Difference between revisions of "Google Summer of Code"

From gem5
Jump to: navigation, search
Line 10: Line 10:
 
# Directory Protocol
 
# Directory Protocol
 
# Real In-order core model
 
# Real In-order core model
#* Kevin has one that works to some degree in SE, it doesn't have functional units yet, but it does have variable latency stages and such
+
#* There is code to start with but nothing that is fully fleshed out.
#* Korey has one he did at MIPS, I don't know about it's features, but it's SE only  as well
 
 
# Write a PLI interface to connect Verilog CPUs to the memory system.
 
# Write a PLI interface to connect Verilog CPUs to the memory system.
 
# Sampling/fast-forwarding techniques
 
# Sampling/fast-forwarding techniques

Revision as of 19:25, 11 March 2008

All the ideas listed here will require some familiarity with Python and a good grasp of advanced C++ concepts.

  1. Build a direct execution CPU model based on the Linux Kernel Virtual Machine
  2. Parallelize M5
    • Use the Wisconsin Wind Tunnel as a guide
    • This actually isn't as bad as it sounds as all objects schedule their own events and there are limited ways they can interact with other objects in the system.
  3. Memory network models
    • (e.g. Crossbar or Mesh)
  4. Directory Protocol
  5. Real In-order core model
    • There is code to start with but nothing that is fully fleshed out.
  6. Write a PLI interface to connect Verilog CPUs to the memory system.
  7. Sampling/fast-forwarding techniques
    • This would have the most impact if it was coupled with (1)
    • Using SMARTS work would be a good guide
  8. Flash memory device model (seems popular nowadays)
    • This could be a hard drive based model like we're seeing in laptops now or a memory device model like several research papers have suggested as storage in between DRAM and disk.