hello,
walking through the examples of 10.glsl (0.92.3) i noticed that 08 and 09 stopped working for me, meaning the consume so much cpu that they hardly perform anymore (esp. 09)
i cant remember the exact time anymore but on the same machine they have been working without problems over half a year ago.
it seems what was hw-accelerated back then is not anymore.
any idea what could be the reason ?
i run ubuntu 9.10 card: Quadro FX Go1400 NVIDIA Driver Version: 185.18.36
it might have stopped working after upgrading to 9.10 but this is only a guess now.
otherwise i can see no performance issue
best
erich
hello,
what Gem version are you using?
can you try to compile the curent svn version?
are they working like they used to (exept for the CPU used)?
c
erich a écrit :
hello,
walking through the examples of 10.glsl (0.92.3) i noticed that 08 and 09 stopped working for me, meaning the consume so much cpu that they hardly perform anymore (esp. 09)
i cant remember the exact time anymore but on the same machine they have been working without problems over half a year ago.
it seems what was hw-accelerated back then is not anymore.
any idea what could be the reason ?
i run ubuntu 9.10 card: Quadro FX Go1400 NVIDIA Driver Version: 185.18.36
it might have stopped working after upgrading to 9.10 but this is only a guess now.
otherwise i can see no performance issue
best
erich
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
hi cyrille,
what Gem version are you using?
Pd version 0.42.5-extended-20100116
with
GEM: ver: 0.92.2 GEM: compiled: Jan 16 2010
can you try to compile the curent svn version?
yes but not instantly, i might need some time
are they working like they used to (exept for the CPU used)?
yes they work, except not accelerated with high cpu consumption
erich
c
erich a écrit :
hello,
walking through the examples of 10.glsl (0.92.3) i noticed that 08 and 09 stopped working for me, meaning the consume so much cpu that they hardly perform anymore (esp. 09)
i cant remember the exact time anymore but on the same machine they have been working without problems over half a year ago.
it seems what was hw-accelerated back then is not anymore.
any idea what could be the reason ?
i run ubuntu 9.10 card: Quadro FX Go1400 NVIDIA Driver Version: 185.18.36
it might have stopped working after upgrading to 9.10 but this is only a guess now.
otherwise i can see no performance issue
best
erich
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
hello,
the only explanation i can find is that this Gem version was not compile with the good openGL header. but i'm not a specialist.
i think (hope) that recompiling will solve your problem.
the Gem version provided by pd extended never worked for me.
Cyrille
erich a écrit :
hi cyrille,
what Gem version are you using?
Pd version 0.42.5-extended-20100116
with
GEM: ver: 0.92.2 GEM: compiled: Jan 16 2010
can you try to compile the curent svn version?
yes but not instantly, i might need some time
are they working like they used to (exept for the CPU used)?
yes they work, except not accelerated with high cpu consumption
erich
c
erich a écrit :
hello,
walking through the examples of 10.glsl (0.92.3) i noticed that 08 and 09 stopped working for me, meaning the consume so much cpu that they hardly perform anymore (esp. 09)
i cant remember the exact time anymore but on the same machine they have been working without problems over half a year ago.
it seems what was hw-accelerated back then is not anymore.
any idea what could be the reason ?
i run ubuntu 9.10 card: Quadro FX Go1400 NVIDIA Driver Version: 185.18.36
it might have stopped working after upgrading to 9.10 but this is only a guess now.
otherwise i can see no performance issue
best
erich
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
On Mon, 2010-04-12 at 13:24 +0200, cyrille henry wrote:
hello,
the only explanation i can find is that this Gem version was not compile with the good openGL header. but i'm not a specialist.
after compiling the latest without performance gain i thought the same and found out that i did not have the nvidia devs which i downloaded and then recompiled gem - without success.
i also checked the lib /usr/lib/xorg/modules/extensions/libglx.so -> libglx.so.185.18.36 this is ok.
is there some other thing i could/need to check ?
best
erich
On 2010-04-12 13:48, erich wrote:
On Mon, 2010-04-12 at 13:24 +0200, cyrille henry wrote:
hello,
the only explanation i can find is that this Gem version was not compile with the good openGL header. but i'm not a specialist.
the idea of openGL is, to allow for runtime (as opposed to compile time) use of hardware acceleration. so the headers should(!) not matter at all.
after compiling the latest without performance gain i thought the same and found out that i did not have the nvidia devs which i downloaded and then recompiled gem - without success.
i also checked the lib /usr/lib/xorg/modules/extensions/libglx.so -> libglx.so.185.18.36 this is ok.
is there some other thing i could/need to check ?
do you have "libglew" installed? if so, does it help to uninstall it (or disable it with "--without-GLEW")?
do you have an old Gem lying around that is known to have had hw acceleration in these patches? does it still have? (e.g. can we rule out that the problem is due to some driver upgrade?)
fga,sdr IOhannes
do you have "libglew" installed? if so, does it help to uninstall it (or disable it with "--without-GLEW")?
deinstalled and recompiled but now success.
do you have an old Gem lying around that is known to have had hw acceleration in these patches? does it still have? (e.g. can we rule out that the problem is due to some driver upgrade?)
no i dont have, but i am sure it was a pd-extended out of the box. also i belief i am pretty sure i have HW acceleration, its is only the examples 08 and 09 of 10.glsl which are affected. all my other stuff and also shaders in general are working fine.
erich
fga,sdr IOhannes
On Mon, 2010-04-12 at 16:05 +0300, erich wrote:
do you have "libglew" installed? if so, does it help to uninstall it (or disable it with "--without-GLEW")?
deinstalled and recompiled but now success.
ha - what a typo !
i meant : NO success
its still not working
i belief now it might be a driver problem. i have the relevant extension for the 08 and 09 example showing in my glxinfo.
i am pretty sure i have HW acceleration, its is only the examples 08 and 09 of 10.glsl which are affected. all my other stuff and also shaders in general are working fine.
any idea how to check this ?
best
erich
hello,
both examples 08 and 09 of 10.glsl work fine now when i send a |quality 0( message to the |pix_texture| on the |mesh_square| render chain before starting rendering.
the message is there in the original help patches, but not connected to a |loadbang|.
best
erich
On Mon, 2010-04-12 at 14:07 +0200, IOhannes m zmoelnig wrote:
On 2010-04-12 13:48, erich wrote:
On Mon, 2010-04-12 at 13:24 +0200, cyrille henry wrote:
hello,
the only explanation i can find is that this Gem version was not compile with the good openGL header. but i'm not a specialist.
the idea of openGL is, to allow for runtime (as opposed to compile time) use of hardware acceleration. so the headers should(!) not matter at all.
after compiling the latest without performance gain i thought the same and found out that i did not have the nvidia devs which i downloaded and then recompiled gem - without success.
i also checked the lib /usr/lib/xorg/modules/extensions/libglx.so -> libglx.so.185.18.36 this is ok.
is there some other thing i could/need to check ?
do you have "libglew" installed? if so, does it help to uninstall it (or disable it with "--without-GLEW")?
do you have an old Gem lying around that is known to have had hw acceleration in these patches? does it still have? (e.g. can we rule out that the problem is due to some driver upgrade?)
fga,sdr IOhannes
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
hello,
i changed the fetching.vert shader so that i would be able to access rgb individually for shifting the vertices, but the result does not suggest that.
is vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st) not a vector with RGBA of a single vertex ?
best
erich
here the shader:
uniform float cr, cg, cb; uniform sampler2D MyTex; varying vec4 C; void main() { vec4 v = vec4(gl_Vertex); vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st); v.z = color.r*cr + color.g*cg +color.b*cb; // v.x += (color.b-0.5)/2.; // v.y += (color.g-0.5)/2.;
C=color; gl_Position = gl_ModelViewProjectionMatrix * v; }
Hello Erich,
No problem here on MacOSX.5.8 with pd-ext 0.41.4 and GEM 0.91.3 'tigital'. My graphic card here : ATI Radeon HD 2600 Pro What is your graphic card ? Do you have errors in Pd console ?
[glsl_program]: uvar#0: "MyTex": [glsl_program]: unknown (0x8B5E) [glsl_program]: uvar#1: "cr": [glsl_program]: GL_FLOAT [glsl_program]: uvar#2: "cg": [glsl_program]: GL_FLOAT [glsl_program]: uvar#3: "cb": [glsl_program]: GL_FLOAT [glsl_program]: uvar#4: "gl_ModelViewProjectionMatrix": [glsl_program]: GL_FLOAT_MAT4_ARB
Here my patch where you need to change the messages cr, cg, cb. I use your vertex shader and the fragment shader with 'gl_FragColor = C;' ++
Jack
hello,
i changed the fetching.vert shader so that i would be able to access rgb individually for shifting the vertices, but the result does not suggest that.
is vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st) not a vector with RGBA of a single vertex ?
best
erich
here the shader:
uniform float cr, cg, cb; uniform sampler2D MyTex; varying vec4 C; void main() { vec4 v = vec4(gl_Vertex); vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st); v.z = color.r*cr + color.g*cg +color.b*cb; // v.x += (color.b-0.5)/2.; // v.y += (color.g-0.5)/2.;
C=color; gl_Position = gl_ModelViewProjectionMatrix * v;
}
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
hi jack,
sorry i should have been a bit more precise.
the shader works without errors and the patch u sent me is the same like mine.
my question refers to that it seems to me that the displacement is not dependent on color but on brightness.
best
erich
On Thu, 2010-04-29 at 17:46 +0200, jack@rybn.org wrote:
Hello Erich,
No problem here on MacOSX.5.8 with pd-ext 0.41.4 and GEM 0.91.3 'tigital'. My graphic card here : ATI Radeon HD 2600 Pro What is your graphic card ? Do you have errors in Pd console ?
[glsl_program]: uvar#1: "cr": [glsl_program]: GL_FLOAT [glsl_program]: uvar#2: "cg": [glsl_program]: GL_FLOAT [glsl_program]: uvar#3: "cb": [glsl_program]: GL_FLOAT [glsl_program]: uvar#4: "gl_ModelViewProjectionMatrix": [glsl_program]: GL_FLOAT_MAT4_ARB
Here my patch where you need to change the messages cr, cg, cb. I use your vertex shader and the fragment shader with 'gl_FragColor = C;' ++
Jack
hello,
i changed the fetching.vert shader so that i would be able to access rgb individually for shifting the vertices, but the result does not suggest that.
is vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st) not a vector with RGBA of a single vertex ?
best
erich
here the shader:
uniform float cr, cg, cb; uniform sampler2D MyTex; varying vec4 C; void main() { vec4 v = vec4(gl_Vertex); vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st); v.z = color.r*cr + color.g*cg +color.b*cb; // v.x += (color.b-0.5)/2.; // v.y += (color.g-0.5)/2.;
C=color; gl_Position = gl_ModelViewProjectionMatrix * v;
}
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
can you try to add a pix_rgba after pix_image? c
erich a écrit :
hi jack,
sorry i should have been a bit more precise.
the shader works without errors and the patch u sent me is the same like mine.
my question refers to that it seems to me that the displacement is not dependent on color but on brightness.
best
erich
On Thu, 2010-04-29 at 17:46 +0200, jack@rybn.org wrote:
Hello Erich,
No problem here on MacOSX.5.8 with pd-ext 0.41.4 and GEM 0.91.3 'tigital'. My graphic card here : ATI Radeon HD 2600 Pro What is your graphic card ? Do you have errors in Pd console ?
[glsl_program]: uvar#1: "cr": [glsl_program]: GL_FLOAT [glsl_program]: uvar#2: "cg": [glsl_program]: GL_FLOAT [glsl_program]: uvar#3: "cb": [glsl_program]: GL_FLOAT [glsl_program]: uvar#4: "gl_ModelViewProjectionMatrix": [glsl_program]: GL_FLOAT_MAT4_ARB
Here my patch where you need to change the messages cr, cg, cb. I use your vertex shader and the fragment shader with 'gl_FragColor = C;' ++
Jack
hello,
i changed the fetching.vert shader so that i would be able to access rgb individually for shifting the vertices, but the result does not suggest that.
is vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st) not a vector with RGBA of a single vertex ?
best
erich
here the shader:
uniform float cr, cg, cb; uniform sampler2D MyTex; varying vec4 C; void main() { vec4 v = vec4(gl_Vertex); vec4 color = texture2D(MyTex, gl_MultiTexCoord0.st); v.z = color.r*cr + color.g*cg +color.b*cb; // v.x += (color.b-0.5)/2.; // v.y += (color.g-0.5)/2.;
C=color; gl_Position = gl_ModelViewProjectionMatrix * v;
}
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
On 2010-04-29 20:07, cyrille henry wrote:
can you try to add a pix_rgba after pix_image?
actually this shouldn't matter.
whenever the image is uploaded as a texture (and you can only access textures from within the shader), it has (not so) magically turned into an RGBA/float image.
fgmadr IOhannes
IOhannes m zmoelnig a écrit :
On 2010-04-29 20:07, cyrille henry wrote:
can you try to add a pix_rgba after pix_image?
actually this shouldn't matter.
ok.
so, i have no idea why it could not work. c
whenever the image is uploaded as a texture (and you can only access textures from within the shader), it has (not so) magically turned into an RGBA/float image.
fgmadr IOhannes
GEM-dev mailing list GEM-dev@iem.at http://lists.puredata.info/listinfo/gem-dev
On 2010-04-29 19:56, erich wrote:
hi jack,
sorry i should have been a bit more precise.
the shader works without errors and the patch u sent me is the same like mine.
my question refers to that it seems to me that the displacement is not dependent on color but on brightness.
you probably should test with simpler images, where you can clearly distinguish between the colors.
fmgasdr IOhannes
-
cyrille henry -
erich -
IOhannes m zmoelnig -
jack＠rybn.org