| Age | Commit message (Collapse) | Author |
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12709]
|
|
learn what)
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12704]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12696]
|
|
replay eddy.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12666]
|
|
more helpful.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12654]
|
|
benchmark is compiled with CUDA4 and cuobjdump is not enabled
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12653]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12650]
|
|
them)
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12636]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12635]
|
|
Fixing shift/reduce conflicts
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12630]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12628]
|
|
A bit of cleanup
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12627]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12625]
|
|
- updating README to move instructions for merging changes to end and add lots of warnings
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12622]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12621]
|
|
allocate on miss (so that it is consistent with the current L1 config).
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12619]
|
|
classes
Some changes to remove warnings
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12609]
|
|
- Added a missed support for the perfect memory configuration option which
models a memory system with 1 cycle latency per request.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12608]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12606]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12581]
|
|
Better wording
Correcting paths
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12576]
|
|
Update gcc, flex and bison versions
Adding a paragraph about how to obtain the simulator via git and how to pull latest changes
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12572]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12564]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12563]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12562]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12558]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12554]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12552]
|
|
Cleaner make clean
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12546]
|
|
Major rework for the Makefile
Deleting decuda_to_ptxplus
Removing all dependencies on boost and removing it from the dependencies of the simulator
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12545]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12523]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12518]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12517]
|
|
Now, the highest sm version below forced_max_capability is selected for each cu file individually
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12516]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12491]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12431]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12384]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12383]
|
|
supported before
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12372]
|
|
just print the stats whenever a kernel is done.
This requires decoupling updating the stats from printing them and modifying the printing code to accomodate this change.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12366]
|
|
could have a few cycles left (e.g. memory).
To prevent the problem with the race condition, no stream operations are allowed to launch until the kernel that has finished finished executing. The loop will exit, the kernel prints its stats, then in the next iteration the next operation is allowed to launch.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12362]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12358]
|
|
Right now this doesn't cause any trouble becaue increment_thread_id guarantees that this particular part of the condition is never false (in correct operation)
Normal kernel configuration have the z dimention smaller than the x dimention which causes it to be always true, but with a weird configuration, this is a potential place for failure.
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12357]
|
|
Merging
//depot/gpgpu_sim_research/fermi_replay/distribution/src/gpgpusim_entrypoint.cc
to //depot/gpgpu_sim_research/fermi/distribution/src/gpgpusim_entrypoint.cc
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12356]
|
|
- FUs depended on the result bus to know if they are going to be used on a certain cycle, this is not the case anymore, occupied bitvectors are added
- A configurable number of result buses is added (the number of buses is equal to the EX_WB pipe width)
- Modified the Fermi config file to add two ports to the operand collector
IPC with a theoretical limit of number_of_SMs*64 is achievable using this configuration
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12349]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12343]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12323]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12311]
|
|
The observed latencies are reduced by 5 to account for other stages in the SM pipeline
Eventually this should be calibrated against the microbenchmarks
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12310]
|
|
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12305]
|