Instead of sending puts back-to-back, separate the two puts with a get.
Also, stall a bit between each transaction. This makes sure the puts and
intermediate get are sent to the same transactor, which will cause the
data buffer to get overwritten between the two puts.
(1) Introduce tracegen.py, a script that invokes the emulator (built
with TraceGenConfig), sending a SIGTERM once all cores are finished.
(2) Update toaxe.py to gather some statistics about the trace.
(3) Introduce tracestats.py, which displays the stats in a useful way.
(4) Introduce tracegen+check.py, a top-level script that generates
traces, checks them, and emits stats. If this commit is pulled, it
should be done after pulling my latest groundtest commit.
To elide several races between reading and writing the metadata array for different types of transactions, all L2XactTrackers can now sink Voluntary Releases (writebacks from the L1 in the current implementation). These writebacks are merged with the ongoing transaction, and the merging tracker supplies an acknowledgment of the writeback in addition to its ongoing activities. This change involved another refactoring of the control logic for allocating new trackers and routing incoming Acquires and Releases. BroadcastHub uses the new routing logic, but still processes all voluntary releases through the VoluntaryReleaseTracker (not a problem because there are no metadata update races).
Closes#18Closes#20
This uses an reorder queue but doesn't check to ensure that the data it fetches
from the queue is actually in the queue before using it. It seems that during
correct operation this never breaks, but I'm trying to get the backup memory
port working again and this assertion fails with it enabled (without the
assertion the core just gets a bogus data beat dies).
Closes#16
Exposes new parameter field SplitMetadata to determine whether the metadata array in a particular cache is stored in a single SeqMem or with one SeqMem per way.
Closes#14
