Nvidia’s 40nm mobile slides in detail

Agonizingly boring as usual

IT ALWAYS SEEMS to be the case that the quality of a slide deck is inversely proportional to the amount of interest it generates, and if my email is any indication, there are a lot of people interested in the Nvidia mobile 40nm deck. So, agonizingly boring as it is, we will share the notes we took on the presentation.

There are 20 slides, but of the whole presentation, only one, slide 18, is worth mentioning. We did earlier, here, in an attempt to save you from the pain. It didn’t work, people just love to be abused, so on we go.

Slide 1: This one has a picture of a cracked open notebook, with “Introducing Nvidia Geforce 200M Series Notebook GPUs”, with “Under embargo until June 15 2009, 6AM PST”. Luckily they don’t brief us or ask us to hold to embargoes, so we don’t have handcuffs. Our embargo ended when we were shown the deck a bit ago.

Slide 2: Here is where the Nvidia trademark unsubstantiated numbers kick off, with a slide called “Notebook Sales Continue to Outpace Desktop”, and a graph showing notebook units going from about 80K in 2006 to 180K in 2010. At the same time, desktop goes from 150-160K down to 120K. 2009 and 2010 have asterisks, but there is no explanation of it, nor of the numbers.

Slide 3: More self-congratulatory back patting. It is called “Big Week for Nvidia-Powered Notebooks”, with a subtitle of, “New Game-Changing Notebooks Across All Form Factors”. They are wrong, but given how prickly their egos are, we won’t tell them that repeating nonsense does not make it true. Three notebooks are listed, the Lenovo Ideapad S12, Dell Studio 14z and the Alienware M17x, with dates and pompous screed above.

Slide 4: This one tries to toe the line of good taste, but fails. They compare their Ion platform on the Ideapad S12 to ‘Intel UMA’, with a third column of “Nvidia Benefit”. All of the columns have checks in the NV column, but on the Intel one, only “Long Battery Life” is checked for both. On that, the “Nvidia benefit” is “Nearly identical power consumption”. This is flat out untrue, but when has NV ever been bothered by such things? Other Nvidia ‘benefits’ include “Powered by Nvidia cuda”. That is like Ford crowing that Chevy doesn’t have a Ford logo on the Corvette. Actually, it is more akin to Nvidia crowing that the wheels fall off their cars more often at higher speeds. Intel should sue over this.

Slide 5: A splash page for the Alienware M17x, calling it the fastest notebook ever. Once again, it isn’t, any of the ones powered by Intel desktop quads will eat it alive, but who cares about the truth?

Slide 6: This one is all about the GTX 280M, and at the bottom says they were launched at CeBIT. You know Nvidia is desperate for filler material when they have to trot out 3 month old parts to pad presentations. Nothing more to see here, but they try to say how great Physx and Cuda are again. We are still not buying it.

Slide 7: This one says that Windows 7 is coming. Ironically, none of Nvidia’s parts can do DX11 or Shader Model 5.0. Own goal guys. ATI was showing off running chips two weeks ago, contrary to rumors, NV still hasn’t taped out a DX11 part.

Slide 8: The claim here is that the new 200M parts use half the power of and have twice the performance. The slide doesn’t seem to finish the sentence so we left it hanging much as they do. This one, as you will see later is shoveling it on pretty thick, and using two disparate states that don’t relate to ‘prove’ a point. It is much akin to saying that your car can get 30MPG while moving at 25MPH, and also do 150MPH, then claiming it can get 30MPG at 150MPH. In any case, the numbers are totally unsubstantiated.

Slide 9: It claims 100+ design wins with ‘Top OEMs’, but won’t name a single one. This should clue you in about how loose the Nvidia definition of ‘top’ is.

Slide 10: Another bar graph that goes from 80%-280%, with the G100M at 100%, and the G210M at about twice that. If you recall, the G110M is the bottom of the barrel integrated graphics chipset, so they are comparing a discrete part to an integrated one. Luckily, they don’t point this out, then people would know what a joke of a comparison this is.

Slide 11: More G110M to G210M comparisons. Here is where they talk about lower battery life, claiming 50% less power used with the G210. At idle. They won’t disclose actual numbers though, that would blow out the claims of Slide 4. Intel publishes TDPs, Nvidia for some reason won’t. They also won’t publish the same slide with max power usage. Curious, isn’t it, almost as if they have a lot to hide.

Slide 12: It claims they are 5x faster at converting movies. No word as compared to what, or power consumed in doing so.

Slide 13: This one is puzzling, is it titled “Consumer GPU Computing Roadmap”, and lists programs by category totaling 385. No word as to what they do, or why, but there are 385 of them!!!1!!one!!1!!

Slide 14: Once again Nvidia really has nothing to say, so it is time to talk about PhysX. There are four companies, Sega, THQ, 2K and EA, and some engines and middleware listed. Nothing new, just a desperate attempt to make people think Physx matters. It doesn’t. Ask anyone from EA how much they are putting into Physx now that the Nvidia TWIMTBP money has dried up.

Slide 15: High humor here, the title is “6 new titles with breakthrough effects!”, and they list 13 titles. Of those, only one has actual gameplay that uses Physx, Mirror’s Edge. The rest are eye candy that can safely be ignored. It is pretty sad to see how few titles actually use Physx for anything other than particle effects. Enthusiasts seem to agree.

Slide 16: Once again, they retread the same old quotes about Nvidia having notebook driver releases. Nothing new, but Nvidia will repeat it again because the alternative is actually talking about their products.

Slide 17: This lists the new lineup in terms of market segments. The Enthusiast segment has the GTX 280M/260M carrying over and slotting in above the GTS 260M and GTS 250M. The GTS 160M seems to not carry over. Performance has the GT 240M and 230M slotting in above the GT 130M, which again does not carry over. Mainstream has the G210M above the now dead G110M. Wow. Or not.

Slide 18: Nvidia PR is learning, albeit slowly. You put the only slide worth a damn in the entire deck in near the end so people have to sit through the nauseating pablum to get the info. This one lists the same info we did earlier on the chips. See here for the full scoop, but prepare to be underwhelmed. This is the only slide in the deck worth bothering with should someone give you the set.

Slide 19: Repeats the bogus claim of half the power, twice the performance, along with the one about Windows 7 coming. They still don’t mention they don’t and won’t have DX11 product for the Windows 7 launch.

Slide 20: The end, titled “Thank You!”. Whenever an Nvidia presentation ends, there is reason to be thankful, and this tradition lives on today.

In the end, there is little to talk about 3 new dies, 5 new SKUs, and no new architecture. If it weren’t for the fact that these parts are made on the TSMC 40nm process, they would be forgettable. These are dull filler, the only parts of the line that are worth bothering with were put out at CeBIT, and even they underwhelmed all onlookers. Bleak times ahead for Nvidia.S|A

The following two tabs change content below.

Charlie Demerjian

Roving engine of chaos and snide remarks at SemiAccurate
Charlie Demerjian is the founder of Stone Arch Networking Services and SemiAccurate.com. SemiAccurate.com is a technology news site; addressing hardware design, software selection, customization, securing and maintenance, with over one million views per month. He is a technologist and analyst specializing in semiconductors, system and network architecture. As head writer of SemiAccurate.com, he regularly advises writers, analysts, and industry executives on technical matters and long lead industry trends. Charlie is also available through Guidepoint and Mosaic. FullyAccurate