Subject: Graphics Cards | May 16, 2012 - 10:40 PM | Tim Verry
Tagged: nvidia, gtx 680m, gpu, mobile, kepler
Videocardz.com managed to get their hands on some rumored details about an upcoming NVIDIA mobile graphics card–the GTX 680M. According to rumors, the mobile chip will be launched at Computex 2012 in Taiwan next month.
There aren’t many details about the mobile chip, but it is set up to be a scaled down version of it’s Kepler based GTX 680 desktop counterpart. The GTX 680M will have approximately half as many CUDA cores at either 744 or 768 cores depending on the source. Either way, the card keeps the same 256-bit memory interface and can support SLI configurations. In addition, the 680M will be able to have up to 4GB of GDDR5 memory. Reportedly, it can use as much as 100 Watts of power.
When paired with an Intel Core i7 3720QM processor, the GPU was able to get a score of 4,905 points in 3DMark 11’s Performance present benchmark. It is supposed to be as much as 37 percent faster than the GTX 670M, which is not surprising considering that chip has only 336 CUDA cores and is clocked at 598 MHz (no word yet on what the GTX 680M will be clocked at).
No matter what the GTX 680M turns out to be, you can bet it will only be found in the highest end gaming notebooks where performance is more important than battery life. Until then, feel free to brush up on your Kepler architecture knowledge by visiting our GTX 680 (desktop) review.
When the Fermi architecture was first discussed in September of 2009 at the NVIDIA GPU Technology Conference it marked an interesting turn for the company. Not only was NVIDIA releasing details about a GPU that wasn’t going to be available to consumers for another six months, but also that NVIDIA was building GPUs not strictly for gaming anymore – HPC and GPGPU were a defining target of all the company’s resources going forward.
Kepler on the other hand seemed to go back in the other direction with a consumer graphics release in March of this year without discussion of the Tesla / Quadro side of the picture. While the company liked to tout that Kepler was built for gamers I think you’ll find that with the information NVIDIA released today, Kepler was still very much designed to be an HPC powerhouse. More than likely NVIDIA’s release schedules were altered by the very successful launch of AMD’s Tahiti graphics cards under the HD 7900 brand. As a result, gamers got access to GK104 before NVIDIA’s flagship professional conference and the announcement of GK110 – a 7.1 billion transistor GPU aimed squarely at parallel computing workloads.
With the Fermi design NVIDIA took a gamble and changed directions with its GPU design betting that it could develop a microprocessor that was primarily intended for the professional markets while still appealing to the gaming markets that have sustained it for the majority of the company’s existence. While the GTX 480 flagship consumer card and the GTX 580 to some degree had overheating and efficiency drawbacks for gaming workloads compared to AMD GPUs, the GTX 680 based on Kepler GK104 has improved on them greatly. NVIDIA has still designed Kepler for high-performance computing though with a focus this time on power efficiency as well as performance though we haven’t seen the true king of this product line until today.
GK110 Die Shot
Built on the 28nm process technology from TSMC, GK110 is an absolutely MASSIVE chip built on 7.1 billion transistors and though NVIDIA hasn’t given us a die size, it is likely coming close the reticle limit of 550 square millimeters. NVIDIA is proud to call this chip the most ‘architecturally complex’ microprocessor ever built and while impressive, it means there is potential for some issues when it comes to producing a chip of this size. This GPU will be able to offer more than 1 TFlop of double precision computing power with greater than 80% efficiency and 3x the performance per watt of Fermi designs.
Subject: Graphics Cards | May 16, 2012 - 05:14 PM | Ryan Shrout
Tagged: nvidia, kepler, GTC 2012, gk110
We are posting live from the "Inside Kepler" talk at NVIDIA's GPU Technology Conference with details on the new GK110 GPU. Here is what we know so far:
- 7.1 billion transistors
- 15 SMX (modified) units
- 2880 available CUDA cores
- Greater than 1 TFLOP FP64 (double precision) compute
- 384-bit GDDR5 memory bus
The block diagram for the GK110 GPU
We will update this post with more photos and information as we have it!
Diagram of the updated SMX for GK110
Don't expect to see this GPU until at least Q4 of this year.
Subject: Shows and Expos | May 15, 2012 - 03:43 PM | Jeremy Hellstrom
Tagged: tesla, nvidia, GTC 2012, kepler, CUDA
SAN JOSE, Calif.—GPU Technology Conference—May 15, 2012—NVIDIA today unveiled a new family of Tesla GPUs based on the revolutionary NVIDIA Kepler GPU computing architecture, which makes GPU-accelerated computing easier and more accessible for a broader range of high performance computing (HPC) scientific and technical applications.
The new NVIDIA Tesla K10 and K20 GPUs are computing accelerators built to handle the most complex HPC problems in the world. Designed with an intense focus on high performance and extreme power efficiency, Kepler is three times as efficient as its predecessor, the NVIDIA Fermi architecture, which itself established a new standard for parallel computing when introduced two years ago.
“Fermi was a major step forward in computing,” said Bill Dally, chief scientist and senior vice president of research at NVIDIA. “It established GPU-accelerated computing in the top tier of high performance computing and attracted hundreds of thousands of developers to the GPU computing platform. Kepler will be equally disruptive, establishing GPUs broadly into technical computing, due to their ease of use, broad applicability and efficiency.”
The Tesla K10 and K20 GPUs were introduced at the GPU Technology Conference (GTC), as part of a series of announcements from NVIDIA, all of which can be accessed in the GTC online press room.
NVIDIA developed a set of innovative architectural technologies that make the Kepler GPUs high performing and highly energy efficient, as well as more applicable to a wider set of developers and applications. Among the major innovations are:
- SMX Streaming Multiprocessor – The basic building block of every GPU, the SMX streaming multiprocessor was redesigned from the ground up for high performance and energy efficiency. It delivers up to three times more performance per watt than the Fermi streaming multiprocessor, making it possible to build a supercomputer that delivers one petaflop of computing performance in just 10 server racks. SMX’s energy efficiency was achieved by increasing its number of CUDA architecture cores by four times, while reducing the clock speed of each core, power-gating parts of the GPU when idle and maximizing the GPU area devoted to parallel-processing cores instead of control logic.
- Dynamic Parallelism – This capability enables GPU threads to dynamically spawn new threads, allowing the GPU to adapt dynamically to the data. It greatly simplifies parallel programming, enabling GPU acceleration of a broader set of popular algorithms, such as adaptive mesh refinement, fast multipole methods and multigrid methods.
- Hyper-Q – This enables multiple CPU cores to simultaneously use the CUDA architecture cores on a single Kepler GPU. This dramatically increases GPU utilization, slashing CPU idle times and advancing programmability. Hyper-Q is ideal for cluster applications that use MPI.
“We designed Kepler with an eye towards three things: performance, efficiency and accessibility,” said Jonah Alben, senior vice president of GPU Engineering and principal architect of Kepler at NVIDIA. “It represents an important milestone in GPU-accelerated computing and should foster the next wave of breakthroughs in computational research.”
NVIDIA Tesla K10 and K20 GPUs
The NVIDIA Tesla K10 GPU delivers the world’s highest throughput for signal, image and seismic processing applications. Optimized for customers in oil and gas exploration and the defense industry, a single Tesla K10 accelerator board features two GK104 Kepler GPUs that deliver an aggregate performance of 4.58 teraflops of peak single-precision floating point and 320 GB per second memory bandwidth.
The NVIDIA Tesla K20 GPU is the new flagship of the Tesla GPU product family, designed for the most computationally intensive HPC environments. Expected to be the world’s highest-performance, most energy-efficient GPU, the Tesla K20 is planned to be available in the fourth quarter of 2012.
The Tesla K20 is based on the GK110 Kepler GPU. This GPU delivers three times more double precision compared to Fermi architecture-based Tesla products and it supports the Hyper-Q and dynamic parallelism capabilities. The GK110 GPU is expected to be incorporated into the new Titan supercomputer at the Oak Ridge National Laboratory in Tennessee and the Blue Waters system at the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign.
“In the two years since Fermi was launched, hybrid computing has become a widely adopted way to achieve higher performance for a number of critical HPC applications,” said Earl C. Joseph, program vice president of High-Performance Computing at IDC. “Over the next two years, we expect that GPUs will be increasingly used to provide higher performance on many applications.”
Preview of CUDA 5 Parallel Programming Platform
In addition to the Kepler architecture, NVIDIA today released a preview of the CUDA 5 parallel programming platform. Available to more than 20,000 members of NVIDIA’s GPU Computing Registered Developer program, the platform will enable developers to begin exploring ways to take advantage of the new Kepler GPUs, including dynamic parallelism.
The CUDA 5 parallel programming model is planned to be widely available in the third quarter of 2012. Developers can get access to the preview release by signing up for the GPU Computing Registered Developer program on the CUDA website.
Subject: Graphics Cards | May 11, 2012 - 04:57 PM | Ryan Shrout
Tagged: sli, nvidia, kepler, gtx 670, GK104, geforce
In our launch review of the GeForce GTX 670 2GB graphics card this week, we had initially mentioned that these $399 graphics cards would support SLI, 3-Way SLI and even 4-Way SLI configurations thanks to the pair of SLI connections on the PCB. We received an update from NVIDIA later on that day that in fact it would NOT support 4-Way SLI.
The message from NVIDIA was pretty clear cut:
"As I’m sure you can imagine, we have to QA every feature that we claim support for and this takes a tremendous amount of time/resources. For the GTX 680 and GTX 690, we do support Quad SLI and take the time to QA it, as it makes sense for the extreme OC’ers and ultra-enthusiasts who are shooting to break world records."
But with the similarities between the GTX 680 and the GTX 670, is there really any QA addition required to enable quad for 670? Seems like a cop-out to me man...
I saw it mostly as a reason to differentiate the GTX 670 and the GTX 680 with a feature since the performance between the cards was very similar; maybe too similar for NVIDIA's tastes with the $100 price difference.
Well this afternoon we received some good news from our contact at NVIDIA:
"Change in plans.....we will be offering 4-Way SLI support for GTX 670 in a future driver."
So while the 301.34 driver will not support 4-Way configurations with the GTX 670, 4-Way SLI will in fact be enabled after all in a future version. We'll be sure to keep you in the loop when that happens and the super-extreme enthusiasts can rejoice.
This does go to show that the fundamental differences between AMD's license-free and seemingly more "open" CrossFire technology and NVIDIA's for-fee SLI technology. With enough feedback and prodding in the right direction, NVIDIA can and does do the right thing, just look at the success we had convincing them to support SLI on AMD CPU platforms last year.
Feet to the fire everyone!
Subject: General Tech | May 11, 2012 - 11:38 AM | Jeremy Hellstrom
Tagged: sales, Q2, nvidia, kepler
NVIDIA made $925m this quarter down from the $1,002m they made 12 months ago and profit is even dimmer with profits falling from $137m to a hair over $60m. This marks the third year NVIDIA's Q1 revenue has been less than in the previous year and that is going to deeply trouble investors. Even if GTX680s and 690s had flooded the market and were sitting on store shelves hoping that someone would come along and buy them that would not have helped sales in the first quarter, though if Kepler had been released early and in great quantities NVIDIA might have turned this distressing trend around.
Q2 could be peachy, three models of GTX670 are still available at NewEgg after the initial sales and if the GTX680's production can be ramped up without much in the way of associated costs we could see some nice financials in the summer. After all they do have the best cards on the market right now. Hit up The Inquirer for more.
"Nvidia is on a high after a successful Kepler GPU launch but its financials paint a very different picture. The firm's first quarter of its 2013 fiscal year yielded revenue of $924.9m, just under four per cent lower than the same period a year previously, however its net income took a beating as profits fell by 55 per cent to $60.4m."
Here is some more Tech News from around the web:
- Tesla gun will probably kill something, sometime @ Hack a Day
- Asustek unveils US$799 ultrabook @ DigiTimes
- If data can be lost, it will @ The Tech Report
- Win HIS HD7870 IceQ Turbo and IceQ X Turbo X Video Cards @ Kitguru
- Weekly Gaming Giveaway #1: Conflict of Heroes: Awakening The Bear @ eTeknix
Subject: Graphics Cards | May 10, 2012 - 02:13 PM | Ryan Shrout
Tagged: nvidia, kepler, gtx 670, gtx 570, GK104, geforce
In the comments to our full review of the GeForce GTX 670 2GB graphics card a user asked for a comparison of the new GTX 670 against the Fermi-based GTX 570. I had some numbers for that already made up but ran out of time and space for it in our review that went up this morning. I thought that there might be others interested in this so I decided to put a news post with the results.
These performance graphs pit a reference clocked GTX 570 1.25GB card against the new reference GTX 670 2GB card.
Without a doubt the new GTX 670 is a faster GPU than the GTX 570:
- 3DMark11: +45%
- Battlefield 3: +34%
- DiRT 3: +28%
- Skyrim: +32%
- Metro 2033: +31%
- Dues Ex: +29%
- Batman: AC: +39%
- Power Consumption: -14%
With an average performance delta of 30% or more, the GTX 670 makes a solid upgrade for GTX 570 users but maybe more interesting, it does this while using 14% less power as well.
Subject: Graphics Cards | May 10, 2012 - 01:30 PM | Jeremy Hellstrom
Tagged: nvidia, kepler, gtx 670, GK104, geforce
The GTX 670 is a scaled down GTX 680 in every aspect including price. 1344 CUDA Cores is 192 less than the GTX 680, baseclock is reduced 91MHz to 915MHz, 16 less texture units for a total of 112; the memory remains the same at 6GHz. Most important to consumers is the reduction in price, down $100 to an MSRP of $400, targeting it directly against the HD7950 at its new price. [H]ard|OCP's testing does not favour AMD as the GTX 670 shows an obvious performance advantage over the HD7950 as well as the still available GTX 580 and does it at a price point which matches AMD's new prices.
Ryan's full review can be read here, where he tests out Galaxy's GTX 670.
"NVIDIA's next generation Kepler GPU continues with the launch of the GeForce GTX 670. This GPU is positioned to provide great performance at a price level $100 less than the GeForce GTX 680. Could this be the best performing $399 video card, besting even the Radeon HD 7950? We tell you all you need to know if $400 is your price range."
Here are some more Graphics Card articles from around the web:
- GeForce GTX 670 @ The Tech Report
- ZOTAC GeForce GTX 670 AMP! Edition 2 GB @ techPowerUp
- NVIDIA GeForce GTX 670 DirectX 11 Video Card Review @Hi Tech Legion
- NVIDIA GeForce GTX 670 Review Feat. EVGA: Bringing GK104 Down To $400 @ AnandTech
- Gigabyte GTX670 OC Video Card Review @ Ninjalane
- EVGA GeForce GTX 670 SuperClocked Video Card Review @ Hardware Secrets
- Palit JetStream GEFORCE GTX 670 2GB @ Tweaktown
- NVIDIA GeForce GTX 670 Video Card Tests @ Benchmark Reviews
- Palit GeForce GTX 670 JetStream 2 GB @ techPowerU
- EVGA GTX 670 Superclocked @ Overclockers.com
- NVIDIA GeForce GTX 670 2GB Video Card @ Bjorn3D
- GIGABYTE GTX 670 OC @ Bjorn3D
- Gainward GeForce GTX 670 Phantom @ Techspot
- Gigabyte GeForce GTX 670 Windforce 3X OC 2GB DirectX 11 Video Card Review @Hi Tech Legion
- NVIDIA GTX 670 Review @ OCC
- NVIDIA & EVGA GeForce GTX 670 2GB Video Card Review @ Legit Reviews
- GEFORCE GTX 670 @ Hardware Heaven
- NVIDIA GeForce GTX 670 Launch Review @ Neoseeker
- ASUS GeForce GTX 670 Direct CU II 2 GB @ techPowerUp
- Gainward GeForce GTX 670 Phantom @ Legion Hardware
- Nvidia GTX 670 @ lanOC Reviews
- NVIDIA GeForce GTX 670 2 GB @ techPowerUp
- NVIDIA GeForce GTX 670 Review @ Hardware Canucks
- NVIDIA GeForce GTX 670 @ Guru of 3D
- GeForce GTX 670 2 and 3-way SLI @ Guru of 3D
- Palit GeForce GTX 670 JetStream @ Guru of 3D
- ASUS GeForce GTX 670 DirectCU II TOP @ Guru of 3D
- Nvidia GeForce GTX 670 review plus SLI and 3-way SLI @ Hardware.Info
- NVIDIA GEFORCE GTX 670 2GB Video Card Overclocked @ Tweaktown
- MSI GTX 680 Twin Frozr Graphics Card Review @ HardwareHeaven
- ZOTAC GeForce GTX 680 AMP! Edition 2 GB @ techPowerUp
- Palit GeForce GTX 680 4GB Jetstream @ Guru of 3D
- NVIDIA GEFORCE GTX 670 2GB @ Tweaktown
- MSI Radeon HD 7870 HAWK 2GB @ Tweaktown
- MSI R7950 Twin Frozr III @ Kitguru
- Ivy Bridge PCI-Express Scaling with HD 7970 and GTX 680 @ techPowerUp
- HIS Radeon HD 7770 iCooler 1 GB @ techPowerUp
- XFX HD 7870 and 7850 Double Dissipation Review @ OCC
- MSI R7970 Lightning / R7870 Hawk review @ Hardware.Info
- HIS Radeon HD 7870 IceQ X Turbo X 2GB Graphics Card @ eTeknix
GK104 takes a step down
While the graphics power found in the new GeForce GTX 690, the GeForce GTX 680 and even the Radeon HD 7970 are incredibly impressive, if we are really honest with ourselves the real meat of the GPU market buys options much lower than $999. Today's not-so-well-kept-secret release of the GeForce GTX 670 attempts to bring the price to entry of the NVIDIA Kepler architecture down to a more attainable level while also resetting the performance per dollar metrics of the GPU world once again.
The GeForce GTX 670 is in fact a very close cousin to the GeForce GTX 680 with only a single SMX unit disabled and a more compelling $399 price tag.
The GTX 670 GPU - Nearly as fast as the GTX 680
The secret is out - GK104 finds its way onto a third graphics card in just two months - but in this iteration the hardware has been reduced slightly.
The GTX 670 block diagram we hacked together above is really just a GTX 680 diagram with a single SMX unit disabled. While the GTX 680 sported a total of 1536 CUDA cores broken up into eight 192 core SMX units, the new GTX 670 will include 1344 cores. This will also drop the texture units to 112 (from 128 on the GTX 680) though the ROP count stays at 32 thanks to the continued use of a 256-bit memory interface.
Subject: Graphics Cards | May 5, 2012 - 06:28 PM | Tim Verry
Tagged: nvidia, msi, kepler, gtx 670, asian colorful
Earlier this week we reported on a rumored MSI GTX 670 graphics card with a “DispalyPort” that was coming out soon. A new photo of the MSI card is below along with the box (though the typo is covered up by the card). Swelclockers has now gotten their hands on even more photos of the alleged GPU–this time from manufacturer Asia Colorful. The two rumored cards are said to be based on the NVIDIA reference design, and supporting the rumors, both cards do look very similar in design.
The alleged MSI GTX 670 GPU
The cards that appear to be Kepler based GTX 670s have a black PCB that extends about 3/4ths of the length of the heatsink and fan cooler. Except for a bit of overlap, the heatsink covers the PCB and the fan is located in the part of the card that hangs past the edge of the PCB. It also looks like they have moved the PCI-E power connectors to the outside edge of the card–a different rumored GTX 670 had the PCI-E power on the back edge of the card.
The underside of the Asia Colorful GTX 670 graphics card.
As far as specifications, the cards support PCI-E 3.0, two DVI outputs, an HDMI and what looks like a DisplayPort output. Beyond that, the card is said to be based on a scaled down version of NVIDIA’s Kepler architecture based GTX 680 GPU. For the GTX 670, NVIDIA has allegedly disabled 192 CUDA cores for a total of 1344. They have also reduced the base GPU core clockspeed from 1,006MHz on the GTX 680 to 900MHz. Although the card still uses the same 256-bit memory interface, the GTX 670 has a reduced memory (GDDR5) clockspeed of 5GHz versus 6GHz on the GTX 680. Because of the scaled back nature of the card, it has a lower TDP of approximately 150W.
Because they have been spotted in retailer’s hands, the GTX 670 is likely very close to release. Rumors are now suggesting that the card will be sold for somewhere between $349 and $379 USD.
Also on the Kepler front, is this cool looking single slot Kepler-class GPU that may or may not be a custom designed GTX 670. What Kepler rumors have you heard, and are you looking forward to the cheaper GTX 600 series cards?
Get notified when we go live!