This contains my bachelors thesis and associated tex files, code snippets and maybe more. Topic: Data Movement in Heterogeneous Memories with Intel Data Streaming Accelerator
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

352 lines
13 KiB

#pragma once
#include <iostream>
#include <unordered_map>
#include <shared_mutex>
#include <mutex>
#include <memory>
#include <sched.h>
#include <numa.h>
#include <numaif.h>
#include <dml/dml.hpp>
#include "cache-data.hpp"
namespace dsacache {
// cache class will handle access to data through the cache
// by managing the cache through work submission, it sticks
// to user-defined caching and copy policies, is thread
// safe after initialization and returns copies of
// cache data class to the user
class Cache {
public:
// cache policy is defined as a type here to allow flexible usage of the cacher
// given a numa destination node (where the data will be needed), the numa source
// node (current location of the data) and the data size, this function should
// return optimal cache placement
// dst node and returned value can differ if the system, for example, has HBM
// attached accessible directly to node n under a different node id m
typedef int (CachePolicy)(const int numa_dst_node, const int numa_src_node, const size_t data_size);
// copy policy specifies the copy-executing nodes for a given task
// which allows flexibility in assignment for optimizing raw throughput
// or choosing a conservative usage policy
typedef std::vector<int> (CopyPolicy)(const int numa_dst_node, const int numa_src_node);
private:
// mutex for accessing the cache state map
std::shared_mutex cache_mutex_;
// map from [dst-numa-node,map2]
// map2 from [data-ptr,cache-structure]
std::unordered_map<uint8_t, std::unordered_map<uint8_t*, CacheData>> cache_state_;
CachePolicy* cache_policy_function_ = nullptr;
CopyPolicy* copy_policy_function_ = nullptr;
// function used to submit a copy task on a specific node to the dml
// engine on that node - will change the current threads node assignment
// to achieve this so take care to restore this
dml::handler<dml::mem_copy_operation, std::allocator<uint8_t>> ExecuteCopy(
const uint8_t* src, uint8_t* dst, const size_t size, const int node
) const;
// allocates the required memory on the destination node
// and then submits task to the dml library for processing
// and attaches the handlers to the cache data structure
void SubmitTask(CacheData* task, const int dst_node, const int src_node);
// querries the policy functions for the given data and size
// to obtain destination cache node, also returns the datas
// source node for further usage
// output may depend on the calling threads node assignment
// as this is set as the "optimal placement" node
void GetCacheNode(uint8_t* src, const size_t size, int* OUT_DST_NODE, int* OUT_SRC_NODE) const;
// checks whether the cache contains an entry for
// the given data in the given memory node and
// returns it, otherwise returns nullptr
std::unique_ptr<CacheData> GetFromCache(uint8_t* src, const size_t size, const int dst_node);
public:
// initializes the cache with the two policy functions
// only after this is it safe to use in a threaded environment
void Init(CachePolicy* cache_policy_function, CopyPolicy* copy_policy_function);
// function to perform data access through the cache
std::unique_ptr<CacheData> Access(uint8_t* data, const size_t size);
// flushes the cache of inactive entries
// if node is -1 then the whole cache is
// checked and otherwise the specified
// node - no checks on node validity
void Flush(const int node = -1);
};
}
inline void dsacache::Cache::Init(CachePolicy* cache_policy_function, CopyPolicy* copy_policy_function) {
cache_policy_function_ = cache_policy_function;
copy_policy_function_ = copy_policy_function;
// initialize numa library
numa_available();
// obtain all available nodes
// and those we may allocate
// memory on
const int nodes_max = numa_num_configured_nodes();
const bitmask* valid_nodes = numa_get_mems_allowed();
// prepare the cache state with entries
// for all given nodes
for (int node = 0; node < nodes_max; node++) {
if (numa_bitmask_isbitset(valid_nodes, node)) {
cache_state_.insert({node,{}});
}
}
std::cout << "[-] Cache Initialized" << std::endl;
}
inline std::unique_ptr<dsacache::CacheData> dsacache::Cache::Access(uint8_t* data, const size_t size) {
// get destination numa node for the cache
int dst_node = -1;
int src_node = -1;
GetCacheNode(data, size, &dst_node, &src_node);
// TODO: at this point it could be beneficial to check whether
// TODO: the given destination node is present as an entry
// TODO: in the cache state to see if it is valid
// check whether the data is already cached
std::unique_ptr<CacheData> task = GetFromCache(data, size, dst_node);
if (task != nullptr) {
return std::move(task);
}
// at this point the requested data is not present in cache
// and we create a caching task for it
task = std::make_unique<CacheData>(data, size);
{
std::unique_lock<std::shared_mutex> lock(cache_mutex_);
const auto state = cache_state_[dst_node].emplace(task->src_, *task);
// if state.second is false then no insertion took place
// which means that concurrently whith this thread
// some other thread must have accessed the same
// resource in which case we return the other
// threads data cache structure
if (!state.second) {
std::cout << "[!] Found another cache instance for 0x" << std::hex << (uint64_t)task->src_ << std::dec << std::endl;
return std::move(std::make_unique<CacheData>(state.first->second));
}
}
SubmitTask(task.get(), dst_node, src_node);
return std::move(task);
}
inline void dsacache::Cache::SubmitTask(CacheData* task, const int dst_node, const int src_node) {
std::cout << "[+] Allocating " << task->size_ << "B on node " << dst_node << " for " << std::hex << (uint64_t)task->src_ << std::dec << std::endl;
// allocate data on this node and flush the unused parts of the
// cache if the operation fails and retry once
// TODO: smarter flush strategy could keep some stuff cached
uint8_t* dst = reinterpret_cast<uint8_t*>(numa_alloc_onnode(task->size_, dst_node));
if (dst == nullptr) {
std::cout << "[!] First allocation try failed for " << task->size_ << "B on node " << dst_node << std::endl;
// allocation on dst_node failed so we flush the cache for this
// node hoping to free enough currently unused entries to make
// the second allocation attempt successful
Flush(dst_node);
dst = reinterpret_cast<uint8_t*>(numa_alloc_onnode(task->size_, dst_node));
if (dst == nullptr) {
std::cerr << "[x] Second allocation try failed for " << task->size_ << "B on node " << dst_node << std::endl;
return;
}
}
task->incomplete_cache_ = dst;
// querry copy policy function for the nodes to use for the copy
const std::vector<int> executing_nodes = copy_policy_function_(dst_node, src_node);
const size_t task_count = executing_nodes.size();
// each task will copy one fair part of the total size
// and in case the total size is not a factor of the
// given task count the last node must copy the remainder
const size_t size = task->size_ / task_count;
const size_t last_size = size + task->size_ % task_count;
std::cout << "[-] Splitting Copy into " << task_count << " tasks of " << size << "B 0x" << std::hex << (uint64_t)task->src_ << std::dec << std::endl;
// save the current numa node mask to restore later
// as executing the copy task will place this thread
// on a different node
bitmask* nodemask = numa_get_run_node_mask();
for (uint32_t i = 0; i < task_count; i++) {
const size_t local_size = i + 1 == task_count ? size : last_size;
const size_t local_offset = i * size;
const uint8_t* local_src = task->src_ + local_offset;
uint8_t* local_dst = dst + local_offset;
task->handlers_->emplace_back(ExecuteCopy(local_src, local_dst, local_size, executing_nodes[i]));
}
// restore the previous nodemask
numa_run_on_node_mask(nodemask);
numa_free_nodemask(nodemask);
}
inline dml::handler<dml::mem_copy_operation, std::allocator<uint8_t>> dsacache::Cache::ExecuteCopy(
const uint8_t* src, uint8_t* dst, const size_t size, const int node
) const {
dml::const_data_view srcv = dml::make_view(src, size);
dml::data_view dstv = dml::make_view(dst, size);
numa_run_on_node(node);
return dml::submit<dml::automatic>(dml::mem_copy.block_on_fault(), srcv, dstv);
}
inline void dsacache::Cache::GetCacheNode(uint8_t* src, const size_t size, int* OUT_DST_NODE, int* OUT_SRC_NODE) const {
// obtain numa node of current thread to determine where the data is needed
const int current_cpu = sched_getcpu();
const int current_node = numa_node_of_cpu(current_cpu);
// obtain node that the given data pointer is allocated on
*OUT_SRC_NODE = -1;
get_mempolicy(OUT_SRC_NODE, NULL, 0, (void*)src, MPOL_F_NODE | MPOL_F_ADDR);
// querry cache policy function for the destination numa node
*OUT_DST_NODE = cache_policy_function_(current_node, *OUT_SRC_NODE, size);
}
inline void dsacache::Cache::Flush(const int node) {
std::cout << "[-] Flushing Cache for " << (node == -1 ? "all nodes" : "node " + std::to_string(node)) << std::endl;
// this lambda is used because below we have two code paths that
// flush nodes, either one single or all successively
const auto FlushNode = [](std::unordered_map<uint8_t*,CacheData>& map) {
// begin at the front of the map
auto it = map.begin();
// loop until we reach the end of the map
while (it != map.end()) {
// if the iterator points to an inactive element
// then we may erase it
if (it->second.Active() == false) {
// erase the iterator from the map
map.erase(it);
// as the erasure invalidated out iterator
// we must start at the beginning again
it = map.begin();
}
else {
// if element is active just move over to the next one
it++;
}
}
};
{
// we require exclusive lock as we modify the cache state
std::unique_lock<std::shared_mutex> lock(cache_mutex_);
// node == -1 means that cache on all nodes should be flushed
if (node == -1) {
for (auto& nc : cache_state_) {
FlushNode(nc.second);
}
}
else {
FlushNode(cache_state_[node]);
}
}
}
inline std::unique_ptr<dsacache::CacheData> dsacache::Cache::GetFromCache(uint8_t* src, const size_t size, const int dst_node) {
// the best situation is if this data is already cached
// which we check in an unnamed block in which the cache
// is locked for reading to prevent another thread
// from marking the element we may find as unused and
// clearing it
// lock the cache state in shared-mode because we read
std::shared_lock<std::shared_mutex> lock(cache_mutex_);
// search for the data in our cache state structure at the given node
const auto search = cache_state_[dst_node].find(src);
// if the data is in our structure we continue
if (search != cache_state_[dst_node].end()) {
// now check whether the sizes match
// TODO: second.size_ >= size would also work
if (search->second.size_ == size) {
std::cout << "[+] Found Cached version for 0x" << std::hex << (uint64_t)src << std::dec << std::endl;
// return a unique copy of the entry which uses the object
// lifetime and destructor to safely handle deallocation
return std::move(std::make_unique<CacheData>(search->second));
}
else {
std::cout << "[!] Found Cached version with size missmatch for 0x" << std::hex << (uint64_t)src << std::dec << std::endl;
// if the sizes missmatch then we clear the current entry from cache
// which will cause its deletion only after the last possible outside
// reference is also destroyed
cache_state_[dst_node].erase(search);
}
}
return nullptr;
}