Back
< next | previous >
23/08/2010

Mandelbrot now timed. 10 lines checked with Simulator and running on the µblaze.
- on the Simulator it suggests that it should take around 0.192 seconds.
- from the timing from the µblaze is says 0.395 seconds.

Need to do some work on the Simulator to get it to be more correct. Also need to test how close the timing function is.
23/08/2010

Tasks:
1. get CoreMark running on the µblaze - done
2. get timer working - done

#include "xtmrctr_l.h"
XTmrCtr_mEnable(XPAR_TMRCTR_0_BASEADDR, 1);
start_time = XTmrCtr_mGetTimerCounterReg(XPAR_TMRCTR_0_BASEADDR, 1);
/* work */
end_time = XTmrCtr_mGetTimerCounterReg(XPAR_TMRCTR_0_BASEADDR, 1);

time in seconds is then (end_time - start_time) / CPU_CORE_FREQUENCY

3. time Mandelbrot
4. Simulator issues (branch losses, decode losses, memory allowance issues)
18/08/2010

got the xilinx kernel running with a test application running 2 threads, 1 checking for button presses, 1 for screen drawing.

need to look into linux for the µblaze, i have had a look at a few (http://www.monstr.eu/wiki/doku.php?id=fdt:fdt & http://xilinx.wikidot.com/microblaze-linux) but they don't seem to work.

also need to add some extra work into the simulator, at the moment the simulator assumes perfect execution. need to add the following:
- branching (for a taken branch need to add 3 stall cycles)
- instruction fetching (need to check boundaries and add stall cycles if a single bundle overlaps)
16/08/2010

got XPS working, the SDK doesn't seem to like importing projects from XPS due to extra files being required. Mark has got it working. We had an issue with the memory system due to sizes and offsets. This has possibly been resolved now, need to look into it more.

The move_main function has now been fixed, passing arguments to main isn't possible due to the addresses needing to be offset by the stack_bottom and causing the values to be too small/big.

Need to finish the second year report and also get CoreMark running on multiple threads on the Simulator.
06/08/2010

Need to wait for Mark or V to get back to get the LE1 code to actually be able to program the board.

The arguments to main still isn't working, i can't figure out why, i need to check the moving main.

Also need to write the second year report:

- A four page report (YR2RSReport) from the student is required in month eleven of the second year. This report will state the work completed and identify significant achievements to date, confirm research novelty and refine the work plan until end of the third year. The report should identify two publications that will be submitted for refereeing during the third year or have already been submitted in the second year. The student should aim to support the Supervisor as joint author on these.

- Research Student and Supervisor should agree an outline thesis plan with Gantt chart and include this in the YR2RSReport.