TweakTown NewsRefine News by Category:
As we near the release date of the nVidia GTX 680, more and more charts are surfacing with purported performance, specifications, and the like. Today is no exception. There appears to be 6 new benchmarks run on the card, using an extreme-cooled Intel Core i7-3960X Extreme Edition processor (running at stock frequency), ASUS Rampage IV Extreme motherboard, 8 GB (4x 2 GB) GeIL EVO 2 DDR3-2200 MHz quad-channel memory, Corsair AX1200W PSU, and Windows 7 x64.
Benchmarks included 3DMark 11 (performance preset), Battlefield 3, Batman: Arkham City, Call of Duty: Modern Warfare 3, Lost Planet 2, and Unigine Heaven (version not mentioned, could be 1). All tests were run at a constant resolution of 1920x1080, with 8x MSAA on some tests (mentioned in the graphs).
NVIDIA are mere days away from releasing their Kepler-based GEFORCE GTX 680, but there is still secrets that are leaking through the cracks. It seems that NVIDIA could have baked a new anti-aliasing (AA) algorithm into the Kepler family of GPUs.
There are plenty of different forms of AA, including FXAA, which have enabled higher levels of image quality, all while not being as taxing on the GPU as MSAA. This leaves one area to improve, to offer better image quality (IQ) while lowering the performance penalty when its enabled. According to the above screenshot, which is rumored to be sourced from NVIDIA's GeForce Kepler press-deck, the new algorithm is compared to 8x MSAA, where it gives superior image quality.
We all know that 8x MSAA is not easy for a GPU and definitely takes an arrow to the knee in terms of performance degradation for a GPU. But, this new screenshot shows us an anti-aliasing method that does indeed look taxing, but if NVIDIA has rubbed their genie lamp, we could expect some amazing AA, where it won't be too taxing on the GPU itself. The latest NVIDIA rumor showed that when Battlefield 3 was running its AA, it had some stellar performance.
Chinese media site PCOnline.com.cn have released what they claim to be an snipper from the press-desk of NVIDIA's GEFORCE GTX 680 launch, which is scheduled for March 22. The specs look interesting indeed.
We should be expecting the GTX 680 to sport 1536 CUDA cores, a base clock of 1006MHz, boost clock of 1058MHz, 2GB of GDDR5 ram with a 256-bit wide bus, 6.0Gbps effective memory speed, power connector-wise we should see dual 6-pin, 195W total TDP, dual dual-link DVI, HDMI, DisplayPort 1.2 and finally, will arrive in PCI-Express 3.0.
If this is true, it goes hand-in-hand with most of the information that has slowly leaked over the last few weeks, and should be some great competition for the already kick ass Radeon HD 7970. Not long to go now everyone! Wave your Team Green flags, now!
Ah, NVIDIA charts. The memories are flooding back now, it feels like so long since I've seen them, but they're back, and in force. NVIDIA's latest charts are for their 28nm-built, Kepler-based NVIDIA GEFORCE GTX 680, caps intended.
If the above charts are true, which most of the time they are not because it's really just a closed-world internal test by NVIDIA themselves, then we're looking at an amazing 40-percent increase over the AMD Radeon HD 7970 in some tests. 15 tests were used, from 7 games/benchmarks. All benchmarks were run with AA, then again without AA.
The one test that got my GPU juices flowing was the Battlefield 3 test with 4xAA, look at those improvements. If only I had the time to play Battlefield 3, I'd probably get 1 or 3 of these new GEFORCE GTX 680's. In the end, it looks as though that crown AMD have graced on their head, is about to get a can of whoop ass thrown at it. Green Team, unite!
More details are arriving for NVIDIA's soon-to-arrive GEFORCE GTX 680, and these new specs are quite interesting. Fudzilla's source are saying that the new GTX 680 will sport two 6-pin PCI-Express power connectors to provide a total TDP of around 190W.
Higher performance per watt is expected from not only users of the card, but NVIDIA, too. Considering the card features 1536 shader units, 2 (or 4GB) of GDDR5 memory clocked at 6GHz, a reported 'very high' GPU clock, we should expect very good things from the GTX 680. Performance is said to be in the area of AMD's Radeon HD 7970, where it wins in some benchmarks, and loses in others. These results are from NVIDIA, so we'll have to wait for some real-world testing.
The GPU itself is said to be around 300mm² in size, too. Price-wise, we should expect a US$549 price, but that is to be expected. NVIDIA is launching a card with more performance than the GTX 580, with much better performance per watt numbers. We should expect some competition in pricing, but that comes down to GPU flavor from other companies such as ASUS, GIGABYTE, MSI, etc as well as heated up competition from the already-released HD 7970.
NVIDIA's [new capitilized text] GEFORCE GTX 680 is beginning to take form, with details leaking out not only about its Dynamic Clocking ability, where it can drive its clock speeds up during times of high-stress, benchmarking, gaming, and more, but its memory.
We should be looking at AIBs offering GTX 680 models with either 2 or 4GB of GDDR5 memory, and for the first time since GDDR5 was introduced, an NVIDIA GPU is able to drive the clock speed of the GDD5 memory higher than AMD, who created the memory standard.
VR-Zone are reporting that the Dynamic Clock Adjustment is similar to the "hot clocks", but the difference with the GK104 GPU is that it comes with many dozen power planes, that will operate on varying clock speeds depending on the computational load, card temperatures, and the power consumption. The full GK104 chip will operate at a lower clock speed, similar to those found on the GTX 480, with low power mode at 300MHz, standard at 705MHz which extends out to 950MHz, while the cores alone will reach 1411MHz when the chip reaches load of 100-percent.
We should expect 192GB/sec from the GK104's 2GHz QDR clocks (6GHz effective), less than the HD 79x0 cards, but more than the HD 78x0 cards.
Yesterday we gave you our first video unboxing of the MSI Z77A-GD65 from the Taipei Test Lab and today we bring you another involving two AMD Radeon 7800 series video cards from Sapphire. Lately we have been very impressed with the cooling solutions Sapphire has used on its video cards and from what we have seen today, this trend looks to continue here. In the video above we cover both the Sapphire Radeon HD 7850 OC Edition as well as the Sapphire Radeon HD 7870 OC Edition video cards.
The 7850 OC Edition features an impressive looking dual fan cooler with a single heatpipe which exhausts hot air out the back of your case. The bundle is rather impressive too including a card to join the Sapphire Select Club, quick installation guide, DVI to VGA connector, molex to PCI Express power connector, mini DisplayPort to DisplayPort adapter, HDMI to DVI connector, CrossFire bridge and a HDMI cable.
Next up in the video above is the 7870 OC Edition which in many ways is quite similar to the 7850 OC Edition, just bigger in all ways - the cooler is pretty much the same except it is bigger just like the retail box. The bundle is exactly the same as the 7850 video card expect it comes with an additional Molex to PCI Express power connector since the video card calls for it with its higher power draw requirements since it is of course a faster video card.
I bet you read that title, and thought "haha, yeah right", well, strap yourself in ladies and gentlemen. NVIDIA's 28nm-based Kepler GK104 GPU is being shown around the Game Developers Conference, as we speak, with Heise.de citing people it spoke to at CeBIT.
Let's do this: some models of the GeForce 600-series from NVIDIA will introduce a load-based clock speed-boost feature, similar to Intel's Turbo Boost, where the GPU will ramp up clock speeds to the GPU when it senses heavy loads. If there's a stressful scene to render, the GPU overclocks itself, gets through the scene, and goes back down to stock speeds. Wow.
This ensures that the GPU has higher minimum and average frame rates, which is a great thing to see. Another thing to point out, is the GK104 does actually sport 1,536 CUDA cores, which should give it a definite kick up the bum for number-crunching scenarios, such as shading, post-processing, and GPGPU-based work. But, the CUDA cores don't automatically give you super amounts of performance, unfortunately.
It's said to give the GTX 680 a 10-percent (or so) advantage of AMD's Radeon HD 7970 in Battlefield 3 for example, but be roughly 10-percent slower than the HD 7970 in 3DMark 11. Something else worth pointing out is that the GeForce GTX 680 will launch this month, and not be a paper-launch. Small quantities will be made available, and through just select AIC partners. Quantities should start building up over the later months.
VR-Zone have a very tasty nugget of information regarding NVIDIA's 28-nm based Kepler GPUs, the first of which is the GeForce GTX 680. We're looking at the GTX 680 sporting 1536 SP, and 256-bit memory. It was rumored to not feature shader hotclocks, but it does.
The core should be clocked in at 705 MHz, with shaders sitting at 1411 MHz, this continues to 2x shader hotclock from Fermi. GK104 also significantly improves upon the GF110/GF114's memory controller, where it pushes boundaries right up against the fence to 6 GHz. What this will provide is a memory bandwidth of 192GB/sec, identical to the GTX 580, even though it has a narrower memory bus.
The GeForce GTX 680 will sport 4 display outputs: stacked dual DVI, HDMI and DisplayPort. One disappointing feature is it only features two SLI connectors, enabling only 3-way SLI. Considering the GTX 580 could handle 4-way SLI, this is quite strange. The GTX 680 is said to be 10-inches long, 0.5-inches shorter than its predecessor. This makes it longer than the GTX 560 Ti or HD 7870, but shorter than the HD 7950 and GTX 570.
NVIDIA will reveal details of the GTX 680 on my birthday of all days, March 12th. We should then expect a hard launch (rumored) for March 23rd.
EXPreview is teasing that NVIDIA could release their Kepler-based GK104 GPU as soon as March 23. You know, as in, less than three weeks from now. Maximum-Tech "confirmed" GK104 to debut in April, but AIC (add-in card) vendors in China have said that GK104 "won't be postponed to April".
Operation Director of ZOTAC in the Asia-Pacific area said at Weibo:
GK104 launch won't be delayed to April, and the relative products have been ready.
The CEO of Galaxy, also revealed at Wiebo:
There are too many things (GK104) happening in March.
Looks like we could expect the GeForce GTX 670 Ti cards to hit the shelves and hearts of Team Green supports across the world, hopefully not in a paper launch way, by the end of this month, or early-April.
A band, Cobra Starship, had an album named after what I think this will be: Hot Mess. Not a hot mess in a bad way, but in a very, very good way.