards make their way to various reviewers. Link to comment Share on other sites More sharing options...
ddavies10693 Posted May 11, 2016 Share Posted May 11, 2016 Was kicking myself over buying the titan x SC recently from Amazon once I saw the announcement this card, but it sounds like you guys are seeing through their marketing and showed that it's a lot less superior than they say. Link to comment Share on other sites More sharing options...
FMod Posted May 11, 2016 Share Posted May 11, 2016 Hope you didn't pay the full $999 for it, much less over that. The Titan X, as all of the Titans, is kinda pointless for gaming. Makes for a great workstation card, though, cheaper than the upper-end Quadros and comes with more RAM. (Most pro apps no longer use quadro-specific features, it's become just about the VRAM now - all games use simplified low-poly models plus maps to save on memory, but the artists and design apps work in high-poly. High goes to millions of vertices, most games don't even support over 64k on a single object, one high-poly model can take more memory than the entire game it's being built for.) Link to comment Share on other sites More sharing options...
obobski Posted May 13, 2016 Share Posted May 13, 2016 Hope you didn't pay the full $999 for it, much less over that. The Titan X, as all of the Titans, is kinda pointless for gaming. Makes for a great workstation card, though, cheaper than the upper-end Quadros and comes with more RAM. (Most pro apps no longer use quadro-specific features, it's become just about the VRAM now - all games use simplified low-poly models plus maps to save on memory, but the artists and design apps work in high-poly. High goes to millions of vertices, most games don't even support over 64k on a single object, one high-poly model can take more memory than the entire game it's being built for.) Agreed. And Titan X, along with the new Maxwell Quadros, have the same FP64 performance as GeForce as well (only Tesla gets the high FP64 performance on the Maxwell; not sure if this is the same for Pascal or not - on Kepler the Quadro, Titan, and Tesla all had higher FP64 performance than GeForce); not really sure what the "big gotcha" for Quadro M6000 is over Titan X these days is as a result. Titan X isn't a bad card, that having been said, its just expensive for what it is. Link to comment Share on other sites More sharing options...
BlackRoseOfThorns Posted May 13, 2016 Share Posted May 13, 2016 GTX 1080 has 40-200% improved performance over stock Titan X in VR. It's not a scheme, this is just Nvidia response to VR hype this season. Judging from clocks screenshot from Zotac's new Firestorm utility tool, gtx 1080 will sit around 360GB/s bandwith (if the formula is same for gddr5x). This translates to roughly 5% improvement over my 980 ti. With custom PCBs and how easy this card is overclocked it might get up to 20% in normal games. Now the question is if you really need that boost to stay at 60fps. There is also a speculation that the new SLI bridge won't limit two cards as mush as the previous version. Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now