gtcasl / eiger Goto Github PK
View Code? Open in Web Editor NEWHome page for the Eiger statistical performance modeling framework.
Home Page: http://gpuocelot.gatech.edu/projects/eiger/
License: Other
Home page for the Eiger statistical performance modeling framework.
Home Page: http://gpuocelot.gatech.edu/projects/eiger/
License: Other
There are no rules in the Makefile in the api directory to install the libraries or binaries.
Flag to turn on verbose logging in Eiger.py would be nice.
For example, the following code does not work:
PERFLOG(SITENAME, IP(its), IP(nx), IP(numReg), IP(numFiles), IP(showProg),
IP(quiet), IP(viz), IP(balance), IP(cost));
but,
PERFLOG(SITENAME, IP(its), IP(nx), IP(numReg), IP(numFiles), IP(showProg), IP(quiet), IP(viz), IP(balance), IP(cost));
does work.
lwperf does not care if a profiling site is commented out, it tries to log it anyway. A poorly constructed lwperf call that is commented out will still generate compilation errors.
The model files don't contain centroid information for determining which model to poll when using kmeans clustering.
Elements of structs do not seem able to be directly used by lwperf. For example, the following code does not work:
opts.nx = 30;
PERFLOG(somesitename, IP(opts.nx)) fails with an error about the variable not being declared
but,
opts.nx = 30;
int q = opts.nx;
PERFLOG(somesitename, IP(q));
works just fine.
Instead of having the user specify what the file name should be when outputting a model, we should just use the data collection name with a suffix like ".model". This way we can synchronize the lwperf instrumentation, model generation, and interface within SST/macro.
Trying to train on multiple result metrics fails with cryptic error about matrices not lining up.
lwperf calls added to applications already have all the information necessary to inject polling calls for SSTmacro polling. Perhaps a flag set for lwperf that tells it you're using sstmacro and desire poll calls to be added, which may also include some way to turn them on/off.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.