Activity for SETI@home
-
SETI@home committed [8b8a2a]
Change to to allow a single instantiation to process multiple files
-
Turned test mode off.
-
Committing some old mods, mainly for FAST data processing.
-
Start of script to generate plots for all new data.
-
Use julian date instead of unix time for seconds into file.
-
New directory for plotting scripts.
-
Include stdint.i.
-
Small mods.
-
Add SNR to print_hits_table().
-
Merge branch 'master' of https://github.com/UCBerkeleySETI/etfits_api
-
Add internal init_meta_data flag.
-
FAST specific init script.
-
Correct number of source nodes for FAST.
-
Add FAST specific code to Makefile.
-
Enable getting the receiver name from redis.
-
Fetch ADC RMS's from redis and place them in the FITS files.
-
Add RUNSECS to status shared memory.
-
Add getStartTimeStamp().
-
Properly pull pointing data from redis for FAST.
-
Add GPU memory allocation timer.
-
Remove the non-production configs.
-
Make cudaThreadSynchronize() easy to turn off/on (and turn it off).
-
More separation of CUB and non-CUB versions of spectroscopy().
-
Correct placement of an #endif.
-
Make cub debugging easier to turn on/off.
-
Convenience script for building for different deployments.
-
Merge branch 'master' of khan.ssl.berkeley.edu:local/src/serendip6_fast
-
Hashpipe's fitshead.h now takes care of int typing.
-
Improve usefulness and readability of timing stats.
-
Add "s6" to start of log file names.
-
More multicasting mods.
-
Make log file names sensible.
-
Add a REALLOC_NONE version of spectroscopy().
-
Prep for separating CUB and nonCUB versions of spectroscopy() for FAST.
-
Move some commented out code to clarify its relationship to running code.
-
Output name of template file on error.
-
Simple script to make the /proc changes necessary for multicasting.
-
Obtain observatory meta-data from the FAST observatory redis server.
-
Move the semaphore post to after freeing all cached memory allocations!
-
Pass source ID (and thus pol number) on to the output thread.
-
Add a free_all_cached() after the FFT.
-
Init script now takes a beam number and uses instances 1 and 2.
-
Add POWTHRSH to hashpipe command line.
-
Mods to allow cub allocator integration.
-
Merge branch 'master' of khan.ssl.berkeley.edu:local/src/serendip6
-
Fixed bug in voltage dumping code.
-
Added status shmem fields for dropped packets percentages.
-
Merge branch 'master' of khan.ssl.berkeley.edu:local/src/serendip6_fast
-
Use unqualified wild card in /sys/class/net/.
-
S6_FAST_LIB_INCLUDES
-
Remove unused transpose data structure allocation.
-
Convert FAST spectroscopy() over to using cub allocation.
-
Sleep only after all instances are up.
-
init_device_vectors() no longer dimensions GPU structures.
-
Drop the R2C FFT "padding" element from down stream processing.
-
Ifdef the entirety of the GPU vector declarations.
-
Add getStartTimeStamp().
-
Move init_device_vectors() to spectroscopy().
-
Get beampol into FITS header for FAST.
-
Restore non-CUB version of device_vectors_t.
-
Checkpoint of code to dump voltages to file.
-
CPU affinity change.
-
Set per interface rp_filter to 0.
-
Improve debug packet header output.
-
No coarse channel powers HDU for FAST.
-
For FAST there is only one missed_pkts element for each input block.
-
Clean up memory management timing tracking.
-
Print GPU memory tracking to stdout rather than stderr.
-
Clean up gpu (thrust calls) timing tracking.
-
Comment out s6_memcpy() stuff for FAST.
-
Beam number as a cmd line parm for FAST init script.
-
Adapt to the fact that FAST source ID indexes beams to 1.
-
Mod the use of status shmem for threshold and nhits.
-
Provide an interface to the CUB caching allocator class.
-
Add beam and pol to FAST status shmem to support multicasting.
1