NVIDIA today announced its latest line of mobile graphics processors (mGPUs) based on the new 40 nm manufacturing technology. The GPUs also come as a surprise, as they embrace two technologies very few expected them to, this early/late. The GPUs support the DirectX 10.1 API, and some feature the new GDDR5 memory interface that doubles effective bandwidth. Enter GeForce GTS 260M, GTS 250M, GT 240M, GT 230M, and G 210M.
The G 210M succeeds the GeForce 9400, at least as far as the manufacturing technology and clock speeds go. It holds 16 shader processors, a 64-bit GDDR3 memory interface, and 512 MB of memory. The GT 230/240M hold 48 shaders, 1 GB of 128-bit GDDR3 memory, and clock-speeds that make up each variant. The GTS 260/250M are especially new, as they feature a 128-bit GDDR5 memory interface. 96 shaders and 1 GB of memory make the rest of their specs., while their clock speeds make for the variants. The table below spreads out the specifications.
The peculiar thing about these GPUs is their timing of entry into the market. This is the fag-end of the DirectX 10(.1) generation of GPUs, with no more high-end GPUs in sight. The company is facing stiff competition from AMD, and needs to cut manufacturing costs, while also decorating specs sheets. The use of GDDR5 allows manufacturers to narrow memory bus width, and reduce the number of memory chips, in turn board footprint, while maintaining the same levels of performance as GDDR3 with higher bus width. The use of 40 nm manufacturing process is certainly a step in the right direction, although we wonder how the company is able to get large yields of these chips, considering AMD isn't able to keep up with demand for its 40 nm GPUs, and is developing alternative SKUs at the same price points.
<div class="table-wrapper"><table border="1" class="resulttable" cellspacing="0" cellpadding="3"><tr><th></th><td>GeForce<br /> G 210M</td><td>GeForce<br /> GT 230M</td><td>GeForce <br /> GT 240M</td><td>GeForce<br> GTS 250M</td><td>GeForce <br /> GTS 260M</td></tr><tr><th>Shader Units</th><td align="center">16</td><td align="center">48</td><td align="center">48</td><td align="center">96</td><td align="center">96</td></tr><tr><th>Memory Size</th><td align="center">512 MB</td><td align="center">1024 MB</td><td align="center">1024 MB</td><td align="center">1024 MB</td><td align="center">1024 MB</td></tr><tr><th>Memory Bus Width/Type </th><td align="center">64 bit GDDR3 </td><td align="center">128 bit GDDR3</td><td align="center">128 bit GDDR3</td><td align="center">128 bit GDDR5</td><td align="center">128 bit GDDR5</td><tr><th>Core/Shader/Memory Clock</th><td align="right">625/1500/800 MHz </td><td align="right">500/1100/800 MHz </td><td align="right">550/1210/800 MHz </td><td align="right">500/1250/800 (3.2 GT/s) MHz </td><td align="right">550/1375/900 (3.6 GT/s) MHz </td></table></div>These GPUs will only be available to the mobile GPU board segment to begin with. There is no indication as to when the company makes desktop graphics cards based on these. AMD launched the RV740 in a similar fashion, with a mGPU board being released weeks ahead of the desktop board (Radeon HD 4770).
62 Comments on NVIDIA Announces its First 40 nm, DirectX 10.1 Compliant GPUs
It's always great fun reading this douchebag bs and failed attempts to look smart and be funny when we all know just how much of a douchebag he is.
Gotta love Charlie the douchebag, he is so stupid and pathetic that he's kinda funny in a way. :nutkick:
:roll:
I suppose that means don't expect a mobile NVIDIA DX11 chip any time soon. They wouldn't be releasing DX10.1 mobile processors if DX11 was almost ready.
The reason that 240 beat 800 is quite simple. We're talking about a Superscalar architecture. Every 5 ALU's only have a single functioning 'brain'. Each of the 240 are SP's and have their own 'brain', as the architecture is Scalar, which is the next level upwards, just as Superscalar is to Vec4 (Vector - 4 ALU/'brain').
ATI doesn't do physics at all, actually. But, if they did, the performance would only be relative in the exact same way as real-world performance. Or, if nothing else, nVidia would have the upper hand. As I've said from the very beginning of G80 - Scalar is better and Superscalar is out-dated. It's not that ATI has 320 and nVidia has 128, or 800 and 240 and whatever comes next. It's that 1600 fail to beat 480.
I remember back in 2006. They had a fully-functional G71M that was identical in almost every way to G71 - the desktop variant and greatest GPU on earth, until R580, of course. R580 still is my favorite GPU.
I suppose they are moving to DX11 and just decided to show that.
Recently, there has been a move to create GPGPU (General-purpose computing on graphics processing units) that can been used for both as needed.
Where is this PhysX chip at exactly? Oh, that's right, there isn't one.
What is the green chip on the card for? I didn't see that on any ATI board I have seen.
7900 GT with 512MB of RAM. That, and the 7900 GTO - same as GTX but with lesser RAM.
Both.. honorable choices. Wish I had one.
The chip exists. Doubt your suspicions. Everything can be proven.
And if not, my 8400 GS can run it. It's healthy enough.