phoenix
phoenix copied to clipboard
an API and runtime environment for data processing with MapReduce for shared-memory multi-core & multiprocessor systems.
## Heap buffer overflow in `string_match` benchmark when `malloc` is used. ### How to reproduce: #### Compile the benchmarks with a patch that enables the `NO_MMAP` flag and `AddressSanitizer` (`ASan`):...
## What is this PR doing: Fixes #9 This PR adapts the buffer allocation for the `string_match` benchmark when the `NO_MMAP` flag is set to eliminate the heap buffer overflow...
In linear_regression example: `max_procs` defined in line 427 (MapReduceScheduler.c) in `#ifdef _SOLARIS_` not defined for later usage in line 473 if not `_SOLARIS_`
For simplicity and reproducible reasons, I built a Docker container for Phoenix 2 that also contains the data-sets. The container is available online via `docker pull ghugo/phoenix-2.0`. More information on...
The archive with inputs for linear_regression: http://csl.stanford.edu/~christos/data/linear_regression.tar.gz is wrong. It contains the inputs for string_match. The README and the source code say that linear_regression "reads the x-y coordinates sequentially from...