Profile cover photo
Profile photo
Timothy Carstens
303 followers -
Climber, Haskell developer, systems-level bit twiddler.
Climber, Haskell developer, systems-level bit twiddler.

303 followers
About
Timothy's posts

Post has attachment

Getting ready to leave Houston and return to Seattle. I've had a great time down here this past week with two of my coworkers, hashing out designs and schedules. Next, I return to the office, exhausted but with Gantt charts in hand, hopefully to land the big contract with a stellar proposal.

I've spent the whole night in Microsoft Project. I am now, officially, dead inside.

Virtualization and high-end 3D performance.

So, the things I've learned so-far on this topic:
* VMware Workstation 8 limits itself to a max of 128MB of videoram. It will emulate more, but as far as I can tell, it will do its own texture swapping between system ram and your video ram. So, no high-end textures if you want good frame rates. I went so far as to find the location in the VMware binary that enforces this limitation, and patched the binary to remove the limit. Unsurprisingly, that didn't quite work, though rather surprisingly, it still booted my VMs just fine :D
* If you are on an Intel processor, and have VT-d, then Xen will let you dedicate a graphics adapter to a VM. I have heard that the same is possible in VMware, but VMware support is pretty awful, so I can't confirm. Also, Intel shipped me one of the defective LGA2011 processors that doesn't have VT-d, so I can't test until I finish my RMA process with them.

For now, I'm building an ever-growing IDA database of the vmware-vmx binary for Linux. I don't expect that I'll actually be able to patch vmware enough to get good graphics, but now I'm on the war path, so we'll see.

Post has attachment
Tonight: topological quantum computers. I'm reading a review article on the use of non-abelian anyons to perform quantum computation without suffering the damages of decoherence. It's a fascinating story that begins with a brief telling of quantum statistics and the perhaps-not-a-dichotomy-after-all between fermions and bosons, and moves from there to describe how the braid group associated to a collection of variously-bound particles could be used to perform computation. I don't know a great deal of quantum mechanics, but I've been able to enjoy enough bits and pieces to really get something out of this. Check it out: http://arxiv.org/abs/0707.1889

Post has attachment
APRS confirmed working.
Photo

Post has attachment
Photo
Photo
View album

Post has attachment
New addition: some weird PowerPC board. Once again, laser cut a mounting board.
Photo

Post has attachment
My talk with +Mikhail Davidov, on the work we did together with +rex warren at Leviathan, has been posted to youtube.

We did this work with funding from DARPA's CINDER initative. From an engineering perspective, this project has code written in C, C#, and Haskell, running on several machines and keeping distributed state. This is not the usual kind of work that I do, but it was a good deal of fun, and I'm sure I'll return to this problem area quite a bit.

Toorcon 13 (2011) - The schadenfreude of unreliable exploits: Going beyond Dr. Watson

Toorcon 13 has come to a close. Personal highlight: answering the "are you releasing code?" question with "if you want code, send a FOIA to Mudge. Also, it's largely written in Haskell." Epic lols. Shouts to my research partner, +Mikhail Davidov, and my project manager, +rex warren. A wonderful team!
Wait while more posts are being loaded