changeset 5ace73846f27 in /z/repo/gem5
details: http://repo.gem5.org/gem5?cmd=changeset;node=5ace73846f27
description:
mem: Just-in-time write scheduling in DRAM controller
This patch removes the untimed while loop in the write scheduling
mechanism and now schedule commands taking into account the minimum
timing constraint. It also introduces an optimization to track write
queue size and switch from writes to reads if the number of write
requests fall below write low threshold.
diffstat:
src/mem/SimpleDRAM.py | 12 ++++-
src/mem/simple_dram.cc | 92 +++++++++++++++++++++++++-------------------------
src/mem/simple_dram.hh | 15 +++++++-
3 files changed, 68 insertions(+), 51 deletions(-)
diffs (233 lines):
diff -r 6ab9ebbebed8 -r 5ace73846f27 src/mem/SimpleDRAM.py
--- a/src/mem/SimpleDRAM.py Fri Nov 01 11:56:24 2013 -0400
+++ b/src/mem/SimpleDRAM.py Fri Nov 01 11:56:25 2013 -0400
@@ -69,12 +69,18 @@
port = SlavePort("Slave port")
# the basic configuration of the controller architecture
- write_buffer_size = Param.Unsigned(32, "Number of read queue entries")
- read_buffer_size = Param.Unsigned(32, "Number of write queue entries")
+ write_buffer_size = Param.Unsigned(32, "Number of write queue entries")
+ read_buffer_size = Param.Unsigned(32, "Number of read queue entries")
# threshold in percent for when to trigger writes and start
# emptying the write buffer as it starts to get full
- write_thresh_perc = Param.Percent(70, "Threshold to trigger writes")
+ write_high_thresh_perc = Param.Percent(70, "Threshold to trigger writes")
+
+ # threshold in percentage for when to stop writes if the read
+ # queue has an entry. An optimisaton to give reads priority if
+ # sufficient number of writes are scheduled and write queue has
+ # sufficient number of free entries
+ write_low_thresh_perc = Param.Percent(0, "Threshold to stop writes")
# scheduler, address map and page policy
mem_sched_policy = Param.MemSched('frfcfs', "Memory scheduling policy")
diff -r 6ab9ebbebed8 -r 5ace73846f27 src/mem/simple_dram.cc
--- a/src/mem/simple_dram.cc Fri Nov 01 11:56:24 2013 -0400
+++ b/src/mem/simple_dram.cc Fri Nov 01 11:56:25 2013 -0400
@@ -67,7 +67,8 @@
banksPerRank(p->banks_per_rank), channels(p->channels), rowsPerBank(0),
readBufferSize(p->read_buffer_size),
writeBufferSize(p->write_buffer_size),
- writeThresholdPerc(p->write_thresh_perc),
+ writeHighThresholdPerc(p->write_high_thresh_perc),
+ writeLowThresholdPerc(p->write_low_thresh_perc),
tWTR(p->tWTR), tBURST(p->tBURST),
tRCD(p->tRCD), tCL(p->tCL), tRP(p->tRP), tRAS(p->tRAS),
tRFC(p->tRFC), tREFI(p->tREFI), tRRD(p->tRRD),
@@ -77,7 +78,8 @@
frontendLatency(p->static_frontend_latency),
backendLatency(p->static_backend_latency),
busBusyUntil(0), writeStartTime(0),
- prevArrival(0), numReqs(0)
+ prevArrival(0), numReqs(0),
+ numWritesThisTime(0), newTime(0)
{
// create the bank states based on the dimensions of the ranks and
// banks
@@ -88,9 +90,10 @@
actTicks[c].resize(activationLimit, 0);
}
- // round the write threshold percent to a whole number of entries
- // in the buffer
- writeThreshold = writeBufferSize * writeThresholdPerc / 100.0;
+ // round the write thresholds percent to a whole number of entries
+ // in the buffer.
+ writeHighThreshold = writeBufferSize * writeHighThresholdPerc / 100.0;
+ writeLowThreshold = writeBufferSize * writeLowThresholdPerc / 100.0;
}
void
@@ -384,29 +387,20 @@
SimpleDRAM::processWriteEvent()
{
assert(!writeQueue.empty());
- uint32_t numWritesThisTime = 0;
- DPRINTF(DRAM, "Beginning DRAM Writes\n");
+ DPRINTF(DRAM, "Beginning DRAM Write\n");
Tick temp1 M5_VAR_USED = std::max(curTick(), busBusyUntil);
Tick temp2 M5_VAR_USED = std::max(curTick(), maxBankFreeAt());
- // @todo: are there any dangers with the untimed while loop?
- while (!writeQueue.empty()) {
- if (numWritesThisTime >= writeThreshold) {
- DPRINTF(DRAM, "Hit write threshold %d\n", writeThreshold);
- break;
- }
+ chooseNextWrite();
+ DRAMPacket* dram_pkt = writeQueue.front();
+ // sanity check
+ assert(dram_pkt->size <= burstSize);
+ doDRAMAccess(dram_pkt);
- chooseNextWrite();
- DRAMPacket* dram_pkt = writeQueue.front();
- // sanity check
- assert(dram_pkt->size <= burstSize);
- doDRAMAccess(dram_pkt);
-
- writeQueue.pop_front();
- delete dram_pkt;
- numWritesThisTime++;
- }
+ writeQueue.pop_front();
+ delete dram_pkt;
+ numWritesThisTime++;
DPRINTF(DRAM, "Completed %d writes, bus busy for %lld ticks,"\
"banks busy for %lld ticks\n", numWritesThisTime,
@@ -415,9 +409,28 @@
// Update stats
avgWrQLen = writeQueue.size();
- // turn the bus back around for reads again
- busBusyUntil += tWTR;
- stopReads = false;
+ if (numWritesThisTime >= writeHighThreshold) {
+ DPRINTF(DRAM, "Hit write threshold %d\n", writeHighThreshold);
+ }
+
+ // If number of writes in the queue fall below the low thresholds and
+ // read queue is not empty then schedule a request event else continue
+ // with writes. The retry above could already have caused it to be
+ // scheduled, so first check
+ if (((writeQueue.size() <= writeLowThreshold) && !readQueue.empty()) ||
+ writeQueue.empty()) {
+ numWritesThisTime = 0;
+ // turn the bus back around for reads again
+ busBusyUntil += tWTR;
+ stopReads = false;
+
+ if (!nextReqEvent.scheduled())
+ schedule(nextReqEvent, busBusyUntil);
+ } else {
+ assert(!writeEvent.scheduled());
+ DPRINTF(DRAM, "Next write scheduled at %lld\n", newTime);
+ schedule(writeEvent, newTime);
+ }
if (retryWrReq) {
retryWrReq = false;
@@ -430,13 +443,6 @@
drainManager->signalDrainDone();
drainManager = NULL;
}
-
- // Once you're done emptying the write queue, check if there's
- // anything in the read queue, and call schedule if required. The
- // retry above could already have caused it to be scheduled, so
- // first check
- if (!nextReqEvent.scheduled())
- schedule(nextReqEvent, busBusyUntil);
}
@@ -565,7 +571,7 @@
accessAndRespond(pkt, frontendLatency);
// If your write buffer is starting to fill up, drain it!
- if (writeQueue.size() >= writeThreshold && !stopReads){
+ if (writeQueue.size() >= writeHighThreshold && !stopReads){
triggerWrites();
}
}
@@ -602,7 +608,7 @@
"Scheduler %s\n" \
"Address mapping %s\n" \
"Page policy %s\n",
- name(), readBufferSize, writeBufferSize, writeThreshold,
+ name(), readBufferSize, writeBufferSize, writeHighThreshold,
scheduler, address_mapping, page_policy);
DPRINTF(DRAM, "Memory controller %s timing specs\n" \
@@ -1119,6 +1125,10 @@
writeRowHits++;
}
+ // Update the minimum timing between the requests
+ newTime = (busBusyUntil > tRP + tRCD + tCL) ?
+ std::max(busBusyUntil - (tRP + tRCD + tCL), curTick()) : curTick();
+
// At this point, commonality between reads and writes ends.
// For writes, we are done since we long ago responded to the
// requestor. We also don't care about stats for writes. For
@@ -1142,23 +1152,13 @@
//time
moveToRespQ();
- // The absolute soonest you have to start thinking about the
- // next request is the longest access time that can occur before
- // busBusyUntil. Assuming you need to precharge,
- // open a new row, and access, it is tRP + tRCD + tCL
-
- Tick newTime = (busBusyUntil > tRP + tRCD + tCL ) ?
- std::max(busBusyUntil - (tRP + tRCD + tCL) , curTick()) :
- curTick();
-
+ // Schedule the next read event
if (!nextReqEvent.scheduled() && !stopReads){
schedule(nextReqEvent, newTime);
} else {
if (newTime < nextReqEvent.when())
reschedule(nextReqEvent, newTime);
}
-
-
}
void
diff -r 6ab9ebbebed8 -r 5ace73846f27 src/mem/simple_dram.hh
--- a/src/mem/simple_dram.hh Fri Nov 01 11:56:24 2013 -0400
+++ b/src/mem/simple_dram.hh Fri Nov 01 11:56:25 2013 -0400
@@ -472,8 +472,10 @@
uint32_t columnsPerRowBuffer;
const uint32_t readBufferSize;
const uint32_t writeBufferSize;
- const double writeThresholdPerc;
- uint32_t writeThreshold;
+ const double writeHighThresholdPerc;
+ uint32_t writeHighThreshold;
+ const double writeLowThresholdPerc;
+ uint32_t writeLowThreshold;
/**
* Basic memory timing parameters initialized based on parameter
@@ -522,6 +524,15 @@
Tick prevArrival;
int numReqs;
+ // Tracks number of writes done to meet the write threshold
+ uint32_t numWritesThisTime;
+
+ // The absolute soonest you have to start thinking about the
+ // next request is the longest access time that can occur before
+ // busBusyUntil. Assuming you need to precharge,
+ // open a new row, and access, it is tRP + tRCD + tCL
+ Tick newTime;
+
// All statistics that the model needs to capture
Stats::Scalar readReqs;
Stats::Scalar writeReqs;
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev