Activity for seti-boinc
-
Eric J. Korpela committed [53d812] on Code
Fix use of tranposed PoT in standalone mode
-
Added new analysis_config field autocorr_min_delay_bins. Should be set to 1
-
Fix required for workunits with more than 8 bits per sample.
-
Updated Makefiles to include splitter_fast/fast_splitter
-
splitter_fast copied from splitter_pfb before merge of Eric's changes
-
Fixes for Ubuntu 20 compile
-
-
-
-
-
-
charlief committed [51c7d2]
Fix Compile error on Mac (memset undefined in this context)
-
charlief committed [32fe4e]
Mac: Update XCode project for V6 graphics; fix compile error on Mac
-
-
Added DevC++ project file for the graphics app.
-
-
charlief committed [dff340]
Put version number definitions in mac_build/sahconfig.h instead of in 2 places (config-i386.h and config-ppc.h)
-
charlief committed [a9a744]
Fix bug in XCode project (ppc targets had Run Script phase after build instead of before.)
-
charlief committed [262908]
Mac: Fix definition of COMPILER_STRING, consolidate version number definitions, set version to 6.00.
-
charlief committed [fc3b6a]
Mac: Fix definition of COMPILER_STRING, consolidate version number definitions, set version to 6.00, fix bug in XCode project.
-
Fix for unitialized data being included in SIMD folding routines. Submitted by
-
charlief committed [05e9ec]
Mac: add icon to graphics application, eliminate unused libraries and search paths from XCode project.
-
Jeff Coburn committed [95b178]
Move repository branches and tags.
-
Fixed location of v5_graphics code.
-
Move repository branches and tags.
-
-
Moved seti_boinc specific tags into tags/seti_boinc
-
Tagged compile for v6.00 beta for windows.
-
-
Sources for Windows 6.02 build.
-
-
Linux 6.02 build code
-
-
Raistmer committed [b6f843]
tagged AP 6.04 release for ATi&NV GPUs under Windows
-
Code used for Windows and Linux S@Hv7 7.00 builds
-
charlief committed [bdbe1c]
Mac: Updates for building with 7.0.58 BOINC libraries
-
tag for AP 7.00 GPU release on beta
-
Added sah_v8_RC1 tag.
-
Minor changes from RC1
-
v8.19 Windows OpenCL GPU release
-
Move repository branches and tags.
-
Bob Johnston 12/17/08 sources...
-
-
-
Fix of triplet bug and optimiztion of cuda app build flags.
-
-
Fix for VLAR bug
-
Checkin of CUDA 2.3 setiathome_enhanced (6.09) client modifications.
-
TuanLe committed [65c3a7]
- add nvcc.rules file. This missing file causes XML parsing error while opening seti_boinc VS++ project
-
TuanLe committed [991607]
- add paths to include directories
-
TuanLe committed [fa5699]
-
TuanLe committed [b9ceff]
-
TuanLe committed [967f1b]
-
TuanLe committed [1ea779]
- complete library linkings and fix include paths
-
TuanLe committed [982a4a]
- fix external linking issues.
-
TuanLe committed [57f66d]
- add #include "str_replace.h"
-
TuanLe committed [fff942]
- add #include "str_replace.h"
-
TuanLe committed [1a3e60]
- clean up code - fix indentation
-
Move repository branches and tags.
-
TuanLe committed [fb9be4]
- add #include "str_replace.h"
-
OpenCL/CUDA/CPU optimized (Lunatics) SaH MultiBeam buids (CPU buid is successor of AKv8, GPU builds contain CPU code from AKv8, some GPU kernels based on initial CUDA_MB release)
-
Initial commit
-
Initial commit
-
Updated BOINC and SETI DB libs required to build new GPU/opt AP and MB apps.
-
optimized AstroPulse V6 for CPU and NV/ATi GPUs corresponding to r534 of Lunatics repositories, team_mod branch
-
Update to match rev375 for MB7 from Lunatics rep.
-
Required updates to match Lunatics rev 417 of team_mod branch for MultiBeam.
-
Few missed files added and unneeded libs removed
-
in AP: Brook config brought to V6 too. Added hint (based on Joe's suggestion) for validator on early exited tasks. Cause GPU builds don't handle best single pulses there will be only 60 and not 70 signals reported on 30/30 overflow for OpenCL builds.
-
in MB OpenCL: implemented refined dechirping based on Jason's changes for CUDA build. This can lead to some slowdown but improves dechirp accuracy that increases validation rate.
-
in AP: fake 100% blanking removed for 30/30 early exit.
-
Numerous changes to match Lunatics AP rev560
-
1)Fixed bug with binary caches and long GPU device names like 9800 GTX/ 9800 GTX+
-
Defaults are changed.
-
MB7 updated to correspond Lunatics r455+
-
1. Switched to BOINC API at rev25550, trunk (to get OpenCL device from BOINC)
-
Reads a file, logical name "cmdline.txt" if it exists and adds its
-
Oops, forgot to add a call to boinc_resolve_filename().
-
persistent cmdline added to MultiBeam too so file names changed to mb_cmdline.txt and ap_cmdline.txt correspondingly.
-
1. -no_cpu_lock switch removed, -cpu_lock switch added (no CPU lock by default now)
-
1. App built with patched BOINC API to recognize OpenCL NV platform
-
1. Default parameters changed again to allow work on ION GPU.
-
FERMI GPUs produce incorrect offsets table in FFA so reverted to CPU for this class of devices for now. Pre-FERMI NV and ATi GPUs produce offsets on device.
-
AP CL file fix
-
Double precision on GPU offsets generation on FERMI path.
-
AP: new switch added:
-
More flexibility added:
-
Made a working directory for cross platform fixes to OpenCL astropulse.
-
Fixed CRLF and tab damage. Put in standard format.
-
Please don't give project header files the same names as standard headers even
-
jason_groothuis committed [84117b]
Test commit of X-branch release docs
-
jason_groothuis committed [3daa54]
Setiathome_enhanced, Cuda Multibeam (V7 enabled), x41z Initial commit.
-
fixes in GPU_TWINCHIRP path
-
jason_groothuis committed [8bb261]
Some Cleanups:
-
jason_groothuis committed [7f8ee8]
Cuda multibeam x41z (Windows)
-
jason_groothuis committed [73f9e8]
Added historical /docs/svnlog.txt. Dates from first full commit post Fermi release, ~17th October 2010 @'x32j', through x41z. Initial commit was made After RAD Fermi compatibility x1 through x32.
-
jason_groothuis committed [01e683]
XBranch: (x41z)
-
AstroPulse:
-
Changes in GPU sync on exit
-
Fixes in non-TWINCHIRP path.
1 >