With notebook sales continuing to outpace the desktop, NVIDIA has its sights set on notebook graphics domination and is today announcing five new additions to its line of GeForce 200M series GPUs.
The new parts are the first NVIDIA GPUs to built on a 40nm process and their notebook debut is testament to the growth of the portable market. NVIDIA's 40nm desktop GPUs, it seems, will have to wait.
So, what does the new line up of GeForce 200M GPUs have to offer? We'll get into detail shortly, but first we'll introduce the five new chips in table form. Existing parts are highlighted in grey, and NVIDIA's five new introductions are highlighted in green.
| GPU | G 110M | G 210M | GT 130M | GT 230M | GT 240M | GTS 160M | GTS 250M | GTS 260M | GTX 260M | GTX 280M |
|---|---|---|---|---|---|---|---|---|---|---|
| GPU clock | Up to 600MHz | 625MHz | 600MHz | 500MHz | 550MHz | 600MHz | 500MHz | 550MHz | 550MHz | 585MHz |
| Processor cores | 16 | 16 | 32 | 48 | 48 | 64 | 96 | 96 | 112 | 128 |
| Processor clock | 1,000MHz | 1,500MHz | 1,500MHz | 1,100MHz | 1,210MHz | 1,500MHz | 1,250MHz | 1,375MHz | 1,375MHz | 1,463MHz |
| Max memory | 1GB, GDDR3 | 512MB, GDDR3 | 1GB, GDDR3 | 1GB, GDDR3 | 1GB, GDDR5 | 1GB, GDDR3 | 1GB, GDDR5 | 1GB, GDDR5 | 1GB, GDDR3 | 1GB, GDDR3 |
| Memory clock | 700MHz | 800MHz | 800MHz | 800MHz | 800MHz | 800MHz | 1,600MHz | 1,800MHz | 1,900MHz | 1,900MHz |
| Memory interface | 64-bit | 64-bit | 128-bit | 128-bit | 128-bit | 256-bit | 128-bit | 128-bit | 256-bit | 256-bit |
| Manufacturing process | 65nm | 40nm | 65nm | 40nm | 40nm | 55nm | 40nm | 40nm | 55nm | 55nm |
| Underlying architecture | G92 | GT200 | G92 | GT200 | GT200 | G92 | GT200 | GT200 | G92 | G92 |
| DirectX / Shader Model | DX10, 4.0 | DX10.1, 4.1 | DX10, 4.0 | DX10.1, 4.1 | DX10.1, 4.1 | DX10, 4.0 | DX10.1, 4.1 | DX10.1, 4.1 | DX10, 4.0 | DX10, 4.0 |
| Peak GFLOPS | 48 | 72 | 144 | 158 | 174 | 288 | 360 | 396 | 462 | 562 |
Keeping things fairly simple, NVIDIA has retained its current nomenclature and its mobile GPUs are identified as G for mainstream parts, GT for performance, GTS for high performance and GTX for enthusiast.
Examining each of the five new products, we see that they're toned-down and shrunken derivatives of NVIDIA's GT200 architecture, albeit with far less potency than their desktop siblings.
Using a 40nm fabrication process and a refined GT200 architecture, NVIDIA's managed to lower the power use of each of the new GPUs, and without going into detail , NVIDIA reckons we'll see double the performance whilst consuming up to 50 per cent less idle power when compared to G100M-series parts.
The power savings, of course, are welcome, but let's not forget that NVIDIA's mobile GT200 implementation pales in comparison to its desktop GT200 parts. Desktop GT200-based GPUs, granted, have a power envelope that's far from suited to notebook form, but we were expecting a little more grunt from NVIDIA's mobile derivatives. The supposedly high-performance GTS 260M, for example, features just 96 stream processors - less than half the number found in the desktop's GTX 260. Certainly nothing to scare AMD's ATI Mobility Radeon HD 4800 series, we reckon.
But let's not jump the gun, could it be that NVIDIA is for once behaving sensibly? We've often lambasted the fact that GPUs simply get bigger, badder and more power hungry with each new launch. This time around, that's not the case - NVIDIA's new introductions are built on a smaller process, offer better performance than their G100M-series predecessors, and claim to use less power, too.
Performance fiends may have been expecting more, but there are a few upgrades worth mentioning. NVIDIA has finally adopted GDDR5 memory, and its latest GPUs finally offer support for DirectX 10.1 - both inevitable upgrades that will have the folks at AMD thinking "about time".
NVIDIA's efforts aren't exactly mind blowing, but its entire series of new mobile GPUs are arriving at an opportune time. With Microsoft's Windows 7 and Apple's Snow Leopard offering built-in GPU acceleration courtesy of DirectX 11's Compute Shaders and OpenCL, respectively, OEMs are keen to equip their notebooks with the latest GPUs - hence NVIDIA already claiming over 100 design wins for its G200M series.
Looking ahead, it's also logical to assume that one of the five new parts will displace the GeForce 9400M found in the majority of Apple systems, and the same part could form of the basis of ION 2 for the next round of netbooks.
NVIDIA might not be setting notebook performance alight, but it's certainly putting itself in a strong position. To hear what the Green Team has to say, check out our exclusive interview with NVIDIA's Matt Wuebbling below.
