diff options
author | Ali Saidi <saidi@eecs.umich.edu> | 2011-09-13 12:58:08 -0400 |
---|---|---|
committer | Ali Saidi <saidi@eecs.umich.edu> | 2011-09-13 12:58:08 -0400 |
commit | 649c239ceef2d107fae253b1008c6f214f242d73 (patch) | |
tree | 06edeb7bfa7b9d1e9167b9cb80f7f9600d41b9a0 /src/cpu/o3/lsq_impl.hh | |
parent | bb921b1459ef3ec55f9cea4ac8d203cd3c801cfd (diff) | |
download | gem5-649c239ceef2d107fae253b1008c6f214f242d73.tar.xz |
LSQ: Only trigger a memory violation with a load/load if the value changes.
Only create a memory ordering violation when the value could have changed
between two subsequent loads, instead of just when loads go out-of-order
to the same address. While not very common in the case of Alpha, with
an architecture with a hardware table walker this can happen reasonably
frequently beacuse a translation will miss and start a table walk and
before the CPU re-schedules the faulting instruction another one will
pass it to the same address (or cache block depending on the dendency
checking).
This patch has been tested with a couple of self-checking hand crafted
programs to stress ordering between two cores.
The performance improvement on SPEC benchmarks can be substantial (2-10%).
Diffstat (limited to 'src/cpu/o3/lsq_impl.hh')
-rw-r--r-- | src/cpu/o3/lsq_impl.hh | 14 |
1 files changed, 10 insertions, 4 deletions
diff --git a/src/cpu/o3/lsq_impl.hh b/src/cpu/o3/lsq_impl.hh index 41b0f3041..ef9167d8c 100644 --- a/src/cpu/o3/lsq_impl.hh +++ b/src/cpu/o3/lsq_impl.hh @@ -90,11 +90,17 @@ LSQ<Impl>::DcachePort::recvTiming(PacketPtr pkt) DPRINTF(LSQ, "Got error packet back for address: %#X\n", pkt->getAddr()); if (pkt->isResponse()) { lsq->thread[pkt->req->threadId()].completeDataAccess(pkt); - } - else { - // must be a snoop + } else { + DPRINTF(LSQ, "received pkt for addr:%#x %s\n", pkt->getAddr(), + pkt->cmdString()); - // @TODO someday may need to process invalidations in LSQ here + // must be a snoop + if (pkt->isInvalidate()) { + DPRINTF(LSQ, "received invalidation for addr:%#x\n", pkt->getAddr()); + for (ThreadID tid = 0; tid < lsq->numThreads; tid++) { + lsq->thread[tid].checkSnoop(pkt); + } + } // to provide stronger consistency model } return true; |