The deluge of NVIDIA leaks continue following the major cyber-attack on the company, with hackers getting away with sensitive information about current and upcoming products. The latest in this series covers the memory bus widths of the next-generation RTX 40-series GPUs based on the "Ada Lovelace" graphics architecture. There is early-information covering the streaming multiprocessor (SM) counts of each GPU, and their large on-die caches.

The top-of-the-line AD102 silicon allegedly has a 384-bit wide memory bus, similar to its predecessor. The next-best AD103 silicon has a 256-bit wide memory bus. Things get very interesting with the AD104, which has a 192-bit wide memory bus. The AD104 is a revelation here, because it succeeds a long line of NVIDIA GPUs with 256-bit memory buses (eg: GA104, TU104, GP104, GM204, etc). This confirms the theory that much like AMD, NVIDIA is narrowing the memory bus widths in the lower segments to cut board costs, and compensate for the narrower bus-width with large on-die caches, high memory data-rates, and other memory-management optimizations.

It looks like NVIDIA bought itself a mountain of unsold GDDR5X memory chips, and is now refreshing its own mountain of unsold GP104 inventory, to make products more presentable to consumers in the wake of its RTX 20-series and real-time ray-tracing lure. First, it was the GP104-based GTX 1060 6 GB with GDDR5X memory, and now it's the significantly faster GeForce GTX 1070, which is receiving the newer memory, along with otherwise unchanged specifications. ZOTAC is among the first NVIDIA add-in card partners ready with one such cards, the GTX 1070 AMP Extreme Core GDDR5X (model: ZT-P10700Q-10P).

Much like the GTX 1060 6 GB GDDR5X, this otherwise factory-overclocked ZOTAC card sticks to a memory clock speed of 8.00 GHz, despite using GDDR5X memory chips that are rated for 10 Gbps. It features 8 GB of it across the chip's full 256-bit memory bus width. The GPU is factory-overclocked by ZOTAC to tick at 1607 MHz, with 1797 MHz GPU Boost, which are below the clock-speeds of the GDDR5 AMP Extreme SKU, that has not just higher 1805 MHz GPU Boost frequency, but also overclocked memory at 8.20 GHz. Out of the box, this card's performance shouldn't be distinguishable from the GDDR5 AMP Core, but the memory alone should serve up a significant overclocking headroom.

ZOTAC began rolling out its first refreshed GeForce GTX 1060 6 GB graphics cards that feature the more advanced GDDR5X memory, dubbed GTX 1060 6 GB G5X Destroyer. Since NVIDIA is carving out this RX 590-rivaling SKU from the larger GP104 silicon, ZOTAC appears to be using the same PCB as its GTX 1080 Mini, with 2/8 memory traces blanked out. The GP104 silicon is configured with the same 1,280 CUDA cores, 80 TMUs, 48 ROPs, and 192-bit wide memory interface as the GP106-based original, but with 6 GB of GDDR5X memory, albeit clocked at the same 8 GHz.

The card sticks to NVIDIA reference clock speeds of 1506 MHz core, 1708 MHz GPU Boost, and 8.00 GHz memory clock, although the 10 Gbps-rated GDDR5X memory chips could come with vast overclocking headroom. The card draws power from a single 8-pin PCIe power connector. Display outputs include three DisplayPorts, and one each HDMI and DVI-D. Although the card physically features SLI-HB fingers, ZOTAC clarifies that SLI is not supported for this SKU.

MSI rolled out one of the many upcoming silently-launched GeForce GTX 1060 6 GB graphics cards to feature the more advanced GDDR5X memory. NVIDIA is designing this SKU to compete with AMD's recently launched Radeon RX 590. Its specifications are very similar to those of the original GTX 1060 6 GB, but GDDR5X lends additional overclocking headroom. NVIDIA is carving this SKU out of the larger GP104 silicon, instead of GP106. You still only get 1,280 CUDA cores, 80 TMUs, 48 ROPs, and a 192-bit wide memory interface.

The MSI Armor OC ships with 1544 MHz GPU clocks, with 1759 MHz GPU Boost frequencies, compared to 1506/1709 MHz reference clock speeds. This factory-overclock is identical to the one the company's original GP106-based Armor OCV1 card ships with. The memory, surprisingly remains at 8.00 GHz, even though we suspect 10 Gbps-rated GDDR5X memory chips are being used in this card. The card draws power from a combination of 6-pin and 8-pin PCIe power connectors. The company's GTX 1060 6 GB Armor OCV1 graphics card only features a single 8-pin. Apparently you get SLI support. The company didn't reveal pricing.

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.

NVIDIA is reportedly giving finishing touches to its first serious GeForce-branded GPU based on a next-generation NVIDIA architecture (nobody knows which), for a late-July product announcement. This involves a limited reference-design "Founders Edition" product launch in July, followed by custom-design graphics card launches in August and September. This chip could be the second-largest client-segment implementation of said architecture succeeding the GP104, which powers the GTX 1080 and GTX 1070.

It's growing increasingly clear that the first product could be codenamed "Turing" after all, and that "Turing" may not be the codename of an architecture or a silicon, but rather an SKU (likely either named GTX 1180 or GTX 2080). As with all previous NVIDIA product-stack roll-outs since the GTX 680, NVIDIA will position the GTX 1080-successor as a high-end product initially, as it will be faster than the GTX 1080 Ti, but the product will later play second-fiddle to a GTX 1080 Ti-successor based on a bigger chip.

NVIDIA has a long history of playing Dr. Frankenstein on their graphics cards. The latest rumors from China suggest that NVIDIA is wheeling the GeForce GTX 1060 into the operating room for the fifth time. Currently, there are already four variants of the GeForce GTX 1060 in the wild: the original model with 6 GB of memory, the refreshed 6 GB model with slightly faster memory (9 Gbps), the cut-down model with 3 GB of memory, and finally, the Chinese-exclusive model with 5 GB of memory. However, NVIDIA's intentions are more ambitious this time. Apparently, their plan is to implement a variant of the existing high-performance GP104 GPU into the GeForce GTX 1060. As a reminder, the more powerful models like the GeForce GTX 1070, 1070 Ti, and 1080 all utilize the GP104 chip. Curiously, we've seen a variant of the GP104 (GP104-140) chip in a lower-end model before, specifically the GTX 1060 3GB.

On this occasion, NVIDIA is going to tailor the GP104-300, the silicon used in the GTX 1070, to the GTX 1060. This new SKU will carry the GP104-150 label. Despite the radical change, GP104-equipped GTX 1060 graphics cards share the same specifications as the original 6 GB model. Therefore, performance should be right in line as well. Although, we can expect some slight, unnoticeable differences in regards to thermals and power consumption. As usual, the new GTX 1060s will be exclusive to the Chinese market. With the next-generation of NVIDIA graphics cards almost among us, the company has been working diligently to clear their Pascal stock.

GIGABYTE today rolled out the GV-NP104D5X-4G, a mining graphics card based on NVIDIA P104-100 mining processor. Derived from the "GP104" silicon, the P104-100 is configured with 1,920 CUDA cores on this card, paired with 4 GB of GDDR5X memory across its 256-bit wide memory interface. What's more interesting, although the GP104 features a PCI-Express 3.0 x16 bus interface, the card has only x4 wiring, although the slot interface itself is of x16 length.

The GPU core clock speed for the GV-NP104D5X-4G is 1607 MHz, with 1733 MHz GPU Boost, and its memory is clocked at 10 GHz (GDDR5X-effective). The card draws power from a single 8-pin PCIe power connector, and has no display outputs. It's cooled by a rather simple aluminium fin-stack heatsink, ventilated by three fans, carried over from the company's latest WindForce 3X cooling solution. Unlike its consumer GeForce graphics cards, this card is backed by only a 3-month warranty. The card will be significantly cheaper than the GTX 1080 or GTX 1070.

GALAX' EU-regional wing KFA2 rolled out the GeForce GTX 1070 Ti EX-SNPR White graphics card. A step above the GTX 1070 Ti EX graphics card, the SNPR White comes with a white metallic cooler shroud, while fan impellers, white LED lighting over the fans, and a white metallic back-plate with RGB LED diffusers. RGB LEDs line the periphery of the back-plate, and the "GeForce GTX" decal on top of the card. The card ships with NVIDIA-reference clock speeds of 1607 MHz core, and 1683 MHz GPU Boost. Its memory ticks at 8.00 GHz (GDDR5-effective).

Drawing power from a combination of 6-pin and 8-pin PCIe power connectors, the card conditions it using a 7-phase VRM. It uses a custom-design aluminium fin-stack heatsink that's ventilated by a pair of 100 mm fans, which stay off when the card is idling. Display outputs include three DisplayPort 1.4, and one each of HDMI 2.0 and DVI-D. Based on the 16 nm "GP104" silicon, the GTX 1070 Ti features 2,432 CUDA cores, 152 TMUs, 64 ROPs, and a 256-bit wide GDDR5 memory interface, holding 8 GB of memory. The KFA2 GeForce GTX 1070 Ti EX-SNPR White is priced at 479€ (including VAT).

NVIDIA today announced a refresh of the performance-segment of its graphics card lineup, with the new GeForce GTX 1070 Ti. This card is positioned to fill the rather large performance and price/performance gap between the GTX 1070 and the GTX 1080. Based on the same "GP104" silicon as those two, the GTX 1070 Ti makes the "Pascal" architecture look fresh again, in the wake of AMD's Radeon RX Vega family launch. The GTX 1070 Ti is endowed with 2,432 CUDA cores, just 128 fewer than the GTX 1080. Its TMU count is proportionally lower at 152 (out of 160). It carries over its entire memory sub-system from the GTX 1070, in featuring 8 GB of GDDR5 memory clocked at 8.00 GHz, which works out to a memory bandwidth of 256 GB/s.

The GeForce GTX 1070 Ti also features marginally lower clock speeds than the GTX 1080, with its core clocked at 1607 MHz, with a GPU Boost frequency of 1683 MHz, compared to the 1506/1683 MHz clocks of the GTX 1070, and the 1607/1733 MHz of the GTX 1080. So it has the nominal clocks of the GTX 1080 and the GPU Boost clocks of the GTX 1070. NVIDIA appears to have a limited stash of GTX 1070 Ti Founders Edition (reference-design) cards, although it's unclear if the company will sell them as a separate SKU. The MSRP for this SKU is USD $429. NVIDIA's AIC (add-in card) partners will roll out their custom-design cards immediately.

Here are some of the first pictures of a custom-design NVIDIA GeForce GTX 1070 Ti graphics card, the KFA2 GTX 1070 Ti EX. For one, these pictures confirm the branding of NVIDIA "almost GTX 1080" SKU, positioned between its performance-segment GeForce GTX 1070, and its high-end GTX 1080. The KFA2 card ships with clock speeds of 1607 MHz core, 1683 MHz GPU Boost, and 8.00 GHz (GDDR5-effective) memory, which works out to a memory bandwidth of 256 GB/s.

Based on the "GP104" silicon, the GTX 1070 Ti packs 2,432 CUDA cores, 152 TMUs, 64 ROPs, and a 256-bit wide GDDR5 memory interface, holding 8 GB of memory. The KFA2 card draws power from a combination of 6-pin and 8-pin PCIe power connectors, conditioning it with a 7-phase VRM. An aluminium fin-stack heatsink to which heat drawn directly from the GPU and ventilated by a pair of 100 mm fans, keeps the card cool. The VRM is cooled by a secondary heatsink, while the memory is cooled passively. Display outputs include three DisplayPort 1.4, and one each of HDMI 2.0 and dual-link DVI-D. The card is expected to launch on October 26.

NVIDIA could severely limit the overclocking capabilities of its upcoming "almost GTX 1080" performance-segment graphics card, the GeForce GTX 1070 Ti. The company will tightly control the non-reference clock-speeds at which its add-in card (AIC) partners ship their custom-design graphics cards; and there could even be tighter limits to which you can overclock these cards. NVIDIA is probably doing this to ensure it doesn't completely cannibalize its GeForce GTX 1080 graphics card, which has been recently refreshed with faster 11 Gbps GDDR5X memory.

The GTX 1070 Ti is based on a "GP104" Pascal silicon with a core-configuration that's vastly higher than the current GTX 1070, and too close to that of the GTX 1080. It features 2,432 CUDA cores, just 128 fewer than the GTX 1080, and core clock speed of 1608 MHz that's on-par with the pricier card, too. The GPU Boost frequency is set to 1683 MHz, which is lower than the 1733 MHz of the GTX 1080. It also features slower GDDR5 memory. The GTX 1070 Ti is expected to launch by the 26th of October, priced at $429.

It turns out that NVIDIA is giving the GeForce GTX 1070 more than a minor refresh. The new performance-segment SKU, which is slated to come out just before Holiday 2017, could perform very very close to the GTX 1080, although sufficiently spaced out from the GTX 1080 refresh (featuring 11 Gbps memory). According to specifications leaked by Chinese tech publication MyDrivers, NVIDIA will give this SKU the coveted "Ti" moniker after all, and carve it out from the "GP104" silicon.

According to the report, the GTX 1070 Ti will be carved out of the "GP104" silicon by disabling just 1 out of 20 streaming multiprocessors, compared to the GTX 1070 desktop, which has 5 out of 20 disabled. This results in a CUDA core count of 2,432, which is just 128 fewer than that of the GTX 1080. The clock speeds of the GTX 1070 Ti are higher than the GTX 1070, too. It comes with a core clock of 1607 MHz, 1683 MHz GPU Boost, and an untouched 8.00 GHz (GDDR5-effective) memory clock. Interestingly, the TDP of this chip is 180W, which is the same as the GTX 1080. NVIDIA will reportedly launch the GeForce GTX 1070 Ti in early-November, 2017, at a price that's 12.5 percent cheaper than the GTX 1080.

NVIDIA is readying a new GeForce GTX 1070 refresh graphics card, according to well-placed sources. Positioned between the current GTX 1070 and the GTX 1080 11 Gbps in performance, the refreshed GTX 1070 could at least displace the current GTX 1070 from its price-point, if not replace it. NVIDIA could carve the new chip out of the latest stepping of the GP104 silicon, and give it more CUDA cores, likely 2,048 (on par with GTX 1070 Mobile), if not higher. It could also get faster memory, likely 9 Gbps GDDR5 or even 10 Gbps GDDR5X. Its core and GPU Boost clock speeds could even be dialed up a little.

NVIDIA's objective here appears to be convincingly outperforming AMD Radeon RX Vega 56, at a lower power-draw. There's a 20 percent performance gap between the current desktop GTX 1070 and GTX 1080, and the new GTX 1070 refresh could find a price-performance equation somewhere in the middle. As NVIDIA's product-stack currently stands, the GTX 1080, which was refreshed with faster 11 Gbps GDDR5X memory, has a wider performance gap with the GTX 1070, creating room for a GTX 1070 refresh SKU somewhere in the middle, which could perform within the 90th percentile of the original GTX 1080 with 10 Gbps memory. What NVIDIA could name the SKU is anybody's guess. Historically, NVIDIA has updated SKU specifications without changing the name. The GTX 1080 and GTX 1060 6 GB were refreshed with faster memory, by simply prominently mentioning the memory clock below the SKU branding, there's also the remote possibility of the GTX 1070 Ti branding to combat the "grandeur" of AMD's RX Vega branding. NVIDIA could have the new GeForce GTX 1070 refresh SKU out in time for Holiday.

GPU-accelerated crypto-currency mining poses a threat to the consumer graphics industry, yet the revenues it brings to GPU manufacturers are hard to turn away. The more graphics cards are bought up by crypto-currency miners, the fewer there are left for gamers and the actual target-audience of graphics cards. This is particularly bad for AMD, as fewer gamers have Radeon graphics cards as opposed to miners; which means game developers no longer see AMD GPU market-share as an amorphous trigger to allocate developer resources in optimizing their games to AMD architectures.

To combat this, both AMD and NVIDIA are innovating graphics cards designed specifically for crypto-currency mining. These cards are built to a cost, lack display outputs, and have electrical and cooling mechanisms designed for 24/7 operation, even if not living up to the durability standards of real enterprise-segment graphics cards, such as Radeon Pro series or Quadro. NVIDIA's "Pascal" GPU architecture is inherently weaker than AMD's "Polaris" and older Graphics CoreNext architectures at Ethereum mining, owing in part to Pascal's lack of industry-standard asynchronous compute. This didn't deter NVIDIA from innovating a lineup of crypto-mining SKUs based on its existing "Pascal" GPUs. These include the NVIDIA P104 series based on the "GP104" silicon (on which the GTX 1080 and GTX 1070 are based); and P106 series based on the "GP106" silicon (GTX 1060 series is based on this chip). NVIDIA didn't tap into its larger "GP102" or smaller "GP107" chips, yet.

NVIDIA's next-generation "Volta" GPU architecture got its commercial debut in the most unlikely class of products, with the Xavier autonomous car processor. The actual money-spinners based on the architecture, consumer GPUs, will arrive some time in 2018. The company will be banking on its old faithful fab TSMC, to build those chips on a new 12 nanometer FinFET node that's currently under development. TSMC's current frontline process is the 16 nm FFC, which debuted in mid-2015, with mass-production following through in 2016. NVIDIA's "GP104" chip is built on this process.

This could also mean that NVIDIA could slug it out against AMD with its current GeForce GTX 10-series "Pascal" GPUs throughout 2017-18, even as AMD threatens to disrupt NVIDIA's sub-$500 lineup with its Radeon Vega series, scheduled for Q2-2017. NVIDIA's "Volta" architecture could see stacked DRAM technologies such as HBM2 gain more mainstream exposure, although competing memory standards such as GDDR6 aren't too far behind.

TechPowerUp today released the latest version of GPU-Z, its popular graphics subsystem information, monitoring, and diagnostic utility, which no PC enthusiast can leave home without. Version 1.17.0 comes with support for new graphics chips, and a host of stability and usability improvements. To begin with, it comes with full support for Intel "Kaby Lake" and "Apollo Lake" HD Graphics (integrated graphics). It also comes with support for NVIDIA GeForce GTX 1050, GTX 1050 Mobile, GP104-based GTX 1060, and GP104-based Quadro Mobile.

GPU-Z 1.17.0 also comes with some user-interface changes, including display of Core architecture codename for Intel iGPUs, UEFI support being shown as available for Intel iGPUs newer than "Sandy Bridge" architecture, fixed ROP count on AMD "Beema" iGPU, and fixed values for Intel "Bay Trail" silicon to be 22 nm silicon process, and DirectX 11 as maximum API support. Grab GPU-Z from the link below.DOWNLOAD: TechPowerUp GPU-Z 1.17.0

Palit Microsystems Ltd, the leading graphics card manufacturer, announcing new flagship Palit GeForce GTX 1080 Dual OC which is the most advanced gaming GPU ever created, powered by the high-performance, energy-efficient new NVIDIA Pascal architecture. This GTX 1080 Dual OC edition delivers amazing performance, compact design, innovative new gaming technologies, and immersive, next-gen VR.

GeForce GTX 1080 is the most advanced gaming GPU ever created , the Palit GeForce GTX 1080 Dual OC is based on the NVIDIA Pascal architecture GP104 GPU armed with 2,560 CUDA cores and comes with ultra-fast FinFET and high-bandwidth GDDR5X technologies, plus support for DirectX 12 features to deliver the fastest, smoothest, most power-efficient gaming experiences. Palit GeForce GTX 1080 Dual OC is designed to deliver 3x more performance than previous-generation video card and its breakthrough innovations in gaming give gamers new possibilities to enjoy via VR experiences.

Due to the usual metrics and happenstances with foundry yields and wafer production, some chips contained in the production wafers are defective, with inoperative sections. This is always taken into account by companies, such as NVIDIA, while designing their product stacks (with the GTX 1070/1080 sharing the same silicon, and GTX 1070 samples being, mostly, defective versions of the fully-enabled Pascal GP104-140 chip). Other times, when supply of defective chips that can guarantee sufficient numbers of lower-tiered products, such as the GTX 1070, is insufficient to guarantee market demands (basically, things go better than expected at production), some sections of fully-operational chips are deactivated, so that it has the same working resources as the (otherwise defective) chips. Sometimes, like it happened with the Radeon HD 6950, these chips' resources can even be unlocked by simple BIOS flashing. According to recent reports, NVIDIA is bound to do something akin to that, by reusing GP104-140 chips on their GTX 1060 3 GB variants.

MSI today unveiled the GeForce GTX 1070 Quick Silver Edition graphics card. Going purely by the factory-overclock on offer, the card appears to be positioned below the GTX 1070 Gaming X graphics card by the company, although one can expect the company to monetize its unique design. The company also has a slightly pricier OC Edition variant of this card, which is identical to the GTX 1070 Gaming X in terms of factory-overclock. The card is characterized by a metallic silver color-scheme that's consistent with the company's GTX 1080 30th Anniversary Edition graphics card, and Gaming Titanium series motherboards, such as the X99A XPower Gaming Titanium. Shimmery silver accents run through bits of the Twin Frozr VI cooler shroud where you'd normally expect fiery red from the Gaming X, while the back-plate is entirely silver, with an illuminated MSI Gaming logo ornament.

The card comes with a factory-overclock of 1518 MHz core and 1708 MHz GPU Boost in Gaming Mode, and 1531 MHz core with 1721 MHz GPU Boost in OC Mode; which is below the factory-overclock of the GTX 1070 Gaming X, which comes with 1582/1771 MHz in Gaming Mode and 1607/1797 MHz in OC Mode, and barely above NVIDIA reference clocks of 1506/1683 MHz. The OC Edition variant, on the other hand, ships with identical clock speeds to the GTX 1070 Gaming X. The card draws power from a combination of 6-pin and 8-pin PCIe power connectors. Based on the 16 nm "GP104" silicon, the GeForce GTX 1070 features 1,920 CUDA cores, 120 TMUs, 64 ROPs, and 8 GB of GDDR5 memory clocked at 8.00 GHz, across a 256-bit wide memory interface. MSI didn't reveal pricing.

AMD CTO, speaking at an investors event organized by Deutsche Bank, recently announced that the company's next-generation "Vega" GPUs, its first high-end parts in close to two years, will be launched in the first half of 2017. AMD is said to have made significant performance/Watt refinements with Vega, over its current "Polaris" architecture. VideoCardz posted probable specs of three parts based on the architecture.

AMD will begin the "Vega" architecture lineup with the Vega 10, an upper-performance segment part designed to disrupt NVIDIA's high-end lineup, with a performance positioning somewhere between the GP104 and GP102. This chip is expected to be endowed with 4,096 stream processors, with up to 24 TFLOP/s 16-bit (half-precision) floating point performance. It will feature 8-16 GB of HBM2 memory with up to 512 GB/s memory bandwidth. AMD is looking at typical board power (TBP) ratings around 225W.

BIOSTAR is proud to announce its latest graphics card for gamers and VR enthusiasts, the BIOSTAR GeForce GTX 1070. Designed to offer an excellent balance of performance and price, consumers can now experience the latest generation of PC game titles and virtual reality experiences with the GeForce GTX 1070. BIOSTAR takes gaming to the next level once again with price, performance and efficiency that is significantly superior than the previous generation graphics cards with the all-new Pascal GPU architecture fabricated on the new 16 nm FinFET process technology.

The BIOSTAR GeForce GTX 1070 features the Pascal-based, high-efficiency GP104 GPU with a total of 1920 CUDA cores capable of running with a Boost core clock of 1683 MHz and base core clock of 1506 MHz. With GPU Boost 3.0, the BIOSTAR GeForce GTX 1070 can run up to and above 1683Mhz and features 8 GB of GDDR5 video memory wired to a wide 256-bit bus interface.

NVIDIA today announced the GeForce GTX 10-series for notebooks. The lineup includes three SKUs - the GeForce GTX 1080 Mobile, the GTX 1070 Mobile, and the GTX 1060 Mobile. Thanks to huge energy-efficiency gains with the "Pascal" architecture and the new 16 nm silicon fab process, this round of NVIDIA's mobile GPUs aren't "gimped out" in comparison its desktop discrete GPU lineup, in that they SKUs don't feature fewer CUDA cores to their corresponding desktop counterparts.

The lineup begins with the GTX 1080 Mobile. Based on the GP104 silicon, this chip features all 2,560 CUDA cores, 160 TMUs, and 64 ROPs physically present on the chip. The chip is endowed with 8 GB of 256-bit GDDR5X memory ticking at 10 Gbps. The core clock speeds remain unknown. The GTX 1070 Mobile is an interesting SKU in that it is better endowed than its desktop counterpart. It features 2,048 CUDA cores (the desktop GTX 1070 features 1,920), 128 TMUs, 64 ROPs, and 8 GB of 256-bit GDDR5 memory running at 8 Gbps. This chip features 1443 MHz core, and 1645 MHz GPU Boost. At the bottom of the pile is the GTX 1060 Mobile. Based on the GP106 silicon, this chip features 1,280 CUDA cores, 80 TMUs, 48 ROPs, and 6 GB of 192-bit GDDR5 memory running at 8 Gbps; with clock speeds of 1405 MHz core, with 1569 MHz GPU Boost.

Late last week, NVIDIA announced the TITAN X Pascal, its fastest consumer graphics offering targeted at gamers and PC enthusiasts. The reign of TITAN X Pascal being the fastest single-GPU graphics card could be short-lived, as NVIDIA announced a Quadro product based on the same "GP102" silicon, which maxes out its on-die resources. The new Quadro P6000, announced at SIGGRAPH alongside the GP104-based Quadro P5000, features all 3,840 CUDA cores physically present on the chip.

Besides 3,840 CUDA cores, the P6000 features a maximum FP32 (single-precision floating point) performance of up to 12 TFLOP/s. The card also features 24 GB of GDDR5X memory, across the chip's 384-bit wide memory interface. The Quadro P5000, on the other hand, features 2,560 CUDA cores, up to 8.9 TFLOP/s FP32 performance, and 16 GB of GDDR5X memory across a 256-bit wide memory interface. It's interesting to note that neither cards feature full FP64 (double-precision) machinery, and that is cleverly relegated to NVIDIA's HPC product line, the Tesla P-series.

BIOSTAR is proud to announce its latest graphics card, the BIOSTAR GeForce GTX 1080. Designed for the latest generation of games and gaming experiences, the GeForce GTX 1080 takes gaming to the next level with performance and efficiency that beats everything before it featuring the all-new Pascal GPU architecture fabricatedon the 16nm FinFET process technology.

The BIOSTAR GeForce GTX 1080 will have an MSRP of USD $799. The BIOSTAR GeForce GTX 1080 features the GP104 GPU sporting 2560 CUDA cores and runs with a base core clock of 1607Mhz. Featuring GPU Boost 3.0, the BIOSTAR GeForce GTX 1080 can run up to 1733Mhz and features 8GB of GDDR5X video memory connected a 256-bit bus interface.