Daniel Carvalho has uploaded this change for review. ( https://gem5-review.googlesource.com/c/public/gem5/+/14295

Change subject: mem-cache: Remove extra latency check
......................................................................

mem-cache: Remove extra latency check

There was a check to see if the difference between the current
tick and the whenReady time was greater than the access latency.
This did not make any sense.

What must be accounted for is the time the block is available
to be accessed, and that is achievable by just applying the
access latency on top of the greater between the tick at which
the access occurs and the tick at which the block is ready.

We assume the tag lookup can only be done when the block's data
is ready to be accessed, although in practice there could be
two different whenReady times, one for the tag and one for the
data. Thus, we are penalizing tag lookups with data fill lat.

Change-Id: I2b2937833f5bf16d8f4e7ca2cb0efbc5635c6219
Signed-off-by: Daniel R. Carvalho <[email protected]>
---
M src/mem/cache/base.cc
1 file changed, 1 insertion(+), 2 deletions(-)



diff --git a/src/mem/cache/base.cc b/src/mem/cache/base.cc
index 4f83b6a..734b2c4 100644
--- a/src/mem/cache/base.cc
+++ b/src/mem/cache/base.cc
@@ -923,8 +923,7 @@

// Check if the block to be accessed is available. If not, apply the
         // access latency on top of block->whenReady.
-        if (blk->whenReady > curTick() &&
-            ticksToCycles(blk->whenReady - curTick()) > lat) {
+        if (blk->whenReady > curTick()) {
             lat += ticksToCycles(blk->whenReady - curTick());
         }
     }

--
To view, visit https://gem5-review.googlesource.com/c/public/gem5/+/14295
To unsubscribe, or for help writing mail filters, visit https://gem5-review.googlesource.com/settings

Gerrit-Project: public/gem5
Gerrit-Branch: master
Gerrit-Change-Id: I2b2937833f5bf16d8f4e7ca2cb0efbc5635c6219
Gerrit-Change-Number: 14295
Gerrit-PatchSet: 1
Gerrit-Owner: Daniel Carvalho <[email protected]>
Gerrit-MessageType: newchange
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to