-
Nvidia reveals Fermi GF100 Chip Details
Techtree News Staff, Jan 18, 2010 1610 hrs ISTNo information on clock speed, performance and pricing of the upcoming Fermi GPUs
Nvidia has revealed some information about its upcoming family of DirectX 11 supporting graphics cards. The Fermi architecture based Nvidia DirectX 11 supporting graphics cards were spotted at CES 2010. A few more details about the much awaited GF100 (Graphics Fermi 100) series graphics were revealed by Nvidia. Stay clear from any confusions for Nvidia clarified that GF100 graphic cards are meant for gaming.
Nvidia hasn't shared all details about the GF100 graphic cards and don't confuse the new details with product launch. At this moment only the GPU die shot is out and we can't count it as the final details of the GF100 cards. Of course, the GF100 cards would support DirectX 11 and direct compete with ATI Radeon HD 5000 series.
Nvidia GF100 has 512 CUDA cores or shader cores compared to the GT200 GPU chip's 240 CUDA cores. The GPU diagram suggests that the chip will carry four graphics processing clusters each will have four stream multiprocessors. Thus, the chip has 16 Stream Multiprocessors (SM) in total, each SM has Polymorph Engine, which basically comprises of geometry setup and processing unit. These Polymorph Engines handle Tessellation, Vertex Fetch, Viewport Transform, Attribute Setup and Stream Output. Each graphics processing cluster has one rasterizer engine.
In GF100, 512 CUDA cores are divided by 32 cores per Stream multiprocessor unit. Besides that the chip will have 128 Texture Mapping Units and 48 Raster Operation Units that are attached to 768KB L2 Cache capable of full read/write access. While the SM cores have 64KB of L1 Cache. This graphics cards will support GDDR5 memory standard obviously for wider memory bandwidth and will use 384-bit memory bus interface.
At this point, Nvidia hasn't announced more crucial information about GF100 cards like the Die Size (we believe it would be 40nm process), Clock speeds, performance, power usage and pricing.
Till Nvidia unveils the clock speeds and the power usage, the performance of these new cards can't be determined. Most important factor for the consumers is the price and that's how they weigh their options most of the time. For in-depth look in to how GF100's architecture works, take a look at AnandTech's article.
SMS REV to 59995 to get expert reviews on latest phones, laptops, cameras and more.
SMS TIPS to 59995 to get Tips and Tricks for phones, laptops, cameras and more.
Rs. 3 per SMS
Nvidia hasn't shared all details about the GF100 graphic cards and don't confuse the new details with product launch. At this moment only the GPU die shot is out and we can't count it as the final details of the GF100 cards. Of course, the GF100 cards would support DirectX 11 and direct compete with ATI Radeon HD 5000 series.
Nvidia GF100 has 512 CUDA cores or shader cores compared to the GT200 GPU chip's 240 CUDA cores. The GPU diagram suggests that the chip will carry four graphics processing clusters each will have four stream multiprocessors. Thus, the chip has 16 Stream Multiprocessors (SM) in total, each SM has Polymorph Engine, which basically comprises of geometry setup and processing unit. These Polymorph Engines handle Tessellation, Vertex Fetch, Viewport Transform, Attribute Setup and Stream Output. Each graphics processing cluster has one rasterizer engine.
In GF100, 512 CUDA cores are divided by 32 cores per Stream multiprocessor unit. Besides that the chip will have 128 Texture Mapping Units and 48 Raster Operation Units that are attached to 768KB L2 Cache capable of full read/write access. While the SM cores have 64KB of L1 Cache. This graphics cards will support GDDR5 memory standard obviously for wider memory bandwidth and will use 384-bit memory bus interface.
At this point, Nvidia hasn't announced more crucial information about GF100 cards like the Die Size (we believe it would be 40nm process), Clock speeds, performance, power usage and pricing.
Till Nvidia unveils the clock speeds and the power usage, the performance of these new cards can't be determined. Most important factor for the consumers is the price and that's how they weigh their options most of the time. For in-depth look in to how GF100's architecture works, take a look at AnandTech's article.
SMS REV to 59995 to get expert reviews on latest phones, laptops, cameras and more.
SMS TIPS to 59995 to get Tips and Tricks for phones, laptops, cameras and more.
Rs. 3 per SMS
Related Links
Tag keywords
Discussion Board
(4) Comments
alex
,Chicago, on Jan 18, 2010 11:15 PM
Report abuse
Ferdinand
,NJ, on Jan 18, 2010 05:24 PM
tipster
,City, on Jan 18, 2010 04:44 PM
