Has Bakerlab examined the suitability of quantum computing as a (future) possible resource for Rosetta calculations?

As a layman, I have only a shallow comprehension of the types of problems that are best suited for (hypothetical) quantum processing. However, my impression is that many of the characteristics of how protein folding is calculated are the *kind* of characteristics (highly parallel, non-deterministic but converging on useful answers, modelling natural phenomena) that could benefit from quantum computing.

Has any consideration been given to this possibility? Is quantum computing on the roadmap at all?

---

Has Bakerlab examined the suitability of quantum computing as a (future) possible resource for Rosetta calculations?

Quantum computing?? I'd settle for gpu client...

---

Ask, and thou shalt receive.

Thanks slicedbread. You're the greatest thing since sl. . . nevermind.

---

I haven't found anywhere regarding "how fast" these D-Wave systems are compared to normal CPUs.

---

A modern CPU is faster compared to DWave 128 Qubit machine.

Ars did an article on it - they're considerably quicker than a CPU in some specific cases:

http://arstechnica.com/science/2013/05/d-waves-quantum-optimizer-pitted-against-traditional-computers/

---

Hello, this is Lucas from the Baker lab. Interesting thread, unfortunately quantum computing is not currently on our radar, but the work from Aspuru-Guzik's group certainly points to interesting possibilities. Two issues here:
1) GPU computing is much closer to reality for Rosetta, as it is completely routine for many other compute tasks. There is actually some preliminary work on this for Rosetta from a couple of developers as many tasks are amenable to GPU algorithms. It is just preliminary, proof-of-concept, and the fact is that as an academic group there is not a ton of man-power around to devote to the coding task of fully implementing this work. Could happen soon though.
2) The Aspuru-Guzik work focuses on a simplified model of protein folding that was popular in the 1990s, lattice proteins. These are a theoretical construct designed to study the process of protein folding, not for structure prediction or design. The big names here are Ken Dill, Jose Onuchic, Peter Wolynes, Eugene Shakhnovich -- lots of interesting work about the theoretical properties of proteins and their folding pathways. It is very beautiful work that related very closely to experimental findings in the 1990s from Alan Fersht's research group, so-called "phi-values" and protein folding transition states. But I digress...
Currently in the Baker lab we are not focusing on the folding problem so much as the structure prediction and design problems. Here lattice models are not useful, but instead we do (drumroll) off-lattice, rotamer-based models. The types of tools Aspuru-Guzik and co-workers developed for quantum computing could be adapted to off-lattice, so thank you for pointing that out. This will be on the radar eventually, but is currently we have a ton of algorithm work to figure out before we start on the quantum computing.

This answers my question nicely (and is probably of interest to those who have posted on other threads concerning GPU processing as well.) Thanks Lucas!

---

1) GPU computing is much closer to reality for Rosetta, as it is completely routine for many other compute tasks. There is actually some preliminary work on this for Rosetta from a couple of developers as many tasks are amenable to GPU algorithms. It is just preliminary, proof-of-concept, and the fact is that as an academic group there is not a ton of man-power around to devote to the coding task of fully implementing this work. Could happen soon though.

!!!!! Rosetta on gpu???

---

I like what I am seeing about possible/probable GPU computing for Rosetta! 'Soon' looks good!

---

The best thing is R@H need to announce open international initiative to let others develop GPU friendly Rosetta client so everyone benefits. For so many years R@H is left behind other GPU benefitting projects.

---

The best thing is R@H need to announce open international initiative to let others develop GPU friendly Rosetta client so everyone benefits.

From Rosetta Commons:
"RosettaCommons is the central hub for over 150 developers from 23 universities and laboratories to contribute and share the Rosetta source code. The RosettaCommons hosts a source code repository, this informational website, a support forum, and annual conferences to facilitate communication between developers."

---

Hi, I'm one of the rosetta devs Lucas mentioned above. Every time I post on forums I seem to get myself in trouble (probably because I do thinks like show people slides, which I'm doing again... I never learn...), so please be nice!

Rosetta for GPU in general is pretty tough for a few reasons, mainly that the it's tough to write large-general purpose GPU code -- and calculations we do in rosetta-land are incredibly diverse. A while back I'd figured out how to do a "refold," one of many basic underpinning of rosetta modeling, efficiently on the GPU. Here's a link to some slides (towards the end, I think) for the curious.

https://www.dropbox.com/s/6qkp6spzpz0cfs4/11_08_13_GPU.pdf

To date, we've had only one "application" of GPU calculations that I'm aware of, doing some entropic analysis of the effect of varying the size of floppy "linkers" attaching large protein components together. Here's some slides going over that:

https://www.dropbox.com/s/9gs20q4kaa0e8wp/12_04_05_GM_pub.pdf

I should note that the task in the above slides is very very simple and special-purpose... for that exact purpose the GPU turned out to work pretty well, but adapting the code to do more proved difficult enough that I'm not personally using it anymore. I've done similar tasks using CPU-based algorithms more recently and have gotten close to the GPU performance through various computational trickery which is impossible on current GPUs.

Thanks for all your support! The BOINC community has been pivotal to the work we do!

-Will

I should note that the task in the above slides is very very simple and special-purpose... for that exact purpose the GPU turned out to work pretty well, but adapting the code to do more proved difficult enough that I'm not personally using it anymore.

Thanks for the upgrades, but....
I cannot understand the "gpu-situation": are you working on very-specific-calculation-gpu-client for Rosetta (and release it -remote future- on gpu's volunteers) or you've completely abandoned this idea??

---

I should note that the task in the above slides is very very simple and special-purpose... for that exact purpose the GPU turned out to work pretty well, but adapting the code to do more proved difficult enough that I'm not personally using it anymore.

Thanks for the upgrades, but....
I cannot understand the "gpu-situation": are you working on very-specific-calculation-gpu-client for Rosetta (and release it -remote future- on gpu's volunteers) or you've completely abandoned this idea??

No answer??

---

boboviz to answer your question: GPU code for Rosetta is not being actively pursued in our lab. Will posted lots of stuff about his work for GPUs a few days ago, but as he says what he did ends up being very specialized and is not of general use. We are pursuing lots of other algorithm improvements, including Will's work, that use standard CPUs and we think that in many cases we are more limited by our ability to come up with correct and fast algorithms than we are by anything that could be solved with a GPU.
That being said, rosettaathome is an incredibly powerful tool for us to develop new algorithms with, and many of us are working on new algorithm development.

boboviz to answer your question: GPU code for Rosetta is not being actively pursued in our lab.

Thanks for the answer!!
I, despite all, continue to dream a future-gpu-client :-P

You are making a very good job with rosetta@home project!!!

---

We are pursuing lots of other algorithm improvements, including Will's work, that use standard CPUs and we think that in many cases we are more limited by our ability to come up with correct and fast algorithms than we are by anything that could be solved with a GPU.

Any roadmap for this improvements??

---

We are pursuing lots of other algorithm improvements, including Will's work, that use standard CPUs and we think that in many cases we are more limited by our ability to come up with correct and fast algorithms than we are by anything that could be solved with a GPU.

Any roadmap for this improvements??

I imagine it's very difficult to create a roadmap when you don't know whether something will be quicker without a lot of effort trying and testing it first. I would guess that if someone has success with porting some part of Rosetta to a GPU then we'll hear about it pretty quickly.

---