Muon1 hits 40 Quad milestone

The Muon1 project today hit its 40 quadrillion particle timestep milestone. That means that in the last 2.5 years, it's processed as much work as it had in the previous 9 years.

Muon1, a 2012 Retrospective

A little late, but I thought I’d bring you the breakdown of results for Muon1 through 2012. Better late than never though, so here are the highs and lows of 2012 for the Muon1 Distributed Particle Accelerator Project.

To start with, let’s start with the stats. The weekly stats continue to be published every Tuesday, and the start and end points this year for this piece are 3 January 2012 to 1 January 2013.

Best Designs - Linac900Ext7Xc2

After a break of several months, the latest entry in the video stable comes from Linac900Ext7Xc2. Started in February, and closed in October, it had the third highest number of simulations returned in recent times.

The best design was 4.142452% by 1 27972 - A Yoyo@home Boinc Wrapper user.

Muon1 Results Broken Down

As the Muon1 Stats server has had a breakdown, and hasn't been updated for two weeks, it's been a good chance to look at the stats themselves, and break them down by lattice.

With no stats box to constantly update it every hour, it's given me a stable dataset to work with for once. It does make a nice change to be able to see the stats and not have them change within an hour. So, while this break is ongoing, I thought about a breakdown of past lattices and their contributions.

Best Designs - Linac900Ext9X

So, after 10 months, and 44 entries, the 'Best Designs' series comes to a temporary close, with Linac900Ext9X. This lattice, the last to be 'finished' brings the series to an end, but will be revisited when more lattices close.

The best design was 2.506735% by TOMAZ

Best Designs - Linac900Ext10tc2

We come to the end of an era, with 10tc2. This lattice, released in July 2011, marked the last of the v4.44 compatable lattices, and the end of the line for the v4.44x clients that have been going since 2007.

The best design was 3.774999% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext6X

A new lattice, and another modification. Would the new client handle the RF phases better than the older client? Would the finer control of the phases make a difference? Only one way to find out.

The best design was 3.440846% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext10td2

The lattices are experimenting with the RF phase, and how that affects the yield and design. When 7tc2 took the worst of the d2 batch and inverted the phase, now we're taking the best, 10d2, and extending the max simulation time, giving 10td2.

The best design was 3.132038% by Team Norway's Steinrar

Best Designs - Linac900Ext6Xc2

The release of v4.45 meant we needed a new lattice. This lattice, a redo of 6tc2, only accepts 4.45 results, rather than 4.44d, in an aim to make it more accurate.

The best design was 3.858423% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext7tc2

Does phase really matter? We increased the yield when we extended the time, and re-centered the phase range with Linac900Ext6tc2, but is that sort of improvement possible on the worst of the d2 lattices? Let's see...

The best design was 3.730295% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext6tc2

The revisiting of a past simulation looks to improve on what was a fairly mediocre design, with a bit more time, and a 'twist'.

The best design was 4.387044% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Removable8c2

A brief return to the Removable series, gives us a quick interlude from the Ext series. Promising results from Linac900Removable6c2 have raised the possibility of taking things further.

The best design was 3.206615% by a Boinc Wrapper user (there was no way to tell who it wsa)

Best Designs - Linac900Ext10d2

The longest result of the series, with Linac900Ext10d2

The best design was 2.883241% by Team Russia's CTPAHHNK

Best Designs - Linac900Ext7d2

The lowest result of the series, with Linac900Ext7d2

The best design was 0.781853% by Team Atomic Milkshake at ArsTechnica's GOD

Best Designs - Linac900Ext8d2

Another Ext simulation gives another Boinc wrapper 'win'.

The best design was 2.670893% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext9d2

As with last weeks 6d2 lattice, we're working on a series of blocks, and phases. Now 9 blocks and the d2 RF phase setup.

The best design was 0.789847% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Ext6d2

Finally, it's time to step away from the "removable" generation of lattices, and into the "ext" generation. These are experimentation with longer linacs and the RF phase (as more fully detailed here)

The best design was 1.432695% by the Dutch Power Cows' Universal Creations~Darkbartje

New lattice files - 7Xc2 & 9Xc2

Those of you paying attention to your clients on Friday would have notices that they were downloading new lattices, and finishing simulations more quickly. That’s because new lattices have been introduced. Also, the v4.44d client has also been finally put to rest. There will be no more lattices that allow it to be used, so after almost a year, the jitter-free 4.45 will finally be the ‘required’ client.

With the retirement of both Linac900Ext10tc2 and Linac900Ext9X we’re moving on to Linac900Ext7Xc2 and Linac900Ext9Xc2. Both are development improvements of previous lattices, and attempt to work on phase handling, as well as investigate new designs without being side-tracked by jitter.

Best Designs - Linac900Removable3

The other alternative to adding complexity, is simplifying. The problem with such huge search-spaces, is that desirable targets are relatively small, and can be missed. By reducing complexity, we can make them more likely to be found.

The best design was 3.068246% by a Boinc Wrapper user (there is no way to tell who it is)

Best Designs - Linac900Removable12

The recent changes in searching for the designs with higher energies and increased yields have been productive, with yield percentages equal to, or above many previous designs, even before the higher energy levels are considered. Yet they've all been a bit simple in the design phase, so it's time we added another layer of complexity.

The best design was 3.670822% by a Boinc Wrapper user (there is no way to tell who it is)