NVIDIA's Linux Fate

Just like everyone else on the globe, I heard/read/watched Linus telling NVIDIA, in absolutely no uncertain terms, that they were bad Linux citizens. At first I thought Linus was telling them that they were "number one!!!" but sadly this turned out not to be the case.

The topic came up from someone in the audience who was trying to get NVIDIA's Optimus tech working on a Linux notebook. This is a bit of a different bird than NVIDIA's usual discrete GPU market, as this requires co-operation between onboard (or on-die) video adapters, the discrete graphics card and the northbridge to turn off the accelerated GPU when not in use. To my knowledge only Windows 7 drivers can accomplish this, and Linux doesn't quite understand the management architecture to leverage it. The open-source Bumblebee Project appears to be doing a great job of trying to set ACPI & BIOS flags to accomplish the same thing, but the entire effort could be greatly accelerated with NVIDIA's help. While NVIDIA has more-or-less sanctioned the Bumblebee project by releasing installers that work with Bumblebee, they haven't contributed to it directly.

Linux called NVIDIA's lack of Linux support the "exception rather than the rule," and extended that comment to mention that "NVIDIA been one of the worst trouble spots" in kernel support and the "single worst company we've ever dealt with." This of course was punctuated with the infamous line: "so NVIDIA, f**k you." He didn't actually speak in astrisks, I'm just experimenting with censorship.

Linus, of course, just "like[s] being outrageous at times... [because] people who get offended should be offended." It was probably meant to be a thrown gauntlet more than a haymaker, and I can definitely see why. I've been a big advocate of NVIDIA, even with their proprietary stance on GPU drivers and increasingly as the nouveau driver has increased its feature parity. These kind of ACPI hacks however are not the same as keeping tight control over intellectual property tied to their GPU hardware. Given how good drivers are the true value behind discrete graphics cards, I understand not wanting to open-source them to competitors ("Hey AMD! Here's how we did tessalation!"). Controlling power management within the OS kernel is a different matter however - and everyone benefits (including NVIDIA) with broader support. What's the barrier to entry of AMD creating something similar that idled the discrete GPU when no accelerated calls were being made? For that matter, what's to stop Microsoft from coming up with a vendor-agnostic solution that does the same thing? It seems not unlike what OS X Lion is doing right now...

NVIDIA's response basically focused on the fact that they permit Bumblebee to co-exist, their proprietary GPU drivers still offer fantastic performance and support (which they do) and they are one of the largest contributors to Linux on ARM. In short: they don't feel like Optimus Linux support is worth doing right now. However there are signs that even this isn't working well - NVIDIA just lost a quarter million dollar GPU deal due to poor MIPS support.

NVIDIA, this is a great time to step up and contribute the solution to the Linux kernel. Show your competitive edge continues to be broad platform support, otherwise people may meander elsewhere.

Post script: If you're interested in getting a sense of how crazy GPU/graphics driver development is on Linux, read Jasper St. Pierre's post on The Linux Graphics Stack. Terrific outline on the steps a polygon goes through to get to your monitor.

openSUSE 12.1 - Good Enough to Blog About

I've been away from the blogging scene for over eight months now. My language skills have languished likewise. It seems only fitting that my return to posting is launched on the same topic I departed with.

For too long it seems like openSUSE has been stuck within an odd, "uncanny valley" of user experience and desktop usability. It did so much well that the stuff that worked poorly really stuck out like a sore thumb. I've never really had the problems with YaST that others of my kin seem to decry ("Omigosh, you use SuSE? Ugh, is YaST any better?"), but other issues like poor integration points (e.g. Eclipse) and the banishment of SCPM and SaX2 were frustrating.

openSUSE 11.4 ended up looking okay however had issues with wireless, the noveau driver wasn't ready for prime time and LibreOffice ended up being crashtastic. Luckily (I guess?) I had a hard drive crash right before openSUSE 12.1 launch and was able to start clean when the mirrors sync'd on November 16th. As soon as openSUSE 12.1 launched I grabbed an ISO and began installing on my ThinkPad W510.

Man, what a difference.

First off, the noveau driver is still a bit buggy but has nearly reached parity with the proprietary NVIDIA driver on desktop compositing and UI effects like wobbly windows. I can't tell you how jazzed I was about this lil' facet. Don't get me wrong, I absolutely adore NVIDIA and the work they've done to bring driver support to the Linux desktop. Still, an OSS solution that has native integration into XOrg provides a seamless experience - especially with xrandr and KDE 4. Now KDE (and to a lesser extent, the bootloader) can natively manage resolution, orientation and multiple desktops. I was a bit apprehensive given I frequently switch between projectors, dual-screen setups with larger monitors and road-warrior setups with just the laptop display however noveau and KDE 4 are able to switch between these environments without so much as a mouse click. Plug in a VGA cable and boom - I'm broadcasting a second desktop over the monitor. Sweet.

LibreOffice is much more stable. No huge features I've been able to discern, but then again I'd likely not even notice them given my superficial use of any office suite. Evolution seems a bit smoother and more responsive - hopefully someday soon my office can switch to Exchange 2008 so I can leverage its new SOAP API via Evolution. That should make a big difference over the webdav connectors currently being used.

I've been trying out new features such as brtfs on alternate workstations, since encrypted brtfs partitions are not supported yet. So far it has been quite nice having snapshot capability and b-tree balancing to a file system, especially given how many crazy files I generate or move around.

Still very happy with Digikam; it has become my mainstay for post-production of RAW photos. It somewhat softens the harsh reality that my wife's iPhone 4S take better photos than my Canon DSLR.

All in all, a major tip o' the hat to the openSUSE team on all fronts. Very happy with the usability of the latest release.

ARM and a Teg

It has been interesting to passively watch news concerning embedded processors and system-on-a-chip designs since I started my irrational rationalizations about smartphones. The landscape is much wilder than conventional notebooks/desktops.

The weird thing is that the Qualcomm MSM7201A gets anecdotal low marks on the speed of its accomplice platforms, but Qualcomm claims that it has an ARM11 processor, an ARM9 modem, Java acceleration and can do 3D acceleration with 4 million triangles per second with a 133 million pixels per second fill rate. Perhaps hardware sporting the Qualcomm chipset has similar problems to my old VIA EDEN - acceleration was largely borked at the driver level. There seems to be a resonant believe with many developers/users that the 3D acceleration just isn't there, although with Android maybe it is. Either way it doesn't inspire a lot of confidence.

NVIDIA's efforts with Google and Tegra are somewhat more reassuring, as NVIDIA is definitely wanting to get its foot in the Android/ChromeOS door with Google. NVIDIA already has great consumer-ready hardware with Tegra on the Zune HD. This sparring match will only get more interesting as Intel launches their systems-on-a-chip, which have a gazillion different chipsets crammed into one. Intel is even wedging the PowerVR chipset in their designs.

I just need to remember that all this nifty hardware comes with a big caveat... without driver support, all of this functionality just sits idle between transistors.

I Heart NVIDIA

Awww... NVIDIA. I love you and your driver updates, especially now that they're coming once a month to *nix.

Retelling Yet Another Myth

Last month I put together my second MythTV box. I've become tired of my VIA Chrome box... video acceleration was a joke. So I thought that an onboard GeForce 7050PV would work since it had XvMC support. Plus NVIDIA's (binary-only) drivers were usually pretty good. Right?



I ordered Shuttle's SN68PTG5 AM2 case, figuring I could re-use an old AMD64 CPU I had and stuff the hard drive in from the previous MythTV box. When I got the case in it was a bit larger... and louder... than I expected. I needed to push my TV forward to get the box to fit behind it. The sheer amount of real estate I was able to work with made the opportunity worthwhile however - there was plenty of space around the motherboard, even with the ginormous heatsink installed.

Maneuvering around the box wasn't bad - I was able to put in the old drive and fill both memory channels with two Corsair XMS2 512M PC2 6400 sticks without a single flesh wound. Putting in the CPU was another issue however - I was short a pin.



Yeah... my old, to-be-recycled CPU was for Socket 939. This was a AM2 socket board - meaning it took 940-pin CPUs. Gah.

I paused for a few days and ordered a 2.1GHz AMD Athlon X2. It was a 65nm Brisbane core and only consumed 45W, so the lower thermals would serve a Myth box fairly well. Plus AMD chips are cheaper than avocados right now... it was an easy sell.



I wrangled the case together, put on some new thermal paste, and re-attached Shuttle's massive (did I mention its massive?) heat pipe for the CPU. Installed my old pcHDTV tuner card to receive HDTV channels from my local cable provider, then sealed everything up. I attached the barebones StreamZap Ir receiver via USB then did a openSUSE install with the MythTV repositories.

lircd actually worked without a hitch... the StreamZap didn't have any issues. Neither did the pcHDTV card - it worked out of the box as well. Things were going well, so I shoved the machine behind the TV, connected it via HDMI and went along my way.



At first things appeared to work well. Thanks to NVIDIA's nvidia-settings application I was able to effortlessly setup my 1080p LCD to work over HDMI. X configuration was TONS easier than with VIA's chipset - there it took me nearly 20-30 hours to get the monitor configuration correct for SVGA out. Of course it doesn't hurt that HDMI is just a DVI formfactor... but I digress.

No, the big pain was XvMC. With acceleration enabled, video would playback (either live or recorded) for anywhere between ten seconds to five minutes, then send X into a complete CPU spin. I'd suddenly have 100% usage on a single core and a complete lockup of X11. X had to be SIGKILL'd - repeatedly - before things would stop spinning out of control.

Luckily I bought a dual-core CPU so I could easily regain control. No matter now hard I tried, however, video acceleration just wouldn't do anything but hard lock MythTV.

Finally I gave up and just told MythTV to use the CPU for decoding. This worked acceptably, even with 720p streams. The CPU was beefy enough - it's a dual core, dual channel rig that has the bandwidth. Just a shame that neither NVIDIA nor VIA were able to provide a chipset that would allow accelerated MPEG2 playback in Linux. C'mon - is it really that bad?

Now things are recording and playing back just fine. The vsync is disabled right now - so I do have image tearing during high-motion decoding. But hopefully NVIDIA's next round of Linux drivers will fix XvMC, give me accelerated video and take all my worries away.

Re-Catching the Gaming Bug

I've re-caught the PC gaming bug. By and large I gave up gaming over a year and a half ago once I retired in Cheydinhal.

Wow... has it really been that long? Eighteen months since I've done any gaming? I've been reduced to occasional dorking around on my n810 (that lil' tablet has worked fantastically, btw), but mostly I've been either designing, developing or otherwise working every waking moment of my day. Man, that's sad.

I received Assassin's Creed as a very loving gift yesterday, so I fired up the good ole Wintendo and attempted an install. It had been so long since I'd turned my home desktop workstation on that I had nearly six months of Windows XP updates backlogged. One hour of driver/NVIDIA/anti-virus/Windows updates, another hour of volume defragmenting, three hours of nTune tweaking, two hours of BIOS tweaks, one hour installing the title and over ten reboots later I had a working install of Assassin's Creed. Whew.

I have an old Athlon 64 X2 and some cheap (but stable) Corsair RAM paired with a doable GeForce 7800 GT. Of course it wasn't quite enough to muscle through Assassin's Creed with all the frames I wanted, so I brought up the tRAS, tRCD and tRP from 8-3-3 to 6-3-3, cranked the frontside bus from 201 to 234 MHz, brought the PCI-E reference clock up to 117 MHz (to bring the bus from 2500 to 2925 MHz) and slightly poked the GPU clock up from 470 MHz to 475 (memory clock wouldn't budge). Surprisingly this actually got me over the fence.

It's kinda fun coming back to the tweak-and-tune days of PC gaming. NVIDIA's nTune app makes tweaking system values a huge deal easier, since it will talk directly to the mobo's NVIDIA chipset via (I'm guessing) ACPI. I can let nTune do its thing, rebooting once its locked up, and just hack merrily away on the laptop in the meantime.

Tomorrow it will be back to work again, but for this weekend I'm having fun. It's overdue.

And So It Begins - Vector Processing Standards Battle!

I'm not sure why I'm such a fan of this topic... maybe just because I enjoy watching the inevitable march towards entirely new CPU architectures.

Nvidia just released something that's been on Apple's wishlist for a while: OpenCL 1.0. Finally a "standard", royalty-free specification for developing applications to leverage vector processing units currently available on GPUs. While the processors on such high-end video cards aren't geared towards general computing per se, they absolutely blaze through certain workloads - especially those that work through sequential processing pipelines.

Microsoft's competing specification for some reason is available for DirectX 11 only, which makes absolutely no blimmin' sense to me. This basically means that your specification is limited to Vista... which rather defeats the concept behind a "standard." Not only do you get vendor lock-in, you get implementation lock-in. Sweet.

Can you imagine what this might do for Nvidia tho? Picture it now: tons of cheap commodity motherboards laid end-to-end, each with six PCI-E slots filled to the brim with Nvidia cards and running OpenCL apps on a stripped-down Linux distro. Supercomputing clusters for cheap.

Although I imagine the electricity bill might suffer.

Chipped Chrome

VIA recently announced that they have opened their reference documentation for their GPUs and are even now actively working with the openChrome project. For me, however, it's too little too late.

I've finally grown tired of even trying to get accelerated video working on my VIA-based MythTV box. XvMC support is simply non-existent and accelerated anything just doesn't work. With standard 480i broadcast TV I had no problem being CPU-bound for MPEG2 decoding, but it just doesn't fly with 720p and a pcHDTV card. I throw in the towel.

I'm looking at what Shuttle has to offer instead with either an Intel, NVIDIA or AMD platform. It appears that the graphic chipset choices break down into either GeForce 8, Intel GMA X4500HD, Intel GMA 3100 or GeForce 7050PV. The best NVIDIA choice appears to be the 7050PV, as it seems to enjoy known XvMC acceleration on MythTV's feature matrix and some have even reported getting it to work in a dual-head environment. The Intel cards should theoretically offer great performance for the power and enjoy good Linux driver support due to Intel's great contributions to X.Org, however Myth's wiki seems to know painfully little about the Intel GMA's. As far as XvMC support goes, Intel's chipsets don't seem to have a great track record.

So GeForce 7050 seems to be the most sane choice for those who are tired of fisticuffs with xorg.conf. But wait! We have audio to worry about too. If I'm going with a Shuttle GeForce 7050, then it looks like I'm going with a Realtek ALC888DD. Here again, MythTV notes that yet another Intel chipset is a true pain to work with. Still, I noticed that the MythTV hardware database notes there was at least one other Shuttle user with a similar setup that was able to get things to work.

Weirdly enough, both of the GeForce 7050 Shuttle boxes I found were AMD boxes. Go figure...

We Thank Thee, O Great NVIDIA...

All the way back in July I mentioned how slow and unstable KDE 4 is with NVIDIA, and why it's not their fault. Through August/September/October I've been living in a time warp and finally emerged out the other side... and behold! NVIDIA has graced our presence with a new driver release that has several KDE 4 compositing & Plasma fixes!!!

I clapped, I was so happy. Take pity on me.

KDE 4 Really Fast, Except When It's Really Slow

I've enjoyed openSUSE 11.0 not necessarily because it works flawlessly, but because it's the best working install of KDE 4. KDE 4 has a lot of great potential, but it's not fully realized until you hit KDE 4.1. Since 4.1 is the actual version of KDE 4 meant for widespread usage, I've been downloading the unstable builds from SuSE's build service.

Things have been working fine except for one extremely nagging thing - the initial draw of a konsole window takes 5-10 seconds. It's extremely obnoxious, especially when you need a terminal so often.

I found out this was due to NVIDIA's drivers on newer boards - as described in the KDE techbase and illustrated in the Ubuntu forums and How-To's. Luckily it was an easy fix - running nvidia-settings -a InitialPixmapPlacement=2 -a GlyphCache=1 as root set the appropriate settings in the NVIDIA driver to allow windows to resize at the speed they should. But I don't necessarily blame NVIDIA - the blame deserves to be cast further.

The Linux Hater's Blog brings up a great point about Xorg not being about to allocate offscreen buffers - something that I didn't realize. Xorg lacks a memory manager, so all the stuff you need for full OpenGL support just can't be done. It simply can't be done with Xorg. All the points made in the rant are absolutely right - the memory management infrastructure for pbuffer and framebuffer objects have to be there, otherwise you're hosed.

The core issue that comes from this deficiency is that X11 in and of itself is inherently unable to support OpenGL. Lack of offscreen buffers means that all the great stuff you should be able to do directly in hardware can only be accomplished with a software renderer. Of course, this defeats NVIDIA's entire business model of making the GPU the most important part of your workstation. So they had to massively replace parts of X11; the NVIDIA Linux drivers must, by sheer necessity, replace huge chunks of the XOrg implementation.

After reading the Linux Hater's post a lot of other stuff made sense - why NVIDIA's drivers are so invasive, why you magically don't need to install Xgl to run Compiz Fusion if you are using the proprietary nvidia driver (because it already replaced Xorg for you, thanks), and why KDE's desktop effects had window resizing slowdowns.

NVIDIA didn't break things - they fixed things. They're just trying to live in our broken world.

Intel Killing the VPU?

I had faintly heard grumblings of Intel loathing Nvidia, but I didn't really put 2 and 2 together until listening to Gordon Mah Ung on this week's No BS Podcast.

Good ole' Gordo spelled out why Nvidia ultimately acquired AGEIA - because Intel deep-sixed Havok's use of GPU's that had been in development for several years. While Havok was an independent company they worked with both ATI and Nvidia to support GPU processing of their physics API. Once Intel bought them the interoperability was tossed in the trash. I'm sure this was a pretty big dig at the GPU makers.

So Nvidia purchases the #2 player in the market to ensure this doesn't happen again. Let's see who enjoys the #1 spot in shipping titles during Christmas of '08.

Make a VPU Socket Already! Get It Over With!

My lands. If the floating point unit took this long to become mainstream I'd be using a Core 2 Duo with a math co-processor still.

Not to go all Halfhill or anything, but it has appeared that a vector co-processor or VPU's on-die were an immediate need for at least the past two or three years. Both Nvidia and AMD are bringing GPU's closer to the CPU, and it at once appeared that AMD's multi-core platform has included VPU/FPU/integer math/memory controller/CISC/RISC/misc./bacon&swiss together to take many types of tasks and integrate them under one die's roof.

And now that Nvidia has wisely acquired AGEIA and their PhysX platform it seems a general purpose vector processing platform is getting closer. A standalone PhysX never took off on the consumer marketplace, and purchasing another Radeon or GeForce just for physics processing (as both AMD and Nvidia were touting at electronics expo's) never caught on either. But a generalized, open-platform physics API that takes advantage of unused GPU cycles would definitely catch on. Spin your GPU fan faster and get real-time smoke effects... sign me up.

Nvidia has been extremely forward-thinking with their Linux drivers, and I hope they continue to be trend setters with the PhysX API. The PhysX engine and SDK was made freely available for Windows and Linux systems prior to Nvidia's acquisition, but hardware acceleration currently only works within Windows. Since Nvidia is porting PhysX to their own CUDA programming interface, it seems entirely probable that the Linux API would plugin to Nvidia's binary-only driver. And why not release the PhysX API under GPL? They could port to CUDA (whose specification is already open, available and widely used) then reduce their future development efforts by letting a wide swath of interested engineers maintain the codebase as needed.

Widely available drivers, development kits and APIs will help drive hardware sales in an era where Vista w/ DirectX 10 adoption isn't exactly stellar. I won't invest in being able to run Crysis in DX10 under native resolution for a 22" LCD, but I will invest to get more particle effects or more dynamic geoms. At that point you're adding to the whole gameplay proposition instead of polishing up aesthetics, with continuously diminishing results.

Nvidia's CPU

Now that ATI and AMD are one, Nvidia is working on a CPU as well. There are a lot of people saying they're working on a "GPU on a chip," but I seriously doubt it. It seems a lot of people are wagering this will go toward the embedded or low-power market, just like when VIA acquired Cyrix. But to do so would be extremely narrow-minded of ATI and Nvidia, even considering the boom of embedded devices and the increasing horsepower needed by set-top boxes.

I'm much more inclined to think that Nvidia is working with Intel (which would be a huge surprise, considering how they've been fighting with SLI on nForce vs. Intel chipsets) to compete with AMD & ATI working on bringing vector processing units to CPU's.

I incessantly bring this topic up, but I had to mention this since it seems that my predictions are actually going to happen. I don't necessarily think this will cause GPU's-on-CPU's to happen, although it might be a by-product. Instead I think this is going to allow for increased parallelization, faster MMX instructions (or 3DNow! if that's your taste) and a movement of putting the work of those physics accelerator cards back on the CPU die.

We may see a CPU with four cores, each with integer and floating point units then a handful of separate vector processors (a la IBM's cell) along side them. Given how the Cell reportedly absolutely sucks for many types of algorithms, this may give developers the best of both worlds. Rapidly branching and conditional logic can be done on the integer units with branch prediction and short instruction pipes while long, grinding algorithms can go to the vector processor. This has already worked for projects like Folding@Home, and could work for many similar algorithms.

Or AMD/ATI and Nvidia could just go for a stupid, embedded GPU's sitting on die with the CPU. But they'd be passing up something much cooler.

Exactly How Many PCI-E Slots Does One Need?

I'm pretty much hammering this subject into the ground, but then again so are component makers.

Recently we saw Ageia's physics accelerator come to market as an add-on expansion card, astride your existing video accelerator and sound card. Next we saw the Killer NIC reviewed by IGN, a "network interface accelerator" which promises to offload the assembly of TCP and UDP packets to take load off the CPU. It's actually an embedded Linux instance, assembling your TCP stack instead of Windows. Windows just sends the Killer NIC raw datagrams, then the Killer NIC does all the work to disassemble the datagrams into bare signals over the wire.

Now there's supposedly an accelerator in the works, the Intia Processor, dedicated to nonplayer character AI acceleration. It would offload AI processing from the CPU and instead have a dedicated API for pathfinding, terrain adaptation and line-of-site detection.

So let's count the possible gaming accelerators here:

Sound (i.e. Sound Blaser X-Fi)

Video (i.e. NVIDIA GeForce)

Video (additional NVIDIA GeForce for SLI)

Physics (Ageia PhysX)

Networking (Killer NIC)

AI (Intia Processor)

So that eats at least six slots, but if you have double-wide video cards, more like eight. Oh yeah, and you'll spend nearly $1,500 on just the above hardware accelerators alone.

This is a cycle that repeats itself in the computer world, tho. Popular software algorithms become API's. API's become baked into hardware. The hardware becomes to inflexible and becomes firmware. The firmware isn't flexible so it goes into software. And on and on we go.

But this hasn't necessarily happened for OpenGL or graphics acceleration. Why? Because there you're not baking an API into silicon, you're using a different type of processing unit altogether to solve a different genre of problems. Multiple vector processing pipelines can streamline linear algorithms much more effectively that complex instruction set CPU's. The same analogy can be made for floating point units when they were introduced along side elder CPU's that could only handle integer math.

Not only that, we're looking at chip makers shy away from raw clock speeds and instead looking at cramming as many CPU cores as possible on a single die. This allows for every CPU stamped out of the factory to be a multiprocessor machine. Once API's become more threadsafe, this crazy specialized hardware will instead just be an API call sent to one of the four idle CPU's on someone's desktop.

I think the eventual result of this crazy hardwareization (feel free to use that one) is that people are going to want to fit their algorithms into one of three processing units: the CPU pool (by "pool" I mean a pool of processors upon which to dump your asymmetric threads), the VPU (vector processing unit) or the FPU. The CPU & FPU have already become as one but as any C/C++ programmer can attest to, the decision to use integer based math as opposed to floating point math still rests heavily on the programmer's mind.

Given people are already compiling & running applications on a GPU that have nothing at all to do with graphics, it's only a matter of time before chipset manufacturers come up with a way to capitalize on the new general purpose processing unit.

GPU - The Next FPU?

What was the difference between a 468/SX and a 486/DX2? The additional floating point processor. Hellz yeah.

In the budding days of personal computers and 486 goodness, CPU's were strictly integer machines. Floating point math was simply too much for the little chicklets that didn't even need active cooling. I recall that back in the olden days of ISA cards and SIMS you could purchase separate FPU's that you could hammer into your brute-force socket... this granted you massive power that your Excel spreadsheets never previously dreamed of.

I recall how up-in-arms everyone was that Quake actually demanded a freakin' FPU. Who'd id think we were, Rockafeller?

Now that graphics cards have become cheap, plentiful and pretty freakin' powerful vector processing has become all the rage. Projects like BrookGPU and GPGPU have made it (relatively) easy to create userspace apps run on a GPU just as it would a CPU. This means your GeForce 7800 could work on certain serialized tasks, crunching numbers while it remained idle between SecondLife sessions.

Of course, GPU's are just good for vector processing, which means only a certain type of algorithms are really suited for it. You don't need to look much farther than critiques of IBM's Cell architecture to see how people feel about a rash of vector processing units versus a CPU that can span and branch effectively.

It was interesting to see that Nvidia and Havok have teamed up to offload physics to the GPU. Bear in mind this doesn't work with an actor's physics since it branches and can't be effectively streamlined or predicted, but does extremely well with cloth and particle physics.

Nvidia could be realizing that Ageia's physics accelerator is a more than vaporware, and is stepping up their offering to get into that product space. However, from the specs it appears that only Ageia's PhysX (*groan*) will be able to handle physics asynchronously. Although, I guess if you run physics calculations like shaders you could send multiple calculations down each pipe on the graphics card itself.

Should be interesting to see how the space pans out. The GPU is quite a hoss nowadays, and needs extremely fast access to memory, so it may not be incorporated as a side-by-side processor or on-die any time soon. But still... one has to wonder if multiple fast vector units are the next big thing to go on a proc.