Added a resource constraint, AtomicALUOperation, to GLC atomics performed in the TCC. The resource constraint uses a new class, ALUFreeList array. The class assumes the following: - There are a fixed number of atomic ALU pipelines - While a new cache line can be processed in each pipeline each cycle, if a cache line is currently going through a pipeline, it can't be processed again until it's finished Two configuration parameters have been used to tune this behavior: - tcc-num-atomic-alus corresponds to the number of atomic ALU pipelines - atomic-alu-latency corresponds to the latency of atomic ALU pipelines Change-Id: I25bdde7dafc3877590bb6536efdf57b8c540a939
800 lines
26 KiB
C++
800 lines
26 KiB
C++
/*
|
|
* Copyright (c) 2020-2021 ARM Limited
|
|
* All rights reserved
|
|
*
|
|
* The license below extends only to copyright in the software and shall
|
|
* not be construed as granting a license to any other intellectual
|
|
* property including but not limited to intellectual property relating
|
|
* to a hardware implementation of the functionality of the software
|
|
* licensed hereunder. You may use the software subject to the license
|
|
* terms below provided that you ensure that this notice is replicated
|
|
* unmodified and in its entirety in all distributions of the software,
|
|
* modified or unmodified, in source code or in binary form.
|
|
*
|
|
* Copyright (c) 1999-2012 Mark D. Hill and David A. Wood
|
|
* Copyright (c) 2013 Advanced Micro Devices, Inc.
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions are
|
|
* met: redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer;
|
|
* 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;
|
|
* neither the name of the copyright holders nor the names of its
|
|
* contributors may be used to endorse or promote products derived from
|
|
* this software without specific prior written permission.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
* "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 THE COPYRIGHT
|
|
* OWNER OR 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.
|
|
*/
|
|
|
|
#include "mem/ruby/structures/CacheMemory.hh"
|
|
|
|
#include "base/compiler.hh"
|
|
#include "base/intmath.hh"
|
|
#include "base/logging.hh"
|
|
#include "debug/HtmMem.hh"
|
|
#include "debug/RubyCache.hh"
|
|
#include "debug/RubyCacheTrace.hh"
|
|
#include "debug/RubyResourceStalls.hh"
|
|
#include "debug/RubyStats.hh"
|
|
#include "mem/cache/replacement_policies/weighted_lru_rp.hh"
|
|
#include "mem/ruby/protocol/AccessPermission.hh"
|
|
#include "mem/ruby/system/RubySystem.hh"
|
|
|
|
namespace gem5
|
|
{
|
|
|
|
namespace ruby
|
|
{
|
|
|
|
std::ostream&
|
|
operator<<(std::ostream& out, const CacheMemory& obj)
|
|
{
|
|
obj.print(out);
|
|
out << std::flush;
|
|
return out;
|
|
}
|
|
|
|
CacheMemory::CacheMemory(const Params &p)
|
|
: SimObject(p),
|
|
dataArray(p.dataArrayBanks, p.dataAccessLatency,
|
|
p.start_index_bit, p.ruby_system),
|
|
tagArray(p.tagArrayBanks, p.tagAccessLatency,
|
|
p.start_index_bit, p.ruby_system),
|
|
atomicALUArray(p.atomicALUs, p.atomicLatency *
|
|
p.ruby_system->clockPeriod()),
|
|
cacheMemoryStats(this)
|
|
{
|
|
m_cache_size = p.size;
|
|
m_cache_assoc = p.assoc;
|
|
m_replacementPolicy_ptr = p.replacement_policy;
|
|
m_start_index_bit = p.start_index_bit;
|
|
m_is_instruction_only_cache = p.is_icache;
|
|
m_resource_stalls = p.resourceStalls;
|
|
m_block_size = p.block_size; // may be 0 at this point. Updated in init()
|
|
m_use_occupancy = dynamic_cast<replacement_policy::WeightedLRU*>(
|
|
m_replacementPolicy_ptr) ? true : false;
|
|
}
|
|
|
|
void
|
|
CacheMemory::init()
|
|
{
|
|
if (m_block_size == 0) {
|
|
m_block_size = RubySystem::getBlockSizeBytes();
|
|
}
|
|
m_cache_num_sets = (m_cache_size / m_cache_assoc) / m_block_size;
|
|
assert(m_cache_num_sets > 1);
|
|
m_cache_num_set_bits = floorLog2(m_cache_num_sets);
|
|
assert(m_cache_num_set_bits > 0);
|
|
|
|
m_cache.resize(m_cache_num_sets,
|
|
std::vector<AbstractCacheEntry*>(m_cache_assoc, nullptr));
|
|
replacement_data.resize(m_cache_num_sets,
|
|
std::vector<ReplData>(m_cache_assoc, nullptr));
|
|
// instantiate all the replacement_data here
|
|
for (int i = 0; i < m_cache_num_sets; i++) {
|
|
for ( int j = 0; j < m_cache_assoc; j++) {
|
|
replacement_data[i][j] =
|
|
m_replacementPolicy_ptr->instantiateEntry();
|
|
}
|
|
}
|
|
}
|
|
|
|
CacheMemory::~CacheMemory()
|
|
{
|
|
if (m_replacementPolicy_ptr)
|
|
delete m_replacementPolicy_ptr;
|
|
for (int i = 0; i < m_cache_num_sets; i++) {
|
|
for (int j = 0; j < m_cache_assoc; j++) {
|
|
delete m_cache[i][j];
|
|
}
|
|
}
|
|
}
|
|
|
|
// convert a Address to its location in the cache
|
|
int64_t
|
|
CacheMemory::addressToCacheSet(Addr address) const
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
return bitSelect(address, m_start_index_bit,
|
|
m_start_index_bit + m_cache_num_set_bits - 1);
|
|
}
|
|
|
|
// Given a cache index: returns the index of the tag in a set.
|
|
// returns -1 if the tag is not found.
|
|
int
|
|
CacheMemory::findTagInSet(int64_t cacheSet, Addr tag) const
|
|
{
|
|
assert(tag == makeLineAddress(tag));
|
|
// search the set for the tags
|
|
auto it = m_tag_index.find(tag);
|
|
if (it != m_tag_index.end())
|
|
if (m_cache[cacheSet][it->second]->m_Permission !=
|
|
AccessPermission_NotPresent)
|
|
return it->second;
|
|
return -1; // Not found
|
|
}
|
|
|
|
// Given a cache index: returns the index of the tag in a set.
|
|
// returns -1 if the tag is not found.
|
|
int
|
|
CacheMemory::findTagInSetIgnorePermissions(int64_t cacheSet,
|
|
Addr tag) const
|
|
{
|
|
assert(tag == makeLineAddress(tag));
|
|
// search the set for the tags
|
|
auto it = m_tag_index.find(tag);
|
|
if (it != m_tag_index.end())
|
|
return it->second;
|
|
return -1; // Not found
|
|
}
|
|
|
|
// Given an unique cache block identifier (idx): return the valid address
|
|
// stored by the cache block. If the block is invalid/notpresent, the
|
|
// function returns the 0 address
|
|
Addr
|
|
CacheMemory::getAddressAtIdx(int idx) const
|
|
{
|
|
Addr tmp(0);
|
|
|
|
int set = idx / m_cache_assoc;
|
|
assert(set < m_cache_num_sets);
|
|
|
|
int way = idx - set * m_cache_assoc;
|
|
assert (way < m_cache_assoc);
|
|
|
|
AbstractCacheEntry* entry = m_cache[set][way];
|
|
if (entry == NULL ||
|
|
entry->m_Permission == AccessPermission_Invalid ||
|
|
entry->m_Permission == AccessPermission_NotPresent) {
|
|
return tmp;
|
|
}
|
|
return entry->m_Address;
|
|
}
|
|
|
|
bool
|
|
CacheMemory::tryCacheAccess(Addr address, RubyRequestType type,
|
|
DataBlock*& data_ptr)
|
|
{
|
|
DPRINTF(RubyCache, "address: %#x\n", address);
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
if (entry != nullptr) {
|
|
// Do we even have a tag match?
|
|
m_replacementPolicy_ptr->touch(entry->replacementData);
|
|
entry->setLastAccess(curTick());
|
|
data_ptr = &(entry->getDataBlk());
|
|
|
|
if (entry->m_Permission == AccessPermission_Read_Write) {
|
|
return true;
|
|
}
|
|
if ((entry->m_Permission == AccessPermission_Read_Only) &&
|
|
(type == RubyRequestType_LD || type == RubyRequestType_IFETCH)) {
|
|
return true;
|
|
}
|
|
// The line must not be accessible
|
|
}
|
|
data_ptr = NULL;
|
|
return false;
|
|
}
|
|
|
|
bool
|
|
CacheMemory::testCacheAccess(Addr address, RubyRequestType type,
|
|
DataBlock*& data_ptr)
|
|
{
|
|
DPRINTF(RubyCache, "address: %#x\n", address);
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
if (entry != nullptr) {
|
|
// Do we even have a tag match?
|
|
m_replacementPolicy_ptr->touch(entry->replacementData);
|
|
entry->setLastAccess(curTick());
|
|
data_ptr = &(entry->getDataBlk());
|
|
|
|
return entry->m_Permission != AccessPermission_NotPresent;
|
|
}
|
|
|
|
data_ptr = NULL;
|
|
return false;
|
|
}
|
|
|
|
// tests to see if an address is present in the cache
|
|
bool
|
|
CacheMemory::isTagPresent(Addr address) const
|
|
{
|
|
const AbstractCacheEntry* const entry = lookup(address);
|
|
if (entry == nullptr) {
|
|
// We didn't find the tag
|
|
DPRINTF(RubyCache, "No tag match for address: %#x\n", address);
|
|
return false;
|
|
}
|
|
DPRINTF(RubyCache, "address: %#x found\n", address);
|
|
return true;
|
|
}
|
|
|
|
// Returns true if there is:
|
|
// a) a tag match on this address or there is
|
|
// b) an unused line in the same cache "way"
|
|
bool
|
|
CacheMemory::cacheAvail(Addr address) const
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
|
|
int64_t cacheSet = addressToCacheSet(address);
|
|
|
|
for (int i = 0; i < m_cache_assoc; i++) {
|
|
AbstractCacheEntry* entry = m_cache[cacheSet][i];
|
|
if (entry != NULL) {
|
|
if (entry->m_Address == address ||
|
|
entry->m_Permission == AccessPermission_NotPresent) {
|
|
// Already in the cache or we found an empty entry
|
|
return true;
|
|
}
|
|
} else {
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
AbstractCacheEntry*
|
|
CacheMemory::allocate(Addr address, AbstractCacheEntry *entry)
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
assert(!isTagPresent(address));
|
|
assert(cacheAvail(address));
|
|
DPRINTF(RubyCache, "address: %#x\n", address);
|
|
|
|
// Find the first open slot
|
|
int64_t cacheSet = addressToCacheSet(address);
|
|
std::vector<AbstractCacheEntry*> &set = m_cache[cacheSet];
|
|
for (int i = 0; i < m_cache_assoc; i++) {
|
|
if (!set[i] || set[i]->m_Permission == AccessPermission_NotPresent) {
|
|
if (set[i] && (set[i] != entry)) {
|
|
warn_once("This protocol contains a cache entry handling bug: "
|
|
"Entries in the cache should never be NotPresent! If\n"
|
|
"this entry (%#x) is not tracked elsewhere, it will memory "
|
|
"leak here. Fix your protocol to eliminate these!",
|
|
address);
|
|
}
|
|
set[i] = entry; // Init entry
|
|
set[i]->m_Address = address;
|
|
set[i]->m_Permission = AccessPermission_Invalid;
|
|
DPRINTF(RubyCache, "Allocate clearing lock for addr: 0x%x\n",
|
|
address);
|
|
set[i]->m_locked = -1;
|
|
m_tag_index[address] = i;
|
|
set[i]->setPosition(cacheSet, i);
|
|
set[i]->replacementData = replacement_data[cacheSet][i];
|
|
set[i]->setLastAccess(curTick());
|
|
|
|
// Call reset function here to set initial value for different
|
|
// replacement policies.
|
|
m_replacementPolicy_ptr->reset(entry->replacementData);
|
|
|
|
return entry;
|
|
}
|
|
}
|
|
panic("Allocate didn't find an available entry");
|
|
}
|
|
|
|
void
|
|
CacheMemory::deallocate(Addr address)
|
|
{
|
|
DPRINTF(RubyCache, "address: %#x\n", address);
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
assert(entry != nullptr);
|
|
m_replacementPolicy_ptr->invalidate(entry->replacementData);
|
|
uint32_t cache_set = entry->getSet();
|
|
uint32_t way = entry->getWay();
|
|
delete entry;
|
|
m_cache[cache_set][way] = NULL;
|
|
m_tag_index.erase(address);
|
|
}
|
|
|
|
// Returns with the physical address of the conflicting cache line
|
|
Addr
|
|
CacheMemory::cacheProbe(Addr address) const
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
assert(!cacheAvail(address));
|
|
|
|
int64_t cacheSet = addressToCacheSet(address);
|
|
std::vector<ReplaceableEntry*> candidates;
|
|
for (int i = 0; i < m_cache_assoc; i++) {
|
|
candidates.push_back(static_cast<ReplaceableEntry*>(
|
|
m_cache[cacheSet][i]));
|
|
}
|
|
return m_cache[cacheSet][m_replacementPolicy_ptr->
|
|
getVictim(candidates)->getWay()]->m_Address;
|
|
}
|
|
|
|
// looks an address up in the cache
|
|
AbstractCacheEntry*
|
|
CacheMemory::lookup(Addr address)
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
int64_t cacheSet = addressToCacheSet(address);
|
|
int loc = findTagInSet(cacheSet, address);
|
|
if (loc == -1) return NULL;
|
|
return m_cache[cacheSet][loc];
|
|
}
|
|
|
|
// looks an address up in the cache
|
|
const AbstractCacheEntry*
|
|
CacheMemory::lookup(Addr address) const
|
|
{
|
|
assert(address == makeLineAddress(address));
|
|
int64_t cacheSet = addressToCacheSet(address);
|
|
int loc = findTagInSet(cacheSet, address);
|
|
if (loc == -1) return NULL;
|
|
return m_cache[cacheSet][loc];
|
|
}
|
|
|
|
// Sets the most recently used bit for a cache block
|
|
void
|
|
CacheMemory::setMRU(Addr address)
|
|
{
|
|
AbstractCacheEntry* entry = lookup(makeLineAddress(address));
|
|
if (entry != nullptr) {
|
|
m_replacementPolicy_ptr->touch(entry->replacementData);
|
|
entry->setLastAccess(curTick());
|
|
}
|
|
}
|
|
|
|
void
|
|
CacheMemory::setMRU(AbstractCacheEntry *entry)
|
|
{
|
|
assert(entry != nullptr);
|
|
m_replacementPolicy_ptr->touch(entry->replacementData);
|
|
entry->setLastAccess(curTick());
|
|
}
|
|
|
|
void
|
|
CacheMemory::setMRU(Addr address, int occupancy)
|
|
{
|
|
AbstractCacheEntry* entry = lookup(makeLineAddress(address));
|
|
if (entry != nullptr) {
|
|
// m_use_occupancy can decide whether we are using WeightedLRU
|
|
// replacement policy. Depending on different replacement policies,
|
|
// use different touch() function.
|
|
if (m_use_occupancy) {
|
|
static_cast<replacement_policy::WeightedLRU*>(
|
|
m_replacementPolicy_ptr)->touch(
|
|
entry->replacementData, occupancy);
|
|
} else {
|
|
m_replacementPolicy_ptr->touch(entry->replacementData);
|
|
}
|
|
entry->setLastAccess(curTick());
|
|
}
|
|
}
|
|
|
|
int
|
|
CacheMemory::getReplacementWeight(int64_t set, int64_t loc)
|
|
{
|
|
assert(set < m_cache_num_sets);
|
|
assert(loc < m_cache_assoc);
|
|
int ret = 0;
|
|
if (m_cache[set][loc] != NULL) {
|
|
ret = m_cache[set][loc]->getNumValidBlocks();
|
|
assert(ret >= 0);
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
void
|
|
CacheMemory::recordCacheContents(int cntrl, CacheRecorder* tr) const
|
|
{
|
|
uint64_t warmedUpBlocks = 0;
|
|
[[maybe_unused]] uint64_t totalBlocks = (uint64_t)m_cache_num_sets *
|
|
(uint64_t)m_cache_assoc;
|
|
|
|
for (int i = 0; i < m_cache_num_sets; i++) {
|
|
for (int j = 0; j < m_cache_assoc; j++) {
|
|
if (m_cache[i][j] != NULL) {
|
|
AccessPermission perm = m_cache[i][j]->m_Permission;
|
|
RubyRequestType request_type = RubyRequestType_NULL;
|
|
if (perm == AccessPermission_Read_Only) {
|
|
if (m_is_instruction_only_cache) {
|
|
request_type = RubyRequestType_IFETCH;
|
|
} else {
|
|
request_type = RubyRequestType_LD;
|
|
}
|
|
} else if (perm == AccessPermission_Read_Write) {
|
|
request_type = RubyRequestType_ST;
|
|
}
|
|
|
|
if (request_type != RubyRequestType_NULL) {
|
|
Tick lastAccessTick;
|
|
lastAccessTick = m_cache[i][j]->getLastAccess();
|
|
tr->addRecord(cntrl, m_cache[i][j]->m_Address,
|
|
0, request_type, lastAccessTick,
|
|
m_cache[i][j]->getDataBlk());
|
|
warmedUpBlocks++;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
DPRINTF(RubyCacheTrace, "%s: %lli blocks of %lli total blocks"
|
|
"recorded %.2f%% \n", name().c_str(), warmedUpBlocks,
|
|
totalBlocks, (float(warmedUpBlocks) / float(totalBlocks)) * 100.0);
|
|
}
|
|
|
|
void
|
|
CacheMemory::print(std::ostream& out) const
|
|
{
|
|
out << "Cache dump: " << name() << std::endl;
|
|
for (int i = 0; i < m_cache_num_sets; i++) {
|
|
for (int j = 0; j < m_cache_assoc; j++) {
|
|
if (m_cache[i][j] != NULL) {
|
|
out << " Index: " << i
|
|
<< " way: " << j
|
|
<< " entry: " << *m_cache[i][j] << std::endl;
|
|
} else {
|
|
out << " Index: " << i
|
|
<< " way: " << j
|
|
<< " entry: NULL" << std::endl;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
void
|
|
CacheMemory::printData(std::ostream& out) const
|
|
{
|
|
out << "printData() not supported" << std::endl;
|
|
}
|
|
|
|
void
|
|
CacheMemory::setLocked(Addr address, int context)
|
|
{
|
|
DPRINTF(RubyCache, "Setting Lock for addr: %#x to %d\n", address, context);
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
assert(entry != nullptr);
|
|
entry->setLocked(context);
|
|
}
|
|
|
|
void
|
|
CacheMemory::clearLocked(Addr address)
|
|
{
|
|
DPRINTF(RubyCache, "Clear Lock for addr: %#x\n", address);
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
assert(entry != nullptr);
|
|
entry->clearLocked();
|
|
}
|
|
|
|
void
|
|
CacheMemory::clearLockedAll(int context)
|
|
{
|
|
// iterate through every set and way to get a cache line
|
|
for (auto i = m_cache.begin(); i != m_cache.end(); ++i) {
|
|
std::vector<AbstractCacheEntry*> set = *i;
|
|
for (auto j = set.begin(); j != set.end(); ++j) {
|
|
AbstractCacheEntry *line = *j;
|
|
if (line && line->isLocked(context)) {
|
|
DPRINTF(RubyCache, "Clear Lock for addr: %#x\n",
|
|
line->m_Address);
|
|
line->clearLocked();
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
bool
|
|
CacheMemory::isLocked(Addr address, int context)
|
|
{
|
|
AbstractCacheEntry* entry = lookup(address);
|
|
assert(entry != nullptr);
|
|
DPRINTF(RubyCache, "Testing Lock for addr: %#llx cur %d con %d\n",
|
|
address, entry->m_locked, context);
|
|
return entry->isLocked(context);
|
|
}
|
|
|
|
CacheMemory::
|
|
CacheMemoryStats::CacheMemoryStats(statistics::Group *parent)
|
|
: statistics::Group(parent),
|
|
ADD_STAT(numDataArrayReads, "Number of data array reads"),
|
|
ADD_STAT(numDataArrayWrites, "Number of data array writes"),
|
|
ADD_STAT(numTagArrayReads, "Number of tag array reads"),
|
|
ADD_STAT(numTagArrayWrites, "Number of tag array writes"),
|
|
ADD_STAT(numTagArrayStalls, "Number of stalls caused by tag array"),
|
|
ADD_STAT(numDataArrayStalls, "Number of stalls caused by data array"),
|
|
ADD_STAT(numAtomicALUOperations, "Number of atomic ALU operations"),
|
|
ADD_STAT(numAtomicALUArrayStalls, "Number of stalls caused by atomic ALU array"),
|
|
ADD_STAT(htmTransCommitReadSet, "Read set size of a committed "
|
|
"transaction"),
|
|
ADD_STAT(htmTransCommitWriteSet, "Write set size of a committed "
|
|
"transaction"),
|
|
ADD_STAT(htmTransAbortReadSet, "Read set size of a aborted transaction"),
|
|
ADD_STAT(htmTransAbortWriteSet, "Write set size of a aborted "
|
|
"transaction"),
|
|
ADD_STAT(m_demand_hits, "Number of cache demand hits"),
|
|
ADD_STAT(m_demand_misses, "Number of cache demand misses"),
|
|
ADD_STAT(m_demand_accesses, "Number of cache demand accesses",
|
|
m_demand_hits + m_demand_misses),
|
|
ADD_STAT(m_prefetch_hits, "Number of cache prefetch hits"),
|
|
ADD_STAT(m_prefetch_misses, "Number of cache prefetch misses"),
|
|
ADD_STAT(m_prefetch_accesses, "Number of cache prefetch accesses",
|
|
m_prefetch_hits + m_prefetch_misses),
|
|
ADD_STAT(m_accessModeType, "")
|
|
{
|
|
numDataArrayReads
|
|
.flags(statistics::nozero);
|
|
|
|
numDataArrayWrites
|
|
.flags(statistics::nozero);
|
|
|
|
numTagArrayReads
|
|
.flags(statistics::nozero);
|
|
|
|
numTagArrayWrites
|
|
.flags(statistics::nozero);
|
|
|
|
numTagArrayStalls
|
|
.flags(statistics::nozero);
|
|
|
|
numDataArrayStalls
|
|
.flags(statistics::nozero);
|
|
|
|
numAtomicALUOperations
|
|
.flags(statistics::nozero);
|
|
|
|
numAtomicALUArrayStalls
|
|
.flags(statistics::nozero);
|
|
|
|
htmTransCommitReadSet
|
|
.init(8)
|
|
.flags(statistics::pdf | statistics::dist | statistics::nozero |
|
|
statistics::nonan);
|
|
|
|
htmTransCommitWriteSet
|
|
.init(8)
|
|
.flags(statistics::pdf | statistics::dist | statistics::nozero |
|
|
statistics::nonan);
|
|
|
|
htmTransAbortReadSet
|
|
.init(8)
|
|
.flags(statistics::pdf | statistics::dist | statistics::nozero |
|
|
statistics::nonan);
|
|
|
|
htmTransAbortWriteSet
|
|
.init(8)
|
|
.flags(statistics::pdf | statistics::dist | statistics::nozero |
|
|
statistics::nonan);
|
|
|
|
m_prefetch_hits
|
|
.flags(statistics::nozero);
|
|
|
|
m_prefetch_misses
|
|
.flags(statistics::nozero);
|
|
|
|
m_prefetch_accesses
|
|
.flags(statistics::nozero);
|
|
|
|
m_accessModeType
|
|
.init(RubyRequestType_NUM)
|
|
.flags(statistics::pdf | statistics::total);
|
|
|
|
for (int i = 0; i < RubyAccessMode_NUM; i++) {
|
|
m_accessModeType
|
|
.subname(i, RubyAccessMode_to_string(RubyAccessMode(i)))
|
|
.flags(statistics::nozero)
|
|
;
|
|
}
|
|
}
|
|
|
|
// assumption: SLICC generated files will only call this function
|
|
// once **all** resources are granted
|
|
void
|
|
CacheMemory::recordRequestType(CacheRequestType requestType, Addr addr)
|
|
{
|
|
DPRINTF(RubyStats, "Recorded statistic: %s\n",
|
|
CacheRequestType_to_string(requestType));
|
|
switch(requestType) {
|
|
case CacheRequestType_DataArrayRead:
|
|
if (m_resource_stalls)
|
|
dataArray.reserve(addressToCacheSet(addr));
|
|
cacheMemoryStats.numDataArrayReads++;
|
|
return;
|
|
case CacheRequestType_DataArrayWrite:
|
|
if (m_resource_stalls)
|
|
dataArray.reserve(addressToCacheSet(addr));
|
|
cacheMemoryStats.numDataArrayWrites++;
|
|
return;
|
|
case CacheRequestType_TagArrayRead:
|
|
if (m_resource_stalls)
|
|
tagArray.reserve(addressToCacheSet(addr));
|
|
cacheMemoryStats.numTagArrayReads++;
|
|
return;
|
|
case CacheRequestType_TagArrayWrite:
|
|
if (m_resource_stalls)
|
|
tagArray.reserve(addressToCacheSet(addr));
|
|
cacheMemoryStats.numTagArrayWrites++;
|
|
return;
|
|
case CacheRequestType_AtomicALUOperation:
|
|
if (m_resource_stalls)
|
|
atomicALUArray.reserve(addr);
|
|
cacheMemoryStats.numAtomicALUOperations++;
|
|
return;
|
|
default:
|
|
warn("CacheMemory access_type not found: %s",
|
|
CacheRequestType_to_string(requestType));
|
|
}
|
|
}
|
|
|
|
bool
|
|
CacheMemory::checkResourceAvailable(CacheResourceType res, Addr addr)
|
|
{
|
|
if (!m_resource_stalls) {
|
|
return true;
|
|
}
|
|
|
|
if (res == CacheResourceType_TagArray) {
|
|
if (tagArray.tryAccess(addressToCacheSet(addr))) return true;
|
|
else {
|
|
DPRINTF(RubyResourceStalls,
|
|
"Tag array stall on addr %#x in set %d\n",
|
|
addr, addressToCacheSet(addr));
|
|
cacheMemoryStats.numTagArrayStalls++;
|
|
return false;
|
|
}
|
|
} else if (res == CacheResourceType_DataArray) {
|
|
if (dataArray.tryAccess(addressToCacheSet(addr))) return true;
|
|
else {
|
|
DPRINTF(RubyResourceStalls,
|
|
"Data array stall on addr %#x in set %d\n",
|
|
addr, addressToCacheSet(addr));
|
|
cacheMemoryStats.numDataArrayStalls++;
|
|
return false;
|
|
}
|
|
} else if (res == CacheResourceType_AtomicALUArray) {
|
|
if (atomicALUArray.tryAccess(addr)) return true;
|
|
else {
|
|
DPRINTF(RubyResourceStalls,
|
|
"Atomic ALU array stall on addr %#x in line address %#x\n",
|
|
addr, makeLineAddress(addr));
|
|
cacheMemoryStats.numAtomicALUArrayStalls++;
|
|
return false;
|
|
}
|
|
} else {
|
|
panic("Unrecognized cache resource type.");
|
|
}
|
|
}
|
|
|
|
bool
|
|
CacheMemory::isBlockInvalid(int64_t cache_set, int64_t loc)
|
|
{
|
|
return (m_cache[cache_set][loc]->m_Permission == AccessPermission_Invalid);
|
|
}
|
|
|
|
bool
|
|
CacheMemory::isBlockNotBusy(int64_t cache_set, int64_t loc)
|
|
{
|
|
return (m_cache[cache_set][loc]->m_Permission != AccessPermission_Busy);
|
|
}
|
|
|
|
/* hardware transactional memory */
|
|
|
|
void
|
|
CacheMemory::htmAbortTransaction()
|
|
{
|
|
uint64_t htmReadSetSize = 0;
|
|
uint64_t htmWriteSetSize = 0;
|
|
|
|
// iterate through every set and way to get a cache line
|
|
for (auto i = m_cache.begin(); i != m_cache.end(); ++i)
|
|
{
|
|
std::vector<AbstractCacheEntry*> set = *i;
|
|
|
|
for (auto j = set.begin(); j != set.end(); ++j)
|
|
{
|
|
AbstractCacheEntry *line = *j;
|
|
|
|
if (line != nullptr) {
|
|
htmReadSetSize += (line->getInHtmReadSet() ? 1 : 0);
|
|
htmWriteSetSize += (line->getInHtmWriteSet() ? 1 : 0);
|
|
if (line->getInHtmWriteSet()) {
|
|
line->invalidateEntry();
|
|
}
|
|
line->setInHtmWriteSet(false);
|
|
line->setInHtmReadSet(false);
|
|
line->clearLocked();
|
|
}
|
|
}
|
|
}
|
|
|
|
cacheMemoryStats.htmTransAbortReadSet.sample(htmReadSetSize);
|
|
cacheMemoryStats.htmTransAbortWriteSet.sample(htmWriteSetSize);
|
|
DPRINTF(HtmMem, "htmAbortTransaction: read set=%u write set=%u\n",
|
|
htmReadSetSize, htmWriteSetSize);
|
|
}
|
|
|
|
void
|
|
CacheMemory::htmCommitTransaction()
|
|
{
|
|
uint64_t htmReadSetSize = 0;
|
|
uint64_t htmWriteSetSize = 0;
|
|
|
|
// iterate through every set and way to get a cache line
|
|
for (auto i = m_cache.begin(); i != m_cache.end(); ++i)
|
|
{
|
|
std::vector<AbstractCacheEntry*> set = *i;
|
|
|
|
for (auto j = set.begin(); j != set.end(); ++j)
|
|
{
|
|
AbstractCacheEntry *line = *j;
|
|
if (line != nullptr) {
|
|
htmReadSetSize += (line->getInHtmReadSet() ? 1 : 0);
|
|
htmWriteSetSize += (line->getInHtmWriteSet() ? 1 : 0);
|
|
line->setInHtmWriteSet(false);
|
|
line->setInHtmReadSet(false);
|
|
line->clearLocked();
|
|
}
|
|
}
|
|
}
|
|
|
|
cacheMemoryStats.htmTransCommitReadSet.sample(htmReadSetSize);
|
|
cacheMemoryStats.htmTransCommitWriteSet.sample(htmWriteSetSize);
|
|
DPRINTF(HtmMem, "htmCommitTransaction: read set=%u write set=%u\n",
|
|
htmReadSetSize, htmWriteSetSize);
|
|
}
|
|
|
|
void
|
|
CacheMemory::profileDemandHit()
|
|
{
|
|
cacheMemoryStats.m_demand_hits++;
|
|
}
|
|
|
|
void
|
|
CacheMemory::profileDemandMiss()
|
|
{
|
|
cacheMemoryStats.m_demand_misses++;
|
|
}
|
|
|
|
void
|
|
CacheMemory::profilePrefetchHit()
|
|
{
|
|
cacheMemoryStats.m_prefetch_hits++;
|
|
}
|
|
|
|
void
|
|
CacheMemory::profilePrefetchMiss()
|
|
{
|
|
cacheMemoryStats.m_prefetch_misses++;
|
|
}
|
|
|
|
} // namespace ruby
|
|
} // namespace gem5
|