Lockstep Programming Model¶

Section author: René Widera, Axel Huebl

The lockstep programming model structures code that is evaluated collectively and independently by workers (physical threads). Actual processing is described by one-dimensional index domains of virtual workers which can even be changed within a kernel. Mathematically, index domains are none-injective, total functions on physical workers.

An index domain is independent from data but can be mapped to a data domain, e.g. one to one or with more complex mappings.

Code which is implemented by the lockstep programming model is free of any dependencies between the number of worker and processed data elements. To simplify the implementation, each index within a domain can be seen as a virtual worker which is processing one data element (like the common workflow to programming CUDA). Each worker \(i\) can be executed as \(N_i\) virtual workers (\(1:N_i\)).

Functors passed into lockstep routines can have three different parameter signatures.

No parameter, if the work is not requiring the linear index within a domain: [&](){ }
An unsigned 32bit integral parameter if the work depends on indices within the domain range [0,domain size): [&](uint32_t const linearIdx){}
lockstep::Idx as parameter. lockstep::Idx is holing the linear index within the domain and meta information to access a context variables: [&](pmacc::mappings::threads::lockstep::Idx const idx){}

pmacc helpers¶

template<uint32_t T_domainSize, uint32_t T_numWorkers, uint32_t T_simdSize> struct Config¶

describe a constant index domain

describe the size of the index domain and the number of workers to operate on a lockstep domain

Template Parameters

T_domainSize: number of indices in the domain
T_numWorkers: number of worker working on T_domainSize
T_simdSize: SIMD width

Subclassed by pmacc::lockstep::ForEach< Config< T_domainSize, T_numWorkers, T_simdSize > >

struct Idx¶: Hold current index within a lockstep domain.

template<uint32_t T_numWorkers> class Worker¶

holds a worker configuration

collection of the compile time number of workers and the runtime worker index

Template Parameters

T_numWorkers: number of workers which are used to execute this functor

Subclassed by pmacc::lockstep::ForEach< Config< T_domainSize, T_numWorkers, T_simdSize > >

template<typename T_Type, typename T_Config> struct Variable : protected pmacc::memory::Array<T_Type, T_Config::maxIndicesPerWorker>, public T_Config ¶

Variable used by virtual worker.

This object is designed to hold context variables in lock step programming. A context variable is just a local variable of a virtual worker. Allocating and using a context variable allows to propagate virtual worker states over subsequent lock steps. A context variable for a set of virtual workers is owned by their (physical) worker.

Data stored in a context variable should only be used with a lockstep programming construct e.g. lockstep::ForEach<>

Warning

doxygenstruct: Cannot find class “pmacc::lockstep::ForEach” in doxygen xml output for project “PIConGPU” from directory: ../xml

Common Patterns¶

Create a Context Variable¶

A context variable is used to transfer information from a subsequent lockstep to another. You can use a context variable lockstep::Variable, similar to a temporary local variable in a function. A context variable must be defined outside of ForEach and should be accessed within the functor passed to ForEach only.

… and initialize with the index of the virtual worker

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
constexpr uint32_t frameSize = 256;
constexpr uint32_t numWorkers = 42;
auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx);
auto vIdx = forEachParticleSlotInFrame(
    [](lockstep::Idx const idx) -> int32_t
    {
        return idx;
    }
);

// is equal to

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
constexpr uint32_t frameSize = 256;
constexpr uint32_t numWorkers = 42;
auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx);
// variable will be uninitialized
auto vIdx = lockstep::makeVar<int32_t>(forEachParticleSlotInFrame);
forEachParticleSlotInFrame(
    [&](lockstep::Idx const idx)
    {
        vIdx[idx] = idx;
    }
);

To default initialize a context variable you can pass the arguments directly during the creation.

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
constexpr uint32_t frameSize = 256;
constexpr uint32_t numWorkers = 42;
auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx, 23);

Data from a context variable can be accessed within independent lock steps. A virtual worker has only access to there own context variable data.

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
constexpr uint32_t frameSize = 256;
constexpr uint32_t numWorkers = 42;
auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx);
auto vIdx = forEachParticleSlotInFrame(
    [](lockstep::Idx const idx) -> int32_t
    {
        return idx;
    }
);

// store old linear index into oldVIdx
auto oldVIdx = forEachExample(
    [&](lockstep::Idx const idx) -> int32_t
    {
        int32_t old = vIdx[idx];
        printf("virtual worker linear idx: %u == %u\n", vIdx[idx], idx);
        vIdx[idx] += 256;
        return old;
    }
);

forEachExample(
    [&](lockstep::Idx const idx)
    {
        printf("nothing changed: %u == %u - 256 == %u\n", oldVIdx[idx], vIdx[idx], idx);
    }
);

Collective Loop¶

each worker needs to pass a loop N times
in this example, there are more dates than workers that process them

// `frame` is a list which must be traversed collectively
while( frame.isValid() )
{
    // assume one dimensional indexing of threads within a block
    uint32_t const workerIdx = cupla::threadIdx(acc).x;
    constexpr uint32_t frameSize = 256;
    constexpr uint32_t numWorkers = 42;
    auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx);
    forEachParticleSlotInFrame(
       [&](lockstep::Idx const idx)
       {
           // independent work, idx can be used to access a context variable
       }
    forEachParticleSlotInFrame(
       [&](uint32_t const linearIdx)
       {
           // independent work based on the linear index only
       }
   );
}

Non-Collective Loop¶

each virtual worker increments a private variable

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
constexpr uint32_t frameSize = 256;
constexpr uint32_t numWorkers = 42;
auto forEachParticleSlotInFrame = lockstep::makeForEach<frameSize, numWorkers>(workerIdx);
auto vWorkerIdx = lockstep::makeVar<int32_t>(forEachParticleSlotInFrame, 0);
forEachParticleSlotInFrame(
    [&](auto const idx)
    {
        // assign the linear index to the virtual worker context variable
        vWorkerIdx[idx] = idx;
        for(int i = 0; i < 100; i++)
            vWorkerIdx[idx]++;
    }
);

Using a Master Worker¶

only one virtual worker (called master) of all available numWorkers manipulates a shared data structure for all others

// example: allocate shared memory (uninitialized)
PMACC_SMEM(
    finished,
    bool
);

// assume one dimensional indexing of threads within a block
uint32_t const workerIdx = cupla::threadIdx(acc).x;
auto onlyMaster = lockstep::makeMaster(workerIdx);

// manipulate shared memory
onlyMaster(
    [&]( )
    {
        finished = true;
    }
);

/* important: synchronize now, in case upcoming operations (with
 * other workers) access that manipulated shared memory section
 */
cupla::__syncthreads(acc);