diff options
| -rw-r--r-- | CMakeLists.txt | 8 | ||||
| -rw-r--r-- | expt_0520.cu | 58 | ||||
| -rw-r--r-- | expt_0525.cu | 837 | ||||
| -rw-r--r-- | skiplistcustom.cuh | 31 | ||||
| -rw-r--r-- | sortlib.cuh | 1 |
5 files changed, 896 insertions, 39 deletions
diff --git a/CMakeLists.txt b/CMakeLists.txt index 118786c..4553b8c 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -222,6 +222,7 @@ set_target_properties(expt_0517 PROPERTIES set_target_properties(expt_0517 PROPERTIES CUDA_ARCHITECTURES "75") set_target_properties(expt_0517 PROPERTIES LINKER_LANGUAGE CUDA) + add_executable(expt_0520 expt_0520.cu skiplistcustom.cuh read_helper.h) target_link_libraries(expt_0520 m stdc++) @@ -230,6 +231,11 @@ set_target_properties(expt_0520 PROPERTIES set_target_properties(expt_0520 PROPERTIES CUDA_ARCHITECTURES "75") set_target_properties(expt_0520 PROPERTIES LINKER_LANGUAGE CUDA) +add_executable(expt_0525 expt_0525.cu read_helper.h) +target_link_libraries(expt_0525 m stdc++) - +set_target_properties(expt_0525 PROPERTIES + CUDA_SEPARABLE_COMPILATION ON) +set_target_properties(expt_0525 PROPERTIES CUDA_ARCHITECTURES "75") +set_target_properties(expt_0525 PROPERTIES LINKER_LANGUAGE CUDA) diff --git a/expt_0520.cu b/expt_0520.cu index 88815a7..d31dc78 100644 --- a/expt_0520.cu +++ b/expt_0520.cu @@ -111,18 +111,12 @@ constexpr size_t FACTOR = 1; // should change this to dynamic next time // constexpr size_t KEY_INDEX_SIZE = 32; -constexpr size_t SAMPLE_SIZE = 1024; +// constexpr size_t SAMPLE_SIZE = 1023; constexpr int block_size = STEP_SIZE; typedef LL key_type; -#ifdef RANDOM_TARGET -constexpr const char *TARGET_STRING = "RANDOM"; -#else -constexpr const char *TARGET_STRING = "PERFECT"; -#endif - #define CUDA_ERROR_CHECK #define CudaSafeCall(err) __cudaSafeCall(err, __FILE__, __LINE__) @@ -158,30 +152,38 @@ inline void __cudaCheckError(const char *file, const int line) { #endif } -__device__ key_type SampleStorage[SAMPLE_SIZE]; +//__device__ key_type SampleStorage[SAMPLE_SIZE]; +__device__ LockFreeSkipList *cudaGlobalSkipList; // Kernel for initializing device memory -__global__ void init(Node **n) { nodes = n; } +__global__ void init(LockFreeSkipList *list, Node **n) { + cudaGlobalSkipList = list; + nodes = n; +} + +__global__ void testFunctions() { cudaGlobalSkipList->testSample(); } // The main kernel -__global__ void kernel(LockFreeSkipList *skipList, const key_type *population, - size_t insertion_length) { +__global__ void kernel(const key_type *population, size_t insertion_length) { // The array items holds the sequence of keys // The array op holds the sequence of operations // The array result, at the end, will hold the outcome of the operations for (int i = 0; i < FACTOR; i++) { // FACTOR is the number of operations per thread - auto tid = - i * gridDim.x * blockDim.x + blockIdx.x * blockDim.x + threadIdx.x; + // auto tid = i * gridDim.x * blockDim.x + blockIdx.x * blockDim.x + + // threadIdx.x; + auto tid = FACTOR * (blockIdx.x * blockDim.x + threadIdx.x) + i; + // printf("%lu\n", tid); if (tid >= insertion_length) return; // Grab the operation and the associated key and execute key_type item = population[tid]; - skipList->Add(item); + // printf("%llu\n", population[tid]); + cudaGlobalSkipList->Add(item); } } @@ -207,20 +209,16 @@ int main(int argc, char **argv) { sample_length); ReadHelper reader("normal_distribution.txt", sample_length, insertion_length); - printf("%d\n", __LINE__); reader.readFile(total_row); - printf("%d\n", __LINE__); key_type *cudaPopulation; - printf("%d\n", __LINE__); cudaMalloc(&cudaPopulation, sizeof(key_type) * insertion_length); std::vector<key_type> _sample, _population; reader.split_into(_sample, _population); cudaMemcpy(cudaPopulation, _population.data(), sizeof(key_type) * insertion_length, cudaMemcpyHostToDevice); - printf("%d\n", __LINE__); // Allocate device memory // cudaMalloc((void **)&Clevels, sizeof(LL) * NUM_ITEMS); @@ -230,8 +228,6 @@ int main(int argc, char **argv) { (Node **)new LL[insertion_length]; // malloc(sizeof(LL) * adds); Node **cudaNodePointers; - printf("%d\n", __LINE__); - // Allocate the pool of free nodes for (int i = 0; i < insertion_length; i++) { @@ -242,14 +238,15 @@ int main(int argc, char **argv) { cudaMemcpyHostToDevice); CudaCheckError(); - printf("%d\n", __LINE__); // Allocate the skip list - LockFreeSkipList *Clist; - auto *list = new LockFreeSkipList(_sample.data(), SAMPLE_SIZE); + // LockFreeSkipList *cudaLockFreeSkipList; + auto *list = new LockFreeSkipList(_sample.data(), sample_length); - cudaMalloc((void **)&Clist, sizeof(LockFreeSkipList)); - cudaMemcpy(Clist, list, sizeof(LockFreeSkipList), cudaMemcpyHostToDevice); + LockFreeSkipList *cudaSkipList = nullptr; + cudaMalloc(&cudaSkipList, sizeof(LockFreeSkipList)); + cudaMemcpy(cudaSkipList, list, sizeof(LockFreeSkipList), + cudaMemcpyHostToDevice); CudaCheckError(); // Calculate the number of thread blocks // NUM_ITEMS = total number of operations to execute @@ -267,7 +264,10 @@ int main(int argc, char **argv) { } // Initialize the device memory - init<<<1, 32>>>(cudaNodePointers); + init<<<1, 32>>>(cudaSkipList, cudaNodePointers); + cudaDeviceSynchronize(); + + testFunctions<<<1, 1>>>(); cudaDeviceSynchronize(); // Launch main kernel @@ -277,7 +277,7 @@ int main(int argc, char **argv) { cudaEventCreate(&stop); cudaEventRecord(start, nullptr); - kernel<<<blocks, NUM_THREADS>>>(Clist, cudaPopulation, insertion_length); + kernel<<<blocks, NUM_THREADS>>>(cudaPopulation, insertion_length); CudaCheckError(); cudaDeviceSynchronize(); cudaEventRecord(stop, nullptr); @@ -289,8 +289,6 @@ int main(int argc, char **argv) { // Print kernel execution time in milliseconds - printf("%s %d ", TARGET_STRING, block_size); - printf("%lu: %lf", NUM_ITEMS, time); #if (defined(MEASURE_TIME) || defined(MEASURE_ACCESS)) @@ -322,7 +320,7 @@ int main(int argc, char **argv) { // printf("%d\n", element); #endif #endif - cudaFree(Clist); + // cudaFree(cudaLockFreeSkipList); for (int i = 0; i < insertion_length; i++) { cudaFree(pointers[i]); } diff --git a/expt_0525.cu b/expt_0525.cu new file mode 100644 index 0000000..af024bc --- /dev/null +++ b/expt_0525.cu @@ -0,0 +1,837 @@ +/* + +Copyright 2012-2013 Indian Institute of Technology Kanpur. All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + +1. Redistributions of source code must retain the above copyright notice, this +list of conditions, and the following disclaimer. +2. Redistributions in binary form must reproduce the above copyright notice, +this list of conditions, and the following disclaimer in the documentation +and/or other materials provided with the distribution. + +THIS SOFTWARE IS PROVIDED BY INDIAN INSTITUTE OF TECHNOLOGY KANPUR ``AS IS'' +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +DISCLAIMED. IN NO EVENT SHALL INDIAN INSTITUTE OF TECHNOLOGY KANPUR OR +THE CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT +OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. + +The views and conclusions contained in the software and documentation are +those of the authors and should not be interpreted as representing official +policies, either expressed or implied, of Indian Institute of Technology Kanpur. + +*/ + +/********************************************************************************** + + Lock-free skip list for CUDA; tested for CUDA 4.2 on 32-bit Ubuntu 10.10 and +64-bit Ubuntu 12.04. Developed at IIT Kanpur. + + Inputs: Percentage of add and delete operations (e.g., 30 50 for 30% add and +50% delete) Output: Prints the total time (in milliseconds) to execute the the +sequence of operations + + Compilation flags: -O3 -arch sm_20 -I ~/NVIDIA_GPU_Computing_SDK/C/common/inc/ +-DNUM_ITEMS=num_ops -DFACTOR=num_ops_per_thread -DKEYS=num_keys + + NUM_ITEMS is the total number of operations (mix of add, delete, search) to +execute. + + FACTOR is the number of operations per thread. + + KEYS is the number of integer keys assumed in the range [10, 9+KEYS]. + The paper cited below states that the key range is [0, KEYS-1]. However, we +have shifted the range by +10 so that the head sentinel key (the minimum key) +can be chosen as zero. Any positive shift other than +10 would also work. + + The include path ~/NVIDIA_GPU_Computing_SDK/C/common/inc/ is needed for +cutil.h. + + Related work: + + Prabhakar Misra and Mainak Chaudhuri. Performance Evaluation of Concurrent +Lock-free Data Structures on GPUs. In Proceedings of the 18th IEEE International +Conference on Parallel and Distributed Systems, December 2012. + +***************************************************************************************/ + +// #include"cutil.h" // Comment this if cutil.h is not available +// #include "cuda_runtime.h" +#include "read_helper.h" +#include "sortlib.cuh" +#include <algorithm> +#include <cassert> +#include <cstdio> +#include <cstdlib> +#include <random> +#include <set> + +#if __WORDSIZE == 64 +typedef unsigned long long LL; +#else +typedef unsigned int LL; +#endif + +#ifndef BUILD_SIZE +#define BUILD_SIZE 1048576 +#endif + +#ifndef STEP_SIZE +#define STEP_SIZE 2 +#endif + +// #define MEASURE_TIME +// #define MEASURE_ACCESS + +#if (defined(MEASURE_ACCESS) && defined(MEASURE_TIME)) +#error "Shouldn't define MEASURE_TIME and MEASURE_ACCESS at the same time" +#endif + +#ifdef MEASURE_TIME +#undef BUILD_SIZE +#define BUILD_SIZE 1024 +#endif + +// Maximum level of a node in the skip list +// #define MAX_LEVEL 32 +constexpr size_t MAX_LEVEL = 16; + +// Number of threads per block +// #define NUM_THREADS 512 +constexpr size_t NUM_THREADS = 512; + +constexpr size_t NUM_ITEMS = BUILD_SIZE; +// constexpr size_t KEYS = 1048576; +constexpr size_t FACTOR = 1; + +// should change this to dynamic next time +// constexpr size_t KEY_INDEX_SIZE = 32; +// constexpr size_t SAMPLE_SIZE = 1023; + +// constexpr int block_size = STEP_SIZE; + +// Supported operations +/*constexpr int ADD = 0; +constexpr int DELETE = 1; +constexpr int SEARCH = 2;*/ + +typedef LL key_type; + +#ifdef RANDOM_TARGET +constexpr const char *TARGET_STRING = "RANDOM"; +#else +// constexpr const char *TARGET_STRING = "PERFECT"; +#endif + +#define CUDA_ERROR_CHECK + +#define CudaSafeCall(err) __cudaSafeCall(err, __FILE__, __LINE__) +#define CudaCheckError() __cudaCheckError(__FILE__, __LINE__) + +inline void cudaSafeCall_(cudaError err, const char *file, const int line) { +#ifdef CUDA_ERROR_CHECK + if (cudaSuccess != err) { + fprintf(stderr, "cudaSafeCall() failed at %s:%i : %s\n", file, line, + cudaGetErrorString(err)); + exit(-1); + } +#endif +} + +inline void __cudaCheckError(const char *file, const int line) { +#ifdef CUDA_ERROR_CHECK + cudaError err = cudaGetLastError(); + if (cudaSuccess != err) { + fprintf(stderr, "cudaCheckError() failed at %s:%i : %s\n", file, line, + cudaGetErrorString(err)); + exit(-1); + } + + // More careful checking. However, this will affect performance. + // Comment away if needed. + err = cudaDeviceSynchronize(); + if (cudaSuccess != err) { + fprintf(stderr, "cudaCheckError() with sync failed at %s:%i : %s\n", file, + line, cudaGetErrorString(err)); + exit(-1); + } +#endif +} + +class Node; + +// Definition of generic node class + +class __attribute__((aligned(16))) Node { +public: + int topLevel; // Level of the node + LL key; // Key value + LL next[MAX_LEVEL + 1]{}; // Array of next links + + // Create a next field from a reference and mark bit + static __device__ __host__ LL CreateRef(Node *ref, bool mark) { + LL val = (LL)ref; + val = val | mark; + return val; + } + + __device__ __host__ void SetRef(int index, Node *ref, bool mark) { + next[index] = CreateRef(ref, mark); + } + + // Extract the reference from a next field + __device__ Node *GetReference(int index) { + LL ref = next[index]; + return (Node *)((ref >> 1) << 1); + } + + // Extract the reference and mark bit from a next field + __device__ Node *Get(int index, bool *marked) { + marked[0] = next[index] % 2; + return (Node *)((next[index] >> 1) << 1); + } + + // CompareAndSet wrapper + __device__ bool CompareAndSet(int index, Node *expectedRef, Node *newRef, + bool oldMark, bool newMark) { + LL oldVal = (LL)expectedRef | oldMark; + LL newVal = (LL)newRef | newMark; + LL *ref = &(next[index]); + LL oldValOut = atomicCAS(ref, oldVal, newVal); + if (oldValOut == oldVal) + return true; + return false; + } + + // Constructor for sentinel nodes + explicit Node(LL k) { + key = k; + topLevel = MAX_LEVEL; + int i; + for (i = 0; i < MAX_LEVEL + 1; i++) { + next[i] = CreateRef((Node *)nullptr, false); + } + } +}; + +/*struct MapNode { + LL key; + // Node *point[MAX_LEVEL + 1]; + Node *point; +}; + +class MemMap { +public: + size_t size; + size_t real_size; + MapNode *store; + + MemMap() : size(0), store(nullptr), real_size(0) {} + + __device__ bool insert(MapNode node) { + bool need_extend = this->size + 1 > this->real_size; + if (need_extend) { + bool need_copy = this->real_size == 0; + if (!need_copy) { + this->real_size += 1; + } + this->real_size *= 2; + MapNode *old = this->store; + this->store = new MapNode[this->real_size]; + if (need_copy) { + memcpy(this->store, old, this->real_size * sizeof(MapNode *)); + } + delete[] old; + } + // need sort after insert + this->store[size] = node; + this->size += 1; + } + + __device__ Node *search(LL key) { + for (int offset = 0; offset < this->size; offset++) { + if (this->store[offset].key >= key) { + return this->store[offset].point; + } + } + return nullptr; + } + + __device__ ~MemMap() { delete[] store; } +};*/ + +// Definition of lock-free skip list + +class LockFreeSkipList { + + key_type *sample; + size_t sampleLength; + + CustomSort customSort; + +public: + Node *head; + Node *tail; + LockFreeSkipList(key_type *_sample, size_t sample_length) + : sampleLength(sample_length), + customSort(sample_length, sizeof(key_type) * 8) { + Node *h = new Node(0); + // size_ = 0; +#if __WORDSIZE == 64 + Node *t = new Node(std::numeric_limits<key_type>::max() - 1); +#else + Node *t = new Node((LL)0xffffffff); +#endif + cudaMalloc(&head, sizeof(Node)); + + cudaMalloc(&tail, sizeof(Node)); + for (auto i = 0; i < h->topLevel + 1; i++) { + h->SetRef(i, tail, false); + } + cudaMemcpy(head, h, sizeof(Node), cudaMemcpyHostToDevice); + + cudaMemcpy(tail, t, sizeof(Node), cudaMemcpyHostToDevice); + + cudaMalloc(&this->sample, sizeof(key_type) * sampleLength); + cudaMemcpy(this->sample, _sample, sizeof(key_type) * sampleLength, + cudaMemcpyHostToDevice); + } + __device__ bool find(LL, Node **, Node **); // Helping method + __device__ bool Add(LL); + __device__ bool Delete(LL); + __device__ bool Search(LL); + + __device__ size_t searchIndex(key_type key) { + auto result = customSort.binary_search(this->sample, key); + auto index = customSort.calculate_index(result - this->sample); + return index; + } + + __device__ unsigned trailing_zeroes(size_t index) { + constexpr auto block_size = 2; + unsigned bits = 0; + LL x = index / block_size; + + if (x) { + while (x % block_size == 0) { + ++bits; + x /= block_size; + } + } + return bits; + } + +#ifdef MEASURE_ACCESS + unsigned access_times = 0; + + __device__ unsigned getAccessCount() const { return this->access_times; } + __device__ void increaseAccessCount(unsigned count = 1) { + atomicAdd(&this->access_times, count); + } +#else + __device__ void increaseAccessCount(unsigned _count = 1) {} +#endif + +#ifdef MEASURE_TIME + unsigned round = 0; + __device__ void increaseRoundCount(unsigned count = 1) { + atomicAdd(&this->round, count); + } + int spend_time[NUM_ITEMS]{0}; + unsigned long long total_time = 0; + __device__ unsigned getRoundCount() const { return this->round; } +#endif +}; + +__device__ Node **nodes; // Pool of pre-allocated nodes +__device__ unsigned int pointerIndex = 0; // Index into pool of free nodes +__device__ LL + *randoms; // Array storing the levels of the nodes in the free pool + +// Function for creating a new node when requested by an add operation + +__device__ Node *GetNewNode(LL key, size_t topLevel) { + LL ind = atomicInc(&pointerIndex, NUM_ITEMS); + Node *n = nodes[ind]; + n->key = key; + // n->topLevel = randoms[ind]; + n->topLevel = topLevel; + int i; + for (i = 0; i < n->topLevel + 1; i++) { + n->SetRef(i, nullptr, false); + } + return n; +} + +__device__ LockFreeSkipList *lockFreeSkipList; // The lock-free skip list + +//__device__ LL KeyIndex[KEY_INDEX_SIZE]; + +//__device__ key_type SampleStorage[SAMPLE_SIZE]; + +// Kernel for initializing device memory + +__global__ void init(LockFreeSkipList *l1, Node **n, LL *rands) { + randoms = rands; + nodes = n; + lockFreeSkipList = l1; +} + +// Find the window holding key +// On the way clean up logically deleted nodes (those with set marked bit) + +__device__ bool +LockFreeSkipList::find(LL key, Node **preds, + Node **succs) { // preds and succs are arrays of pointers + int bottomLevel = 0; + bool marked[] = {false}; + bool snip; + Node *pred; + Node *curr = nullptr; + Node *succ; + bool beenThereDoneThat; + while (true) { + beenThereDoneThat = false; + pred = head; + int level; + for (level = MAX_LEVEL; level >= bottomLevel; level--) { + curr = pred->GetReference(level); + while (true) { + succ = curr->Get(level, marked); + while (marked[0]) { + snip = pred->CompareAndSet(level, curr, succ, false, false); + beenThereDoneThat = true; + if (!snip) + break; + curr = pred->GetReference(level); + succ = curr->Get(level, marked); + beenThereDoneThat = false; + // printf("find key is %d \n",(int)key); + } + if (beenThereDoneThat) + break; + if (curr->key <= key) { + pred = curr; + curr = succ; + } else { + break; + } + } + if (beenThereDoneThat) + break; + preds[level] = pred; + succs[level] = curr; + } + if (beenThereDoneThat) + continue; + return ((curr->key == key)); + } +} + +__device__ bool LockFreeSkipList::Search(LL key) { + int bottomLevel = 0; + bool marked = false; + Node *pred = head; + Node *curr = nullptr; + Node *succ; + int level; + for (level = MAX_LEVEL; level >= bottomLevel; level--) { + curr = pred->GetReference(level); +#ifdef MEASURE_ACCESS + this->increaseAccessCount(); +#endif + while (true) { + succ = curr->Get(level, &marked); +#ifdef MEASURE_ACCESS + this->increaseAccessCount(); +#endif + while (marked) { + curr = curr->GetReference(level); + succ = curr->Get(level, &marked); +#ifdef MEASURE_ACCESS + this->increaseAccessCount(2); +#endif + } + if (curr->key < key) { + pred = curr; + curr = succ; + } else { + break; + } + } + } + return (curr != nullptr && curr->key == key); +} + +__device__ bool LockFreeSkipList::Delete(LL key) { + int bottomLevel = 0; + Node *preds[MAX_LEVEL + 1]; + Node *succs[MAX_LEVEL + 1]; + Node *succ; + bool marked[] = {false}; + while (true) { + bool found = find(key, preds, succs); + if (!found) { + return false; + } else { + Node *nodeToDelete = succs[bottomLevel]; + int level; + for (level = nodeToDelete->topLevel; level >= bottomLevel + 1; level--) { + succ = nodeToDelete->Get(level, marked); + while (!marked[0]) { + nodeToDelete->CompareAndSet(level, succ, succ, false, true); + succ = nodeToDelete->Get(level, marked); + } + } + succ = nodeToDelete->Get(bottomLevel, marked); + while (true) { + bool iMarkedIt = + nodeToDelete->CompareAndSet(bottomLevel, succ, succ, false, true); + succ = succs[bottomLevel]->Get(bottomLevel, marked); + if (iMarkedIt) { + find(key, preds, succs); + // size_ -= 1; + // atomicDec(&size_, 1); + return true; + } else if (marked[0]) { + return false; + } + } + } + } +} + +__device__ bool LockFreeSkipList::Add(LL key) { + Node *newNode = + GetNewNode(key, this->trailing_zeroes(this->searchIndex(key))); + int topLevel = newNode->topLevel; + int bottomLevel = 0; + Node *preds[MAX_LEVEL + 1]; + Node *succs[MAX_LEVEL + 1]; + int level; + while (true) { + bool found = find(key, preds, succs); + if (found) { + return false; + } else { + Node *pred; + Node *succ; + for (level = bottomLevel; level <= topLevel; level++) { + succ = succs[level]; + newNode->SetRef(level, succ, false); + } + pred = preds[bottomLevel]; + succ = succs[bottomLevel]; + bool t; + // printf("--- key is %d pred is %d succ is %d level is %d + // \n",(int)key,(int)pred->key,(int)succ->key,0); + t = pred->CompareAndSet(bottomLevel, succ, newNode, false, false); + if (!t) { + continue; + } + for (level = bottomLevel + 1; level <= topLevel; level++) { + + while (true) { + pred = preds[level]; + succ = succs[level]; + newNode->SetRef(level, succ, false); + // printf("-- key is %d pred is %d succ is %d level is %d + // \n",(int)key,(int)pred->key,(int)succ->key,(int)level); + if (pred->CompareAndSet(level, succ, newNode, false, false)) { + break; + } + // printf("key is %d pred is %d succ is %d level is %d + // \n",(int)key,(int)pred->key,(int)succ->key,(int)level); + find(key, preds, succs); + } + } + // size_ += 1; + // this->key_map.insert(MapNode(ll, newNode)); + // atomicAdd(&size_, 1); + return true; + } + } +} + +__global__ void print() { + // For debugging + int tid = blockIdx.x * blockDim.x + threadIdx.x; + if (tid == 0) { + Node *p = lockFreeSkipList->head; + bool marked = false; + while (p != nullptr) { +#if __WORDSIZE == 64 + printf("%#llx, %u, marked=%u, address is %p : ", p->key, p->topLevel, + marked, p); +#else + printf("%#x, %u, marked=%u, address is %p\n", p->key, p->topLevel, marked, + p); +#endif + for (int i = 0; i < p->topLevel + 1; i++) { + printf(" %d ", (int)(p->GetReference(i)->key)); + } + printf("\n"); + p = p->Get(0, &marked); + } + printf("\n"); + } +} + +// The main kernel + +__global__ void kernel(const LL *items, size_t search_length, LL *result) { + // The array items holds the sequence of keys + // The array op holds the sequence of operations + // The array result, at the end, will hold the outcome of the operations + + for (int i = 0; i < FACTOR; + i++) { // FACTOR is the number of operations per thread + auto tid = + i * gridDim.x * blockDim.x + blockIdx.x * blockDim.x + threadIdx.x; + if (tid >= search_length) + return; + + // Grab the operation and the associated key and execute + LL item = items[tid]; +#ifdef MEASURE_TIME + unsigned long long start_time = clock64(); +#endif + result[tid] = lockFreeSkipList->Search(item); +#ifdef MEASURE_TIME + unsigned long long end_time = clock64() - start_time; + if (lockFreeSkipList->spend_time[tid]) { + printf("conflict: %d\n", tid); + } + lockFreeSkipList->spend_time[tid] = (int)end_time; +#endif + } +} + +__global__ void kernelAdd(LL *item, size_t insertion_length) { + + for (int i = 0; i < FACTOR; + i++) { // FACTOR is the number of operations per thread + auto tid = + i * gridDim.x * blockDim.x + blockIdx.x * blockDim.x + threadIdx.x; + + if (tid >= insertion_length) + return; + lockFreeSkipList->Add(item[tid]); + } +} + +/*LL Randomlevel() { + LL v = 1; + double p = 0.5; + while (((rand() / (double)(RAND_MAX)) < p) && (v < MAX_LEVEL)) + v++; + return v; +}*/ + +// Generate the level of a newly created node + +__global__ void print_function() { +#ifdef MEASURE_ACCESS + printf("count: %u\n", l->getAccessCount()); +#endif +} + +/*__global__ void copy_function(int *spend_time) { + memcpy(spend_time, lockFreeSkipList->spend_time, sizeof(int) * NUM_ITEMS); +}*/ + +inline size_t calcBlocks(size_t input) { + return (input % (NUM_THREADS * FACTOR) == 0) + ? input / (NUM_THREADS * FACTOR) + : (input / (NUM_THREADS * FACTOR)) + 1; +} + +int main(int argc, char **argv) { + if (argc != 4) { + printf("Need two arguments: percent add ops and percent delete ops (e.g., " + "30 50 for 30%% add and 50%% delete).\nAborting...\n"); + exit(1); + } + + auto sample_length = strtol(argv[2], nullptr, 10); + auto insertion_length = strtol(argv[2], nullptr, 10); + auto search_length = strtol(argv[2], nullptr, 10); + auto total_row = 0UL; + + ReadHelper readHelper("normal_distribution.txt", sample_length, + insertion_length + search_length); + readHelper.readFile(total_row); + std::vector<key_type> _sample, _population; + readHelper.split_into(_sample, _population); + printf("Create search vector: %ld\n", + _population.end() - _population.begin() + insertion_length); + std::vector<key_type> _search(_population.begin() + insertion_length, + _population.end()); + _population.resize(insertion_length); + + // Allocate necessary arrays + // LL *op = new LL[NUM_ITEMS]; //(LL *)malloc(sizeof(LL) * NUM_ITEMS); + // LL *levels = new LL[NUM_ITEMS]; //(LL *)malloc(sizeof(LL) * NUM_ITEMS); + // LL *items = new LL[NUM_ITEMS]; //(LL *)malloc(sizeof(LL) * NUM_ITEMS); + LL *result = new LL[NUM_ITEMS]; //(LL *)malloc(sizeof(LL) * NUM_ITEMS); + + // Allocate device memory + + LL *cudaOperatorItems; + // LL *Cop; + LL *Cresult; + LL *Clevels; + + cudaMalloc(&Cresult, sizeof(LL) * NUM_ITEMS); + cudaMalloc(&cudaOperatorItems, sizeof(LL) * insertion_length); + // cudaMalloc(&Cop, sizeof(LL) * NUM_ITEMS); + cudaMalloc(&Clevels, sizeof(LL) * NUM_ITEMS); + // cudaMemcpy(Clevels, levels, sizeof(LL) * NUM_ITEMS, + // cudaMemcpyHostToDevice); + cudaMemcpy(cudaOperatorItems, _population.data(), + sizeof(key_type) * insertion_length, cudaMemcpyHostToDevice); + // cudaMemcpy(Cop, op, sizeof(LL) * NUM_ITEMS, cudaMemcpyHostToDevice); + Node **pointers = + (Node **)new LL[insertion_length]; // malloc(sizeof(LL) * adds); + Node **Cpointers; + + // Allocate the pool of free nodes + + for (int i = 0; i < insertion_length; i++) { + cudaMalloc(&pointers[i], sizeof(Node)); + } + cudaMalloc(&Cpointers, sizeof(Node *) * insertion_length); + cudaMemcpy(Cpointers, pointers, sizeof(Node *) * insertion_length, + cudaMemcpyHostToDevice); + + // Allocate the skip list + + LockFreeSkipList *Clist; + auto *list = new LockFreeSkipList(_sample.data(), _sample.size()); + + cudaMalloc(&Clist, sizeof(LockFreeSkipList)); + cudaMemcpy(Clist, list, sizeof(LockFreeSkipList), cudaMemcpyHostToDevice); + // Calculate the number of thread blocks + // NUM_ITEMS = total number of operations to execute + // NUM_THREADS = number of threads per block + // FACTOR = number of operations per thread + + size_t blocks = calcBlocks(insertion_length); + + CudaCheckError(); + + // Initialize the device memory + init<<<1, 32>>>(Clist, Cpointers, Clevels); + cudaDeviceSynchronize(); + + // Insertion to skiplist + kernelAdd<<<blocks, NUM_THREADS>>>(cudaOperatorItems, insertion_length); + cudaDeviceSynchronize(); + + // Re-allocate memory for search + cudaFree(cudaOperatorItems); + cudaMalloc(&cudaOperatorItems, sizeof(key_type) * search_length); + cudaMemcpy(cudaOperatorItems, _search.data(), + sizeof(key_type) * search_length, cudaMemcpyHostToDevice); + + // Launch main kernel + + cudaEvent_t start, stop; + cudaEventCreate(&start); + cudaEventCreate(&stop); + cudaEventRecord(start, nullptr); + + blocks = calcBlocks(search_length); + kernel<<<blocks, NUM_THREADS>>>(cudaOperatorItems, search_length, Cresult); + CudaCheckError(); + cudaDeviceSynchronize(); + cudaEventRecord(stop, nullptr); + cudaEventSynchronize(stop); + float time; + cudaEventElapsedTime(&time, start, stop); + cudaEventDestroy(start); + cudaEventDestroy(stop); + + // Print kernel execution time in milliseconds + + printf("%lu: %lf", NUM_ITEMS, time); + + // LL *Cop2; + // cudaMalloc((void **)&Cop2, sizeof(LL) * NUM_ITEMS); + // cudaMemcpy(Cop2, op, sizeof(LL) * NUM_ITEMS, cudaMemcpyHostToDevice); + + cudaEventCreate(&start); + cudaEventCreate(&stop); + cudaEventRecord(start, nullptr); + kernel<<<blocks, NUM_THREADS>>>(cudaOperatorItems, search_length, Cresult); + CudaCheckError(); + cudaDeviceSynchronize(); + cudaEventRecord(stop, nullptr); + cudaEventSynchronize(stop); + cudaEventElapsedTime(&time, start, stop); + cudaEventDestroy(start); + cudaEventDestroy(stop); + + // Print kernel execution time in milliseconds + + printf(" %lf\n", time); + + // Check for errors + + // Move results back to host memory + + cudaMemcpy(result, Cresult, sizeof(LL) * NUM_ITEMS, cudaMemcpyDeviceToHost); + + // Uncomment the following for debugging + // print<<<1,32>>>(); + cudaDeviceSynchronize(); + +#if (defined(MEASURE_TIME) || defined(MEASURE_ACCESS)) + + print_function<<<1, 1>>>(); + + cudaDeviceSynchronize(); +#ifdef MEASURE_TIME + { + int *cuda_tmp = nullptr; + cudaMalloc(&cuda_tmp, sizeof(int) * NUM_ITEMS); + copy_function<<<1, 1>>>(cuda_tmp); + cudaDeviceSynchronize(); + CudaCheckError(); + FILE *file = fopen("spend_time.txt", "w"); + int *tmp = new int[NUM_ITEMS]; + cudaMemcpy(tmp, cuda_tmp, sizeof(int) * NUM_ITEMS, cudaMemcpyDeviceToHost); + // memcpy(tmp, SpendTime, sizeof(int) * NUM_ITEMS); + for (int i = 0; i < NUM_ITEMS; i++) { + if (tmp[i] == 0) + break; + fprintf(file, "%d\n", tmp[i]); + } + // printf("%d\n", i); + delete[] tmp; + fclose(file); + } + // for (auto element : SpendTimeVec) + // printf("%d\n", element); +#endif +#endif + /*cudaFree(Clist); + cudaFree(Cop2); + cudaFree(Clevels); + cudaFree(Cop); + cudaFree(Citems); + cudaFree(Cresult); + free(pointers); + delete [] op; + delete [] levels; + delete [] items; + delete [] result;*/ + return 0; +} diff --git a/skiplistcustom.cuh b/skiplistcustom.cuh index a16e80a..80ecc19 100644 --- a/skiplistcustom.cuh +++ b/skiplistcustom.cuh @@ -1,5 +1,6 @@ #pragma once #include "sortlib.cuh" +#include <limits> #ifndef LOCKFREE_SKIPLIST_CUH_ #define LOCKFREE_SKIPLIST_CUH_ @@ -85,7 +86,7 @@ public: // Definition of lock-free skip list class LockFreeSkipList { - static constexpr size_t SAMPLE_LENGTH = 1024; + // static constexpr size_t SAMPLE_LENGTH = 1023; public: Node *head; @@ -99,20 +100,21 @@ public: double slice_size; + size_t sampleLength; + LockFreeSkipList(key_type *samples, size_t sample_length) - : searcher(sample_length, sizeof(key_type)) { + : searcher(sample_length, sizeof(key_type)), sampleLength(sample_length) { Node *h = new Node(0); // size_ = 0; #if __WORDSIZE == 64 - Node *t = new Node((key_type)NUM_ITEMS + 10); + Node *t = new Node(std::numeric_limits<key_type>::max()); #else Node *t = new Node(0xffffffffULL); #endif cudaMalloc(&head, sizeof(Node)); cudaMalloc(&tail, sizeof(Node)); - int i; - for (i = 0; i < h->topLevel + 1; i++) { + for (int i = 0; i < h->topLevel + 1; i++) { h->SetRef(i, tail, false); } cudaMemcpy(head, h, sizeof(Node), cudaMemcpyHostToDevice); @@ -123,8 +125,10 @@ public: initDeviceVariable(); this->slice_size = 1.0 / sample_length; + // printf("%lf\n", this->slice_size); - cudaMalloc(&samples, sizeof(key_type) * sample_length); + printf("%ld\n", sample_length); + cudaMalloc(&this->samples, sizeof(key_type) * sample_length); cudaMemcpy(this->samples, samples, sizeof(key_type) * sample_length, cudaMemcpyHostToDevice); } @@ -150,9 +154,11 @@ public: return bits; } - __device__ size_t calcLevel(key_type k) { + __device__ size_t calcLevel(key_type k) const { + // printf("Samples\n"); auto index = this->searcher.sample_cdf_custom_version(this->samples, k) / slice_size; + // printf("Samples finish\n"); auto level = trailing_zeroes(index); return level; } @@ -162,6 +168,12 @@ public: //__device__ bool Delete(key_type); __device__ bool Search(key_type) const; + __device__ void testSample() { + for (int i = 0; i < sampleLength; i++) { + printf("%lld\n", this->samples[i]); + } + } + #ifdef MEASURE_ACCESS unsigned access_times = 0; @@ -193,6 +205,7 @@ public: __device__ Node *GetNewNode(key_type key, unsigned int *pointerIndex, int topLevel) { + printf("%d\n", topLevel); key_type ind = atomicInc(pointerIndex, NUM_ITEMS); Node *n = nodes[ind]; n->key = key; @@ -328,7 +341,9 @@ __device__ bool LockFreeSkipList::Search(key_type key) const { }*/ __device__ bool LockFreeSkipList::Add(key_type key) { - Node *newNode = GetNewNode(key, this->pointerIndex, calcLevel(key)); + auto level1 = calcLevel(key) - 1; + printf("%lu\n", level1); + Node *newNode = GetNewNode(key, this->pointerIndex, level1); int topLevel = newNode->topLevel; int bottomLevel = 0; Node *preds[MAX_LEVEL + 1]; diff --git a/sortlib.cuh b/sortlib.cuh index ae0ea67..4c339ca 100644 --- a/sortlib.cuh +++ b/sortlib.cuh @@ -55,6 +55,7 @@ public: const key_type val) const { // int step_limit = (int)fast_log(LENGTH); + // printf("Binary search\n"); key_type *last_known_point = start; auto son = 0UL; |