summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2012-03-31roi: Remove some unneeded G_OBJECT castsBenjamin Franzke1-7/+7
2012-03-31roi: Fail if invalid or no image is givenBenjamin Franzke1-0/+4
2012-03-31roi: Move rubberband bounds check into do_roiBenjamin Franzke1-30/+31
2012-03-31roi: Use GDK_BUTTON_PRIMARY in button_releaseBenjamin Franzke1-4/+6
2012-03-31roi: Trivial reorder in draw_rubberbandBenjamin Franzke1-1/+2
2012-03-31roi: The configure-event is not yet neededBenjamin Franzke1-8/+0
2012-03-31roi: Remove key_press state from roi structBenjamin Franzke1-12/+6
2012-03-31roi: Make all functions staticBenjamin Franzke1-8/+8
2012-03-31roi: Schedule rubberband drawing to draw-handlerBenjamin Franzke1-79/+68
So that we get automatic double buffer handling.
2012-03-31roi: Use gtk_render_{frame,background} for rubberbandBenjamin Franzke1-23/+20
2012-03-31roi: Render only if selected width and height != 0Benjamin Franzke1-1/+3
2012-03-31roi: Use gtk3Benjamin Franzke2-13/+11
2012-03-30roi: Move to a gtk based selecton-interfaceBenjamin Franzke2-45/+254
2012-03-30Add simple roi-algorithmBenjamin Franzke2-2/+103
2012-03-24wimmel_gl: Use GL_BGRA_EXT in readpixelsBenjamin Franzke2-5/+21
That is the native format graphics cards use, so we reach a fastpath for the copy, which saves around 1s.
2012-03-21wimmel_gl: Use luminance texturesBenjamin Franzke2-28/+33
2012-03-21commands: Use a smaller scan areaBenjamin Franzke1-1/+1
Simply to be faster.
2012-03-20wimmel: Convert to calculation on monochrome imageBenjamin Franzke1-3/+21
2012-03-15wimmel: Also use tiff as output formatBenjamin Franzke1-1/+1
2012-03-15wimmel_gl: Reduce newlines in gl framebuffer callsBenjamin Franzke1-7/+3
2012-03-15wimmel_gl: Output as tiff imageBenjamin Franzke1-1/+1
Saves 0.5 seconds.
2012-03-15wimmel_gl: Use a GLES2 contextBenjamin Franzke2-22/+79
Makes it easier to check that only non-deprecated stuff is used.
2012-03-15wimmel_gl: Use byte-alignment for readpixelsBenjamin Franzke1-2/+1
2012-03-15wimmel_gl: Allow compilation with -DGLUTBenjamin Franzke2-17/+20
2012-03-15Makefile: Pass in 2 threads if /proc/cpuinfo can not be readBenjamin Franzke1-1/+1
2012-03-15Makefile: Add g_threads cflags/libsBenjamin Franzke1-2/+2
2012-03-15wimmel_gl: Use a second texture unit for search regionBenjamin Franzke2-8/+43
Speeds up calulation from ~11s to ~10s.
2012-03-15wimmel.frag: Rename search_rect to srectBenjamin Franzke2-8/+7
2012-03-15wimmel_gl: Use EGL instead of GLUT by defaultBenjamin Franzke2-28/+55
That way we can create a context and make it current without creating a window.
2012-03-14Revert "wimmel.frag: Discard fragment early if difference is to significant"Benjamin Franzke1-6/+3
This reverts commit 5bcbaab8ea90655c2496c8df029a9cb991dba0ce. This was 0.2seconds faster on r600g but 4 seconds slower on sandybridge.
2012-03-14wimmel.frag: Discard fragment early if difference is to significantBenjamin Franzke1-3/+6
2012-03-14wimmel_gl: Remove unneeded fps calculationBenjamin Franzke1-10/+0
2012-03-14wimmel_gl: Draw rects around matching subimagesBenjamin Franzke2-17/+26
All red dots in the GL rendering represent a top-left corner of a matching rectangle.
2012-03-14wimmel: Render to a fboBenjamin Franzke2-28/+54
2012-03-14Add wimmel_gl prototypeBenjamin Franzke3-0/+280
Uses a shader to search for matching subimages.
2012-03-14Makefile: No dependency on gtk yetBenjamin Franzke3-4/+4
2012-03-14Makefile: Pass N_THREADS define as number of cpusBenjamin Franzke1-2/+5
2012-03-14wimmel: Dont use get_pixel for permormance reasonsBenjamin Franzke1-23/+18
2012-03-14wimmel: Simplify thread seperationBenjamin Franzke1-29/+25
2012-03-14util: Be faster by disabling asserts and alphaBenjamin Franzke1-3/+11
2012-03-14Add initial wimmel programBenjamin Franzke5-0/+377