Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | avoid division by zero errorHEADmaster | Bjoern B. Brandenburg | 2010-11-12 |
| | |||
* | generate preemptions and migrations with equal probability | Bjoern B. Brandenburg | 2010-11-12 |
| | |||
* | add memtrash, a background workload task | Bjoern B. Brandenburg | 2010-11-09 |
| | | | | This is the old background workload task used for the synthetic experiments. | ||
* | add .gitignore | Bjoern B. Brandenburg | 2010-11-09 |
| | |||
* | Make sure we have the same number of migrations and preemptions. | Bjoern B. Brandenburg | 2010-11-09 |
| | | | | This skips unneeded samples once the desired number of migrations has been observed. | ||
* | cli() and sti() depend on asm/irq.h | Bjoern B. Brandenburg | 2010-11-09 |
| | |||
* | allow cross-compilation for ARM targets | Bjoern B. Brandenburg | 2010-11-09 |
| | | | | This uses the pre-defined make rules from liblitmus. | ||
* | Add "synthetic method" cache affinity loss experiment | Bjoern B. Brandenburg | 2010-04-12 |
| | | | | | | | | Important things: (commit 98d3dac1385deae671a106691294c150ddaf3f26 wip-bbb-cache-test) - Allocate memory buffers from static allocation. This avoids issues with the kernel giving us pre-warmed pages from the LRU list. | ||
* | Fix missing header directory | Andrea Bastoni | 2010-04-12 |
| | | | | - cpmd header directory is missing from include path | ||
* | Add various test scripts | Andrea Bastoni | 2010-04-12 |
| | | | | - also useful for normal debugging, not just PM | ||
* | Add schedule-sensitive PM samples analyzer | Andrea Bastoni | 2010-04-12 |
| | | | | - main data analyzer for preemption and migration (schedule-sensitive method) | ||
* | Add IQR outliers removal | Andrea Bastoni | 2010-04-12 |
| | | | | | - this code is ugly C style code "transformed" in Python should be reworked at some point... | ||
* | Add picke wrapper and csv writer functions | Andrea Bastoni | 2010-04-12 |
| | |||
* | Add module for C 2 Python interaction | Andrea Bastoni | 2010-04-12 |
| | | | | | | | | | | | | | | | | | The following are implemented: static PyMethodDef PmMethods[] = { {"load", pm_load, METH_VARARGS, "Load data from raw files"}, {"getPreemption", pm_get_preemption, METH_VARARGS, "Get preemption overheads - length"}, {"getL2Migration", pm_get_samel2, METH_VARARGS, "Get L2 Migration overheads - length"}, {"getOnChipMigration", pm_get_samechip, METH_VARARGS, "Get Chip (L2 or L3) overheads - length"}, {"getOffChipMigration", pm_get_offchip, METH_VARARGS, "Get Off Chip overheads - length"}, {NULL, NULL, 0, NULL} }; | ||
* | Add best-effort cache polluter | Andrea Bastoni | 2010-04-12 |
| | |||
* | Access WSSs with interrupts disabled | Andrea Bastoni | 2010-04-12 |
| | |||
* | Add main pm test program | Andrea Bastoni | 2010-04-12 |
| | | | | - pm_task and architectural dependent memory access | ||
* | Cache-related Preemption and Migration delays testing programs | Andrea Bastoni | 2010-04-12 |
Add common functionalities - data structures - serialize / read traces - valid samples analyzer - categorize L2 / L3 migrations |