So the wait of several years has been now coming to the end as Today Nvidia has released it’s Maxwell Second Generation Architecture based Graphics Card Gtx 980 and Gtx 970, the new Maxwell architecture is based on lower power consumption and more thermal chip, the GM204 is the latest and improved version equipped with some new technologies. It will be better to take a quick recap of GM204 and new Technologies of architecture before moving to the GTX 900 Series Cards.
NVIDIA Releases Maxwell GM204 Based GTX 980 and GTX 970 Graphics Cards
Nvidia Maxwell GM204 Architecture:
The GM204 is equipped with lots of new features if we compare it with the First Generation of Maxwell Architecture which was previously used in Gtx750/750TI (GM107). Both Cards have won many in terms of performance per core and power efficiency, this is the main reason that this year Maxwell Second Generation Graphics Cards are most anticipated one. The Gtx 980 is the most flagship offering in history if we see such a huge Performance per core also. Also, if we see the prices this time price of both Cards are reasonable. The GM204 Chips have two variants, GM204-400 is used on the Gtx 980 and the GM204-200 is fused on the GTX 970. The fully enabled GM204 chip features 4 GPC (Graphics Processing Clusters) each group contains four SMM GPC units, each SMM unit also contains 128 CUDA cores, 8 texture units, and a polygon engine unit (PolyMorph Engine 3.0), a total of 2048 CUDA cores, 128 texture units, while the GTX 970 SMM group owns 13 units and 1664 CUDA cores thus making it around as fast as the GeForce GTX 780 while the GTX 980 will tackle the GeForce GTX 980 with a good 15-20% performance lead.
GM204 still maintain the GK104 core core four 64bit GDDR5 memory controller, but each group of ROP units has increased from 8 to 16, L2 cache from 512KB to 2MB, total 256bit bit wide, 64 ROP unit. In addition, Maxwell architecture SMM in each unit has a separate polygon unit, so the total number of polygons unit also doubled, which means Tessellation performance will exponentially increase.
Little change in the overall architecture of the cell is the texture units, Kepler architecture in each SM unit has 16 TMU texture units, now reduced to eight, but double the number of SM, so the total number of texture units is 128, but NVIDIA said they improved the efficiency of the texture units, so the texture fill rate performance under the same number of cases also improved to 12%.
Nvidia Maxwell Architecture New Technologies:
A side from hardware enhancements, GM204 is also equipped with some software improvements, Dynamic Super Resolution is the enhanced version of previous technology which will help to increase the video quality of 1080p resolution to the level of 4K. Then there’s Multi-Pixel Programming Sampling technology, which improves randomization of each sample and reduces quantification artifacts for better geometry processing and anti aliasing filtering. In terms of display Gtx 980 has an upgraded version of HDMI, HDMI 2.0 also there are three three Display Ports 1.2, and 1 DVI.
In total there are six major updates that enable the new algorithms of better picture quality if we compared it with previous Generations graphics cards.
- Conservative Raster
- Voxel Global
- Multi Projection
- Shader : Raster Ordered View
- Tilled resources
- Advance Sampling
The new Second Generation Macwell Architecture is equipped with new tiled resources and multi-projection technology for voxel grids (future VXGI) which enhances global illumination. The DirectX 11.2 API makes use of 3D Tiled Resources that allows hardware managed virtual memory for the graphics processing units.
MFAA and NVIDIA Dynamic Super Resolution:
Almost every time the introduction of new hardware, NVIDIA always brings some new technologies to improve the quality synchronization. With the debut of the second generation of Maxwell architecture, NVIDIA has brought a lot of new benefits, we say these two are most important. The first is the Multi-Frame Sampled AA) abbreviated as MFAA. Huge addition of a new anti-aliasing family members.
Technical principle is more complex, sampling mode screen is simply used for each pixel or was the spatial distribution or across multiple time frames interleaved distributed on a separate frames. NVIDIA claims that this “temporal integration” of anti-aliasing algorithm can get the best quality at the same time, performance is better than traditional MSAA (multi-sample anti-aliasing), presumably based on the cost of obtaining 8xAA 4xAA picture quality, 2xAA The cost of obtaining 4xAA quality.
According to official propaganda, anti-aliasing turned on the same grade, MFAA speed can average 30% faster than MSAA.
However, the technology is still being developed, yet currently open. NVIDIA will open later in more detail.
According to NVIDIA’s argument, it can achieve a 4K quality picture on your 1080p display. Independent of the technology and gaming, the vast majority can support very smoothly, only space for a small number of games, the button may be some display dislocation. NVIDIA GeForce Experience will be enabled by default, this feature, R343 series driver started to provide support.
Nvidia Maxwell Based Gtx 980 Specifications:
The GM204 Chips of Gtx 980 fuses 2048 CUDA Cores, 128 TMUs, 64 ROPs. The core clock is maintained at 1126 MHz core and 1216 MHz boost while the Memory clocked is around 7.00 GHz Effective which pumps out 224.4 GB/s bandwidth. The power supply interface for dual 6pin design, thermal design power 165W. The Output interface section provides a DVI, one HDMI 2.0 and three DP1.2. On power consumption ratio, GTX 980 compared to the GTX 780, the performance improvement of 25% in the case, the power consumption is actually reduced from 250W to 165W. From Gtx 980 is a built quality is good, the top notch Graphics Card featuring better performance per core and much power efficient than any other card of the same performance.
The PCB has been modified to a more brute design, NVIDIA can be seen using eight Samsung K4G41325FC-HC28 128M x 32. A total of eight of these modules has been featured, which equate to 4 GB GDDR5 VRAM across a 256-bit bus. The voltage controller has been moved below the power connectors and the power delivery includes 5 Phases compared to 6 on the GeForce GTX 780 Ti. The NVIDIA GeForce GTX 980 will retail at $549 US while non-reference models will retail at around $599 US pricing.
Nvidia Maxwell Based Gtx 970 Specifications:
GM204 Core on Gtx 970 fuses 1664 CUDA Cores, 64 ROPs and 104 TMUs. The Price of Gtx 970 is just $329. The GeForce GTX 970 will be available in both reference and non-reference variants at launch, which will retail at a range of $329 to $349 US. The new GPU is expected to be 20% faster and 30% more power efficient compared to Kepler based GTX 770. Dual 6pin power supply interface design, thermal design power of about 148W. If we compare it with Gtx 770, the GeForce GTX 970 raises the basic clock rate of 101 MHz and 130 MHz in boost clock.
The Gtx 970 is equipped with the DVI, HDMI and three display ports. The card uses HDMI 2.0.