so all the 4GB cards are 8GB and it's an issue?
TYPICAL AMD! Giving more than promised for cheaper prices! I can see why they're going broke
I'm reminded of the phenom 2's.. Where you could get a quad core from dual and tri core chips.
I'm also reminded of like every one of their video cards since the 6000 series...
OH, you bought a 6850, 7850 or 7950. and saved $100... WELL don't worry, you can make that a 70 with no effort and no problems.
What AMD should be doing is making 3.5 GB cards, then SAYING they're 4GB.. that is a strategy that will work!
In other news.. I overclocked the shit out of one of my fury nano's, benched almost as good as a fury X. I promptly put it back to stock because it was hot, loud and I have a second one right there, so it was kind of pointless.
If I should Die Hard Arcade before I Alan Wake.
PREY the Lords of Lore my Soul Calibur to Take-Two Interactive.
ian wrote:
I'm reminded of the phenom 2's.. Where you could get a quad core from dual and tri core chips.
I'm also reminded of like every one of their video cards since the 6000 series...
OH, you bought a 6850, 7850 or 7950. and saved $100... WELL don't worry, you can make that a 70 with no effort and no problems.
Reminds me of what I think was the first AMD card I'd ever bought back in 2003. There was a way you could take a 9500 and essentially turn it into a 9700. Hey I even found an old article about it!
It would take the card from 4 rendering pipelines up to 8. I remembering buying the 9500 just so I could do that. If I remember right it wouldn't work on all 9500s. You had to get lucky and get a good card, otherwise when you enable the extra pipelines you would get some sort of artifacting. Luckily mine changed to a 9700 with no problem! I remember using 3DMark2001 to make sure it worked. Those were the days!
Nvidia unveils new GTX Titan X: 11 teraflops, 12GB GDDR5X, a mere $1,200
Potentially 30 percent faster than GTX 1080; 58 percent faster than the old Titan X.
Forget the GTX 1080: there's a new slab of graphics card hotness on the way from Nvidia, and its name is, er, the GTX Titan X. Yes, Nvidia has taken its most expensive graphics card and given it a Pascal-architecture makeover. $1200—UK price TBC, but probably £1,100—buys you 11 teraflops of FP32 performance, which is a significant 30 percent jump over the 8.9 teraflops of the GTX 1080, and a massive 58 percent higher than the 6.6 teraflops of the original Titan X.
The new Titan X launches on August 2 in the US and Europe. At first it'll only be available from the Nvidia website, but it will percolate down to other retailers soon after.
The Titan X is powered by a new chip, GP102, which packs in 3584 CUDA cores. While Nvidia hasn't revealed the amount of Streaming Multiprocessors (SMs), texture units, and the like, if the company uses a similar architecture to the GP104 chip (as used in the GTX 1080 and GTX 1070), expect a 40 percent boost in SMs over the GTX 1080 to 28. The chip runs at a 1417MHz base clock and 1531MHz boost clock.
Backing up GP102 is 12GB of GDDR5X memory running at an effective 10GHz and attached to a wide 382-bit bus, resulting in a 480GB/s of memory bandwidth, or a 50 percent increase over the GTX 1080. Power is provided by one 8-pin and one 6-pin PCIe connector, with a max TDP of 250W, the same as the previous Titan X. Connectivity consists of DisplayPort 1.4, HDMI 2.0b, and dual-link DVI. Externally, the Titan X uses the same multifaceted cooler of the GTX 1080, albeit now finished in black.
Interestingly, while the new GTX Titan X features the same number of CUDA cores as Nvidia's Tesla P100 GPU—as used in deep learning and science applications—it is using a different chip. With previous Titans, Nvidia has simply taken its biggest chip and given it a prosumer makeover, but that doesn't seem to be the case here. Those hoping for a return of the stellar FP64 performance of the original Titan may be disappointed here, although Nvidia is pushing the card's 44 TOPs of INT8, a measurement for neural network inference performance.
It's also a wee bit disappointing not to see HBM2 being used and Nvidia opting for GDDR5X. Yes, HBM2 is still in its infancy (the P100 is the only GPU that uses it) and costs are high. But given this is Nvidia's flagship GPU, and it costs $1200, its inclusion would have been a pleasant surprise. It would also have given AMD a poke, especially since the company is planning to use HBM2 in its upcoming Vega architecture.
The reveal of the Titan X follows the launch of the GTX 1080, the GTX 1070, and the GTX 1060, all of which were released over the past two months. That's a fierce release schedule by anyone's standards. Quite why Nvidia is so keen to get the Titan X out of the door, particularly when it has zero competition at the top end of the market, is something of a mystery. Still, with 11 teraflops, the Titan X might finally be the card that gives us sweet 60FPS 4K gaming.
Given that the performance of Titan typically trickles down to other cards a year later, expect to be able to able to play in 4K60 for a much more reasonable price next year.
The latest Polaris GPU is a 1080p budget powerhouse.
In our recent RX 480 4GB vs 8GB face-off, we said that while the more expensive card faced a worthy challenger in the form of GTX 1060, AMD's cheaper model was in a class of its own. Well, that's no longer the case - the RX 470 is almost as good and well worth checking out.
if a small rant be allowed, WTF is wrong with updating my gfx? tried to update with the Nvidia updater & it says it cant do it, it said nothing of this when checking for updates, now I cant even reinstall the drivers that came with the card as that gives the same problem so now I'm stuck with the generic Windows one it seems. All I wanted was to update my 960's drivers, guess I shoulda left it alone
£190/$199 price-point takes the fight to AMD's Polaris GPUs.
Nvidia has announced that a new GTX 1060 is due for release over the next few weeks, this time featuring 3GB of framebuffer memory. It's based on the same GP106 processor as the 'full-fat' GTX 1060 6GB, but there's an important difference in addition to the VRAM differential - this one has 1152 CUDA cores compared to the full complement of 1280 in the top tier model. Pricing is interesting - it'll cost $199 in the USA, with a £189 cost in the UK.
It's a bold move on Nvidia's part, and places the cut-down GTX 1060 on collision course with two AMD products - the RX 470, which we rather liked, and the four gig version of the RX 480 - assuming you can find one. Availability there, especially on the versions with an aftermarket cooler appears to be thin on the ground.
Based on the performance we've seen from the fully enabled GTX 1060, this cut-back model should still present plenty of challenge to the brace of sub-£200 AMD Polaris offerings. We're looking at 90 per cent of the core count found in the more expensive model (costing £40-£50 more), and half the VRAM, but memory bandwidth remains unaltered at 192GB/s. The same 8gbps GDDR5 modules are utilised. On top of that, base and boost clocks remain static.
It'll be a fascinating face-off in what is set to be one of the most fiercely competitive market sectors - one that AMD has dominated with its Polaris launches. We expect to see some interesting benchmarks based on the existing GTX 1060 data, but the challenge facing Nvidia is selling a 3GB GPU up against its competition where we see four gigs as the minimum.
Generally speaking, the more memory you have, the more future-proof you are. However, making matters somewhat less clear-cut is the issue of memory compression. Nvidia has been utilising it for some time, while AMD debuted it in its first product based on the Tonga processor - R9 285. Since then, it has been further iterated in its Polaris architecture. However, Nvidia hasn't sat still either - its last-gen Maxwell cards have excellent memory management, and this has been improved still further for Pascal. Based on our recent GTX 780 Ti testing in our GTX 1080 review, memory utilisation on challenging 1080p titles even holds up in the old Kepler cards too.
With that in mind, a key point of enquiry in our upcoming GTX 1060 3GB review will be memory utilisation and how it compares between Pascal and Polaris - and thankfully there are several titles we can turn to here in really putting this element of the card through its paces. A new driver to support the 3GB card is due today and Nvidia expects stock to filter through into retail "over the next few weeks" - our understanding is that the launch has been pulled forward, which may have left third party card vendors a touch unprepared. We aim to get hold of a sample as soon as possible and will roll out a comprehensive review some time next week.
Hawq wrote:if a small rant be allowed, WTF is wrong with updating my gfx? tried to update with the Nvidia updater & it says it cant do it, it said nothing of this when checking for updates, now I cant even reinstall the drivers that came with the card as that gives the same problem so now I'm stuck with the generic Windows one it seems. All I wanted was to update my 960's drivers, guess I shoulda left it alone
It could be worse.. you could have crossfired AMD cards and windows 10!
Thank god I have good enough cards to use one and still play everything maxed out at even 1440P
And I'll be very interested to check this new 1060 out..
If the price is right here, it could be the new 970 in bang for buck.
If I should Die Hard Arcade before I Alan Wake.
PREY the Lords of Lore my Soul Calibur to Take-Two Interactive.
New specs on Nvidia GTX 1050 may have leaked ahead of launch
Nvidia’s recent Pascal refreshes have focused mainly on the high end of the market, though that’s changed in the past month with the GTX 1060 hitting the $200 price point. Now, specs of the upcoming GP107 / GTX 1050 have supposedly surfaced. While all such data must be taken with a grain of salt, the specs make provisional sense.
According to BenchLife, the upcoming part will be a 768:64:32 core (that’s cores, texture units, and ROPS). That’s still a significant step down from the GTX 1060, which offers a base configuration of 1280:80:48 at 6GB and 1152:72:48 at 3GB. The GPU reportedly has a 128-bit memory path for 112GB/s of memory bandwidth altogether.
The GTX 1050 is a very different card than the RX 470 it would presumably compete against. AMD’s 4th-generation GCN is a much wider beast, with a 2048:128:32 configuration. Historically, the GTX 1060 – 1080 have all wielded a significant clock speed difference over GCN, but that may not be the case here — the maximum boost clock for the GTX 1050 is supposedly 1380MHz, and while Pascal GPUs with Nvidia Boost 3.0 tend to hit much higher clocks, we doubt the GPU is designed to leap to 1900MHz+ from a 1380MHz maximum boost. AMD’s RX 470 has 211GB/s of memory bandwidth — on paper, it’s significantly stronger than the GTX 1050 in multiple respects.
To-date, however, Pascal has compensated for seemingly weak paper specs with extremely strong real-world performance. The GTX 1060 we have in-house, for example, is virtually identical to the GTX 980’s performance despite having markedly less memory bandwidth, fewer shader cores, and fewer texture units.
GP107 supposedly won’t drop until mid-October, which gives both AMD and Nvidia some time to fix their pricing and availability problems. As things stand, a GTX 1050 at $150 would compete against the RX 460 4GB rather than the RX 470 4GB — and that’s not a comparison that would end well for AMD. By the same token, however, we’re pretty tired of the limited relationship between what AMD and Nvidia say a GPU will sell for and what it actually sells for. Only the GTX 1060 has managed to hit its recommended price targets within what we consider a reasonable amount of time, and only barely. Both companies need to shore up their availability and hit their MSRPs, and we’re going to treat all launch pricing from both manufacturers as provisional at best until we see proof they can keep their word.
The only good thing about launch pricing from AMD and Nvidia as of this writing is that prices are slowly creeping towards their official levels. By the time the GTX 1050 launches, card prices should be better aligned with where they were supposed to be months ago.
hNVIDIA GeForce GTX 1050 Ti to cost 139 USD, GTX 1050: 109 USD
We can finally reveal the pricing of GeForce GTX 1050 series, which are expected to launch next week.
The GeForce GTX 1050 Ti based on full-fat GP107 GPU will feature 768 CUDA cores, 4GB GDDR5 memory and 75W TDP. According to leaked slide GeForce GTX 1050 Ti will cost 139 USD, which is 20 USD less than GTX 950 at launch.
The GTX 1050 non-Ti will feature 640 CUDA cores and 2GB GDDR5 memory, while TDP will be the same as Ti’s model — 75W. This card however, will cost 109 USD.
Both cards will launch on the same day, which is October 25th.
Now here are GeForce GTX 1050 series from 9 major AIBs:
A non-Kotaku link for anyone refusing to support that site (like myself). Great news, Vulkan is an awesome API, I just wish more developers took advantage of it.
