search for: ferromagnetic

...ve; upon looking I see that they're using problem reductions to Ising models which aren't ferromagnet models (sign is different), and those problems are known to be NP-hard. Haven't thought about what kind of models LLVM is using. I hadn’t noticed that either. Anyway, LLVM’s models are ferromagnetic, the interactions are sums of expected basic block frequencies which are always non-negative numbers. Thanks, /jakob

On 7 August 2013 17:06, Jakob Stoklund Olesen <stoklund at 2pi.dk> wrote: > With the D-Wave computer in the news recently, you may find it interesting that LLVM’s register allocator is using Ising models to compute regions for live range splitting. > > The problem of finding a region for splitting a live range is mapped to an Ising model with the help of the edge bundle graph, see

With the D-Wave computer in the news recently, you may find it interesting that LLVM’s register allocator is using Ising models to compute regions for live range splitting. The problem of finding a region for splitting a live range is mapped to an Ising model with the help of the edge bundle graph, see EdgeBundles.h. A node in the edge bundle graph represents a set of CFG edges that enter or

...#39;t realized LLVM was using that and a quick skim of the file doesn't make me fully confident I understand what model it's implementing.) The cost function does map to the Hamiltonian of an Ising model (give or take a couple of constants), and all interactions are non-negative, making it ferromagnetic. There are a couple of cheats, though: - The nodes are tri-state, {-1, 0, +1}, with the neutral state being used as the starting point. This allows a small spin-up seed to grow into a domain of unbiased nodes until it hits negatively biased nodes. The nearby local minima found without this tweak t...

...e exceptions: Hard disc manufacturers use size tags in MByte, with 1 MByte = 1000000 Bytes... > Early on in the days of computer memory, it was realized that > memory is best organized in powers of 2. > I think that this dates back to the days when "core" memory > was little ferromagnetic cores (like a 0.1 mm donut) storing > one bit It's not that difficult and has nothing to do with history: to address a couple of memory bits, you have to select an address, and you will have to use a fixed set of address lines to transmit this address. Now, all these address lines carry dig...

On 7 August 2013 17:06, Jakob Stoklund Olesen <stoklund at 2pi.dk> wrote: > With the D-Wave computer in the news recently, you may find it interesting that LLVM’s register allocator is using Ising models to compute regions for live range splitting. > > The problem of finding a region for splitting a live range is mapped to an Ising model with the help of the edge bundle graph, see