badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 Opencl is working fine, Screen shot in a moment, I stated this on 1st post -- opencl benchmark multi-sampling on and off - Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 easy way to do it. http://sourceforge.n.../maciasl/files/ add this link to your rep od MACIASL PJALM Intel7 http://repo.pjalm.info/intel7 Credits for patch goes to user PJALM We do not support people who steal others people's work... credits... Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 Ok Rampage dev, going to do it your way as dsdt seems the best way to go... I am looking for pci0 in my dsdt, cant find it lol... any ideas ? dsdt.aml.zip Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 Ok Rampage dev, going to do it your way as dsdt seems the best way to go... I am looking for pci0 in my dsdt, cant find it lol... any ideas ? Also your openCL is almost half of what the base score for you card should be. Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 actually looks like my card is in PEG0 which is in pci0 Ok Lesson 1, dont use small case letters in dsdt editor, lol Please ignore above post Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 Yep, so just expand the list and look down until you see it or just search for it within the app. Command F Link to comment Share on other sites More sharing options...
fau7i Posted May 12, 2013 Share Posted May 12, 2013 Interesting discussion...I managed to boot again with generic ATI framebuffer. I found no obvious difference compare to when using specific framebuffer (Futomaki in my case). Except a very little dropped score in OpenCL benchmark (520 fps > 508 fps). And it seems badaxe2's score is reasonable. Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 Look: Link to comment Share on other sites More sharing options...
iFIRE Posted May 12, 2013 Share Posted May 12, 2013 a Fake? Link to comment Share on other sites More sharing options...
k3nny Posted May 12, 2013 Share Posted May 12, 2013 (edited) Rampage Dev, it would be kind of you to stop shifting the topic. I am stating different facts that you seem to simply ignore. The OpenCL score is perfectly fine if you look at the different benchmarks in the according thread. A 7970 reaches a score of ~700. A score of ~660 is good for a 7950. Let's assume no framebuffer is loaded. The fact is that everything works right now, regardless of the custom framebuffer injection. Until now, badaxe2 did not report any single issue worth mentioning except the anti aliasing bug which is not related to the framebuffer in any way. What I want to see, is some actual proof that injecting a framebuffer will give an advantage in this case. EDIT: The highest OpenCL score is a fake. It was discussed in the according thread. Look at the scores of other cards with a similar performance, it is simply not maintainable! EDIT2: Also there are big fluctuations when running this benchmark. Try it for yourself and run it several times in row. You will get a difference of 20 points for sure, maybe even 40 in the worst case. EDIT3: OpenGL4 won't work as Apple does not fully support it. Edited May 12, 2013 by k3nny Link to comment Share on other sites More sharing options...
fau7i Posted May 12, 2013 Share Posted May 12, 2013 Yeah, I already know that result which is so inspiring. Sure, 7970 is far better card, and I just can't afford Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 @rampage I get a compile problem when pasting in your code, ?? without your code i get no errors Link to comment Share on other sites More sharing options...
k3nny Posted May 12, 2013 Share Posted May 12, 2013 It is hard to tell what's wrong just by looking at the picture. Upload your modified DSDT, someone will fix it for sure. Link to comment Share on other sites More sharing options...
iFIRE Posted May 12, 2013 Share Posted May 12, 2013 delete device HDAU Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 @rampage I get a compile problem when pasting in your code, ?? without your code i get no errors Your pasting over extra code. Rampage Dev, it would be kind of you to stop shifting the topic. I am stating different facts that you seem to simply ignore. The OpenCL score is perfectly fine if you look at the different benchmarks in the according thread. A 7970 reaches a score of ~700. A score of ~660 is good for a 7950. Let's assume no framebuffer is loaded. The fact is that everything works right now, regardless of the custom framebuffer injection. Until now, badaxe2 did not report any single issue worth mentioning except the anti aliasing bug which is not related to the framebuffer in any way. What I want to see, is some actual proof that injecting a framebuffer will give an advantage in this case. EDIT: The highest OpenCL score is a fake. It was discussed in the according thread. Look at the scores of other cards with a similar performance, it is simply not maintainable! EDIT2: Also there are big fluctuations when running this benchmark. Try it for yourself and run it several times in row. You will get a difference of 20 points for sure, maybe even 40 in the worst case. EDIT3: OpenGL4 won't work as Apple does not fully support it. I am not ignoring anything. You just do not understand what a framebuffer does... Also what does luxmark give you? Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 delete device HDAU His video on youtube says to copy hdau ? Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 His video on youtube says to copy hdau ? Just follow as instructed. Link to comment Share on other sites More sharing options...
k3nny Posted May 12, 2013 Share Posted May 12, 2013 Let's assume I do not know what a framebuffer does. Does it change anything badaxe2, fau7i reported or does it prove your point of need for framebuffer injection? No. Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 Just follow as instructed. Im following as instructed and I get compile error ??? Can you try - AMD 7950 DSDT Injection.rtf.zip dsdt.aml.zip Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 Let's assume I do not know what a framebuffer does. Does it change anything badaxe2, fau7i reported or does it prove your point of need for framebuffer injection? No. http://rampagedev.wo...er-personality/ Framebuffer Personality Structure 1. Using Personality Gibba we have the following: Personality: Gibba ConnectorInfo count in decimal: 5 Disk offset in decimal 179984 0000000 00 04 00 00 04 03 00 00 00 01 00 00 12 04 05 01 0000010 00 04 00 00 04 03 00 00 00 01 00 00 22 05 04 02 0000020 00 08 00 00 04 02 00 00 00 01 00 00 11 02 06 04 0000030 00 02 00 00 14 02 00 00 00 01 00 00 00 00 03 05 0000040 04 00 00 00 04 02 00 00 00 01 00 00 11 02 01 03 0000050 2. Taking Port 0000000 we have the following: 00 04 00 00 04 03 00 00 00 01 00 00 12 04 05 01 3. Break it down into four equal parts: 00 04 00 00 04 03 00 00 00 01 00 00 12 04 05 01 4. Then take last two lines and break it down into two equal parts respectively: 00 04 00 00 04 03 00 00 00 01 00 00 12 04 05 01 5. Then once more take last two lines and break it down into two equal parts respectively: 00 04 00 00 04 03 00 00 00 01 00 00 12 04 05 01 6. The following is what each part in step 5 is: Connector Type* 00 04 00 00 ATY,ControlFlags** 04 03 00 00 Features** 00 01 Unknown* 00 00 Transmitter* 12 Encoder* 04 Hotplug ID* 05 Sense ID** 01 Without a FB the card can not and will not function as designed. Also not all the ports may not work with out a proper FB loaded. OpenCL will not work without on and I think that clover is at play in this case and is forcing something to load but the IOreg is not showing it. Why would you start troubleshooting before you had the card working as the designers intended. Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 Let's assume I do not know what a framebuffer does. Does it change anything badaxe2, fau7i reported or does it prove your point of need for framebuffer injection? No. I dont know if it changes anything, would be nice to test though.. Also good to have gpu name displayed correctly Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 I dont know if it changes anything, would be nice to test though.. Also good to have gpu name displayed correctly Here DSDT.aml.zip Link to comment Share on other sites More sharing options...
fau7i Posted May 12, 2013 Share Posted May 12, 2013 Okay, here is Luxmark score with generic framebuffer. Clover might plays its role, but to boot back with that framebuffer I have to use GraphicsInjector=No. Link to comment Share on other sites More sharing options...
badaxe2 Posted May 12, 2013 Author Share Posted May 12, 2013 Thans Rampage, i see its not quite the same as the txt file though... gonna reboot and fresh install, will let you know how it goes later Link to comment Share on other sites More sharing options...
Rampage Dev Posted May 12, 2013 Share Posted May 12, 2013 My test from back in December with a 7970. Link to comment Share on other sites More sharing options...
Recommended Posts