I think you can do that by generating multiple requests along with the main
request to the dram. You can do that in memory controller or the cpu side.
Specific details depends upon how many cache lines you require and their
addresses. But i think its not that much difficult. See the trace of some
high spatial locality benchmark like libquantum, so from the trace you can
reverse engineer how to generate consecutive memory addresses.

On Apr 29, 2017 3:04 AM, "Muzamil Rafique" <[email protected]>
wrote:

Hi All,

When we have a miss in LLC and found that address in row-buffer (row
currently open) in main memory, how can we bring multiple cache-lines
residing in that row buffer (along with requested one) in to caches? Is
there currently any support available in gem5?

Thanks
Muzamil

_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to