-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AIWC #201
Open
BeauJoh
wants to merge
203
commits into
jrprice:master
Choose a base branch
from
BeauJoh:master
base: master
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Open
AIWC #201
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…eated/run, this can be used to bound some of the other metrics in the feature-space
…when a barrier has been hit and cleared in a workitem
…and metrics to AIWC
…ller than the desired history score
… was leading to an unsigned integer underflow
… kernel execution
…ew metrics are printed at the end of the OpenCL program, since they are outside the scope of any kernel. To this end, these results are printed to a new aiwc_memory_transfers_*.csv file.
…labels are named during this stage of IR simulation
Don't store the size of memory operations as it's not used in later analysis. To determine the number of unique memory addresses, just use the keys of the map that's used to store the count of accesses against each address. In calculation of Local Memory Address Entropy, the address was being cast from an unsigned to a signed int, which is incorrect for addresses where the first bit is a one. Instead of doing a division (by float), just use a right shift by the chosen number of bits.
Calculation of granularity was using inverse of number of instructions, but should have been inverse of number of work items. Move granularity and barriers-per-instruction to Thread-Level Parallelism (renamed from Instruction Level Parallelism).
…branch_history_size from memory to control
Merge in changes from original Oclgrind
Recover Benjamin's commit on generic memory space
…ing the OpenCL ICD loader
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Adds Architecture-Independent Workload Characterization (AIWC) plugin
Adds support for -DLLVM_ROOT_DIR=/llvm cmake argument along with the more verbose -DLLVM_DIR=/llvm/lib/cmake/llvm
Added support for Generic address space support --- needed for running DPC++-OpenCL SYCL implementation
p.s. sorry for having such a long convoluted history of commits over many files rather than one nice clean change.