DoubangoTelecom compv issues

DoubangoTelecom / compv

Insanely fast Open Source Computer Vision library for ARM and x86 devices (Up to #50 times faster than OpenCV)

https://doubango.org

GNU General Public License v3.0

194 stars 41 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Use CompVPtrXXXX everywhere

#76 DoubangoTelecom opened 8 years ago
0
Use reference (&) instead of pointer (*) in return args

#75 DoubangoTelecom opened 8 years ago
0
Replace sizeof(float) == 4 with compv_float32_t

#74 DoubangoTelecom opened 8 years ago
0
COMPV_ASSERT must break on release mode

#73 DoubangoTelecom opened 8 years ago
0
Some AVX2 functions should be AVX1

#72 DoubangoTelecom opened 8 years ago
0
Rectangular matrixes, optimize using MCOP

#71 DoubangoTelecom opened 8 years ago
0
Dimensional reduction for the covariance matrix (Eigen values/vectors)

#70 DoubangoTelecom closed 8 years ago
0
FAST detector, pushing interest points: optimization issue

#69 DoubangoTelecom opened 8 years ago
1
Samples

#68 DoubangoTelecom opened 8 years ago
0
When testing mode, memZero after malloc() to make sure to have same MD5 result

#67 DoubangoTelecom opened 8 years ago
0
Decreasing CPU usage / Speedup

#66 DoubangoTelecom opened 8 years ago
0
Add list of customers on doubango.org

#65 DoubangoTelecom opened 8 years ago
0
Disable x86 ASM when X64 version is available

#64 DoubangoTelecom opened 8 years ago
0
add support for fxp brief decription

#63 DoubangoTelecom opened 8 years ago
0
Add support for fxp convolution

#62 DoubangoTelecom closed 8 years ago
0
Promote float to compv_float32_t when sizeof(float) != 4

#61 DoubangoTelecom opened 8 years ago
0
convlt1_verthz: VS2013 intrin faster than our ASM (not the case for Intel C++)

#60 DoubangoTelecom opened 8 years ago
1
Add support for Intel® Integrated Performance Primitives (Intel® IPP)

#59 DoubangoTelecom opened 8 years ago
0
adds support for Short Vector Math Library (SVML) Functions

#58 DoubangoTelecom opened 8 years ago
0
Call vzeroupper after COMPV_YASM_RESTORE_YMM

#57 sarandogou closed 8 years ago
0
Description, uses capacity() and size() to avoid allocating memory for each process

#56 sarandogou closed 8 years ago
0
Remove ComVDescription and replace it with CompVArray

#55 sarandogou closed 8 years ago
0
Win64 calling convention: stack must be aligned on 16bytes *after* params allocation -> SystemV redzone

#54 DoubangoTelecom opened 8 years ago
0
getThreadIdxForNextToCurrentCore not correct, must use getThreadIdxCurrent() instead of getCoreId()

#53 sarandogou opened 8 years ago
0
do not use -march=native on xcode

#52 sarandogou opened 8 years ago
0
Apple LLVM: fatal error: error in backend: Cannot select: intrinsic %llvm.x86.avx.vzeroupper

#51 sarandogou closed 8 years ago
0
Add support for frame-based multi-threading

#50 DoubangoTelecom closed 8 years ago
1
RGBA->I420 with odd width doesn't work or YUV player is buggy

#49 DoubangoTelecom opened 8 years ago
0
ImageConv, add x64 versions

#48 DoubangoTelecom opened 8 years ago
1
Replace COMPV_YASM_SAVE_XMM with COMPV_YASM_SAVE_YMM in AVX code

#47 DoubangoTelecom closed 8 years ago
0
use vzeroall instead of vzeroupper and omit vpxor ymm, ymm

#46 DoubangoTelecom closed 8 years ago
0
Add FastStrenghts32 implementation using AVX-only code and VPHMINPOSUW

#45 DoubangoTelecom closed 8 years ago
0
Use "__stdcall" on all ASM extern functions to avoid surprises

#44 DoubangoTelecom closed 8 years ago
0
Add unit tests with resolution change

#43 DoubangoTelecom opened 8 years ago
0
Make memory checking thread-safe

#42 DoubangoTelecom closed 8 years ago
0
cache optimisations and many other stuffs (e.g cupid) must be tested on AMD CPUs

#41 DoubangoTelecom opened 8 years ago
0
Make box sorting multi-threaded

#40 DoubangoTelecom closed 8 years ago
0
Cache alignment issue on all SIMD code

#39 DoubangoTelecom opened 8 years ago
1
On windows, save XMM6-n registers for avx

#38 DoubangoTelecom closed 8 years ago
0
Compute function pointers onces and update when CPU flags is update

#37 DoubangoTelecom opened 8 years ago
0
replace all cmovs with setcc when possible. The later is supported on all i386

#36 DoubangoTelecom opened 8 years ago
0
False sharing issue on interest points vector for FAST

#35 DoubangoTelecom closed 8 years ago
0
CompVImage::wrap is very slooow

#34 DoubangoTelecom opened 8 years ago
0
RGBA->I420 C++ version doesn't work

#33 DoubangoTelecom opened 8 years ago
0
COMPV_INTERLEAVE_I8_XMM_SSE2, reverse line 2/3

#32 DoubangoTelecom closed 8 years ago
0
Use dr memory for memory checks (works on Windows unlike valgrind)

#31 DoubangoTelecom opened 8 years ago
0
Use valgrind for profiling

#30 DoubangoTelecom opened 8 years ago
0
Products

#29 DoubangoTelecom opened 8 years ago
0
COMPV_YASM_SAVE_RSP_TO_MM0, COMPV_YASM_RESTORE_RSP_FROM_MM0

#28 DoubangoTelecom closed 8 years ago
0
compv_popcnt16 is sloow because of the test.

#27 DoubangoTelecom closed 8 years ago
0

Previous Next