ATI HD4850 error

[hieveryone]

I’ve tried the public release today but I can only record. If I try to reconstruct the scanned object or if I try to use realtime I always get the same error.
My system:
intel core2 quad Q6...@2.4Ghz
8 GB ram DDR3
ATi radeon HD4850

Here’s the log:
[2012/02/29 00:23:13.156250][info ] Welcome to ReconstructMe
[2012/02/29 00:23:13.156250][info ] Running 0.4.0-143
[2012/02/29 00:23:13.156250][info ]
[2012/02/29 00:23:13.171875][info ] Licensed for non-commercial purposes only.
[2012/02/29 00:23:13.171875][info ] See ‘License of ReconstructMe.txt’
[2012/02/29 00:23:13.187500][info ]
[2012/02/29 00:23:14.890625][info ] Preparing environment for Device ATI RV770
[2012/02/29 00:23:15.000000][error ] THROWING EXCEPTION: Throw in function (unknown)
Dynamic exception type: class LibOpenCLBridge::opencl_error
std::exception::what: Device: ATI RV770
C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl(737): error: write to < 32 bits via
pointer not allowed unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl(774): error: write to < 32 bits via
pointer not allowed unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl".

Internal error: clc compiler invocation failed.

[struct LibOpenCLBridge::opencl_error_code *] = -11
[struct LibOpenCLBridge::opencl_error_name *] = CL_BUILD_PROGRAM_FAILURE
[struct LibFundament::tag_robvis_message *] = Device: ATI RV770
C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl(737): error: write to < 32 bits via
pointer not allowed unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl(774): error: write to < 32 bits via
pointer not allowed unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl".

Internal error: clc compiler invocation failed.

[GwebMa]

Hi, nice to meet you here :-)

Just in case your Radeon is a Mobility version (as used in notebooks and AllinOne machines) try these drivers:

http://www2.ati.com/DRIVERS/mobile/11-12_mobility_vista_win7_32_dd_cc…

http://www2.ati.com/DRIVERS/mobile/11-12_mobility_vista_win7_64_dd_cc…

Am 29.02.2012 um 11:27 schrieb hieveryone:

ATi radeon HD4850

[hieveryone]

Hi MagWeb, good to see you here as well. ;)
Thanks for the suggestion, unfortunately my graphic card is not a mobility version.

[Christoph Heindl]

Am Mittwoch, 29. Februar 2012 11:27:21 UTC+1 schrieb hieveryone:

Here’s the log:

C:\DOCUME~2\X\IMPOST~1\Temp\OCL6FE.tmp.cl(737): error: write to < 32 bits via
pointer not allowed unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;

That has already been reported and we are working on a fix.

Thanks,

Christoph

[Christoph Heindl]

Could you try this version

http://reconstructme.wordpress.profactor.at/?wpdmact=process&did=MzUuaG90bGluaw==

Haven't made it public yet, since i don't know if it fixes all of your problems :)

Best,

Christoph

[hieveryone]

Thanks for the quick response.
Still no luck, realtime and reconstruction from files still gives me a similar error; I'm attaching the log.
A strange thing that occurred this time is that I also got the vpu recover error but I didn't even pressed 'p', just opened the recording window.

[Christoph Heindl]

Hi, could you attach the log file that can be found inside the logs directory. It contains more details about your GPU. Thanks.

[Christoph Heindl]

From this:

http://www.ibiblio.org/e-notes/webcl/rep4870.html 

I can see that your GPU does not support the extensions..this means we are probably out of luck :( Do you have access to any other device, CPU?

Am Mittwoch, 29. Februar 2012 16:29:11 UTC+1 schrieb hieveryone:

A strange thing that occurred this time is that I also got the vpu recover error but I didn't even pressed 'p', just opened the recording window.

vpu recover? 

[hieveryone]

Il giorno mercoledì 29 febbraio 2012 17:28:29 UTC+1, Christoph Heindl ha scritto:

I can see that your GPU does not support the extensions..this means we are probably out of luck :( Do you have access to any other device, CPU?

 

Er, I'm not sure I understood you question, do you mean another kinect or something similar? No.

 

vpu recover? 

Yeah, "VPU Recover has reset your graphics accelerator as it was no longer responding to graphics driver commands"; apparently it's a common problem with my graphics card model, still it's the first time I got this warning.

 

Hi, could you attach the log file that can be found inside the logs directory. It contains more details about your GPU. Thanks.

 

Actually the one that I attached was the log from the logs folder, did I missed something? If that's the case I will do another scan later and post all the logs generated inside the folder, just to be sure.

[Martin Ankerl]

>> I can see that your GPU does not support the extensions..this means we are
>> probably out of luck :( Do you have access to any other device, CPU?
>
>
> Er, I'm not sure I understood you question, do you mean another kinect or
> something similar? No.

Usually ReconstructMe uses the graphic card to do all its calculation.
But when this fails (e.g. with an old graphics card), it might still
be possible to use the CPU to calculate everything. We call these
processing units devices. You might be able to install an OpenCL
driver (that's the framework we are using to do the calculations on
either the GPU or the CPU) for your CPU, which is usually very slow,
but might be more stable. In the FAQ is a link to Intel OpenCL driver
that uses the CPU:

http://reconstructme.net/faq/#devicenotshowup

--
Martin Ankerl | http://martin.ankerl.com

[hieveryone]

Oh, I see. I'm downloading the drivers for the cpu, I will try them later and I'll make you know if that works.
I'm attaching all the logs created inside of the logs folder.

[Joel Clemens]

I also have an ATI radeon HD 4850, also get the same error. I tried
recording the video (works) and post processing. using
reconstructme.exe --replay --depths reconstructme.depths.gz --colors
reconstructme.avi

[2012/02/29 22:20:50.045898][info ] Welcome to ReconstructMe
[2012/02/29 22:20:50.050781][info ] Running 0.4.0-153
[2012/02/29 22:20:50.053710][info ]
[2012/02/29 22:20:50.056640][info ] Licensed for non-commercial
purposes only

[2012/02/29 22:20:50.061523][info ] See 'License of
ReconstructMe.txt'
[2012/02/29 22:20:50.066406][info ]
Do you agree to our license terms? (y/n) : y
[2012/02/29 22:20:52.131835][info ] Preparing environment for Device
ATI RV77

[2012/02/29 22:20:52.189453][error ] THROWING EXCEPTION: Throw in
function (un
nown)

Dynamic exception type: class LibOpenCLBridge::opencl_error
std::exception::what: Device: ATI RV770

G:\temp\OCL462F.tmp.cl(737): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

G:\temp\OCL462F.tmp.cl(774): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "G:\temp\OCL462F.tmp.cl".

Internal error: clc compiler invocation failed.


[struct LibOpenCLBridge::opencl_error_code *] = -11
[struct LibOpenCLBridge::opencl_error_name *] =
CL_BUILD_PROGRAM_FAILURE
[struct LibFundament::tag_robvis_message *] = Device: ATI RV770

G:\temp\OCL462F.tmp.cl(737): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

G:\temp\OCL462F.tmp.cl(774): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "G:\temp\OCL462F.tmp.cl".

Internal error: clc compiler invocation failed.

[2012/02/29 22:20:52.251953][error ] This is somewhat embarrassing
[2012/02/29 22:20:52.254882][error ] ..\..\..\source\robvis
\lib_openclbridge\s
c\openclbridge\project.cpp(185): Throw in function class cl::Program
__thiscall
LibOpenCLBridge::project::compile(const class cl::Context &)
Dynamic exception type: class
boost::exception_detail::clone_impl<class LibOpen

LBridge::opencl_error>
std::exception::what: Device: ATI RV770

G:\temp\OCL462F.tmp.cl(737): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

G:\temp\OCL462F.tmp.cl(774): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "G:\temp\OCL462F.tmp.cl".

Internal error: clc compiler invocation failed.


[struct LibOpenCLBridge::opencl_error_code *] = -11
[struct LibOpenCLBridge::opencl_error_name *] =
CL_BUILD_PROGRAM_FAILURE
[struct LibFundament::tag_robvis_message *] = Device: ATI RV770

G:\temp\OCL462F.tmp.cl(737): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
depths[gid] = v;
^

G:\temp\OCL462F.tmp.cl(774): error: write to < 32 bits via pointer not

allowed
unless cl_khr_byte_addressable_store is enabled
dest[gid] = convert_ushort_sat(sum_weighted_depth / sum_weight);
^

2 errors detected in the compilation of "G:\temp\OCL462F.tmp.cl".

[Christoph Heindl]

Am Mittwoch, 29. Februar 2012 19:00:43 UTC+1 schrieb hieveryone:

Oh, I see. I'm downloading the drivers for the cpu, I will try them later and I'll make you know if that works.
I'm attaching all the logs created inside of the logs folder.

Thanks for the logs, I've seen that we have dropped verbose output in the logs, will correct this today. 

[hieveryone]

I downloaded the Intel OpenCL sdk but apparently it doesn't support my cpu, hence it doesn't even install. Just my luck.

[Christoph Heindl]

I have an idea how to prevent that from happen, but it requires more thinking...stay tuned (for another week or so :)

[hieveryone]

I certainly will. Thanks. ;)

[fyrebug]

Just wondering if there's an update to this. my ati 4870 has the same issue as it's the same gpu core.

I'd hate to run off and buy a 6850 just to find out it's something you had a fix for in a week or 2. the 4850 otherwise works fine for me, skyrim, portal2, 3ds max, etc, it's just here that I have issues, but for the ability to have a 3d scanner, it would be worth upgrading... just want to see if I should.. or if I should be patient.

thanks

[Christoph Heindl]

Am Mittwoch, 7. März 2012 05:23:26 UTC+1 schrieb fyrebug:

Just wondering if there's an update to this. my ati 4870 has the same issue as it's the same gpu core.

No update yet.

 

I'd hate to run off and buy a 6850 just to find out it's something you had a fix for in a week or 2. the 4850 otherwise works fine for me, skyrim, portal2, 3ds max, etc, it's just here that I have issues, but for the ability to have a 3d scanner, it would be worth upgrading... just want to see if I should.. or if I should be patient.

Sady, I cannot guarantee anything. I can see what's causing the error, it's just we don't have a quick fix yet. All the fixes we have discussed are more or less hacks that could hit us again at a later point in time.

In case there is a chance, be patient...We might have a great idea in a week or two... It's not that we have forgotten about this problem. But we are a tiny time working hard to provide the best for everyone. This does not always work out...

Best,

Christoph

[fyrebug]

thanks for the reply. I'll play more with the cpu based CL, but I'm having it fail quite a bit, I think it just losses tracking and dies. I'll see how the next little while goes and then possibly buy something. in the least I guess I'd get a 3rd monitor with the 6850..

you guys have created some really amazing software, thank you. I understand you can't support everything so I'm not worried about that, I just didn't want to spend the money if I might not need to. the 4870 is on AMD's list of supported openCL devices, so I thought there might be a chnace. it's still pretty hefty by todays terms even though it's old..

[Christoph Heindl]

Am Mittwoch, 7. März 2012 15:34:16 UTC+1 schrieb fyrebug:

you guys have created some really amazing software, thank you. I understand you can't support everything so I'm not worried about that, I just didn't want to spend the money if I might not need to. the 4870 is on AMD's list of supported openCL devices, so I thought there might be a chnace. it's still pretty hefty by todays terms even though it's old..

I will do my best to fix this issue. Frankly, next week is full of other appointments ...

[Christoph Heindl]

Am Mittwoch, 7. März 2012 15:34:16 UTC+1 schrieb fyrebug:

thanks for the reply. I'll play more with the cpu based CL, but I'm having it fail quite a bit, I think it just losses tracking and dies. I'll see how the next little while goes and then possibly buy something. in the least I guess I'd get a 3rd monitor with the 6850..

you will be better off with record and replay in case you use the cpu since you won't miss a frame.

[Christoph Heindl]

 I think i have some good news for both you. I have just commited a fix that should resolve this issue (had a great idea this afternoon ...). However I'm not sure if all occurrences are resolved, since i cannot compile for your specific cards. Are you interested in testing the software?

Best,

Christoph

[Christoph Heindl]

I think i have resolved your issue. In case you are interested in testing, please let me know. 

[fyrebug]

100% interested! WOOHOO!
I just printed off a list of 6850/6870 cards and was going to go to the store at lunch. ha! great timing!
I'm DYING to use this software.
I won't be able to test though for another 5-6 hours when I get home.
can you post a link?

On Wednesday, March 7, 2012 12:11:01 PM UTC-5, Christoph Heindl wrote:

I think i have resolved your issue. In case you are interested in testing, please let me know. 

[Christoph Heindl]

Am Mittwoch, 7. März 2012 19:04:34 UTC+1 schrieb fyrebug:

100% interested! WOOHOO!
I just printed off a list of 6850/6870 cards and was going to go to the store at lunch. ha! great timing!
I'm DYING to use this software.
I won't be able to test though for another 5-6 hours when I get home.
can you post a link?

Yes, here it is

http://reconstructme.wordpress.profactor.at/?wpdmact=process&did=NDAuaG90bGluaw==

I cannot guarantee that everything will work, but it is a start :) Please report back ASAP.

Best,

Christoph

[fyrebug]

will do as soon as I get home.
is there anything special I should try? or just try realtime scan mode .bat file as normal?


(perhaps someone else will beat me to testing...)


thank you for trying!

[Christoph Heindl]

Am Mittwoch, 7. März 2012 19:53:45 UTC+1 schrieb fyrebug:

will do as soon as I get home.
is there anything special I should try? or just try realtime scan mode .bat file as normal?

nothing special, just reconstruct ;)

[FromSpain]

El miércoles 7 de marzo de 2012 20:04:28 UTC+1, Christoph Heindl escribió:
> Am Mittwoch, 7. März 2012 19:53:45 UTC+1 schrieb fyrebug:<blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">will do as soon as I get home.

> is there anything special I should try? or just try realtime scan mode .bat file as normal?

> </blockquote>
>
> </div>
> nothing special, just reconstruct ;)</div><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">
>
> </div></blockquote></blockquote></blockquote></blockquote></blockquote>

I´m also interested as my 4870 seems to have similar problems.
II have downloaded the V199 but it ask for a pass :(

[fyrebug]

it got furthur and started, but then crashed. and it seems to be looking for files under a user account hudson which must be one of the team members, is that why it's crashing now?

=========================
C:\ReconstructMe199\bin>reconstructme.exe --realtime
[2012/03/07 18:51:17.406351][info   ] Welcome to ReconstructMe
[2012/03/07 18:51:17.421951][info   ] Running 0.4.0-199
[2012/03/07 18:51:17.421951][info   ]
[2012/03/07 18:51:17.421951][info   ] Licensed for non-commercial purposes only.

[2012/03/07 18:51:17.437551][info   ] See 'License of ReconstructMe.txt'
[2012/03/07 18:51:17.437551][info   ]

Do you agree to our license terms? (y/n) : y

[2012/03/07 18:51:19.325155][info   ] Starting sensor
[2012/03/07 18:51:19.325155][info   ] Testing OpenNI driver backend
[2012/03/07 18:51:26.079967][info   ] Preparing environment for Device ATI RV770

[2012/03/07 18:51:32.647578][info   ] Press 'p' to reconstruct/pause recording
[2012/03/07 18:51:32.663178][info   ] Press 'r' to reset volume
[2012/03/07 18:51:32.663178][info   ] Press 'ESC' to stop recording
[2012/03/07 18:51:32.709978][error  ] THROWING EXCEPTION: Throw in function (unk

nown)
Dynamic exception type: class LibOpenCLBridge::opencl_error

std::exception::what: Failed to enqueue run kernel command.
[struct LibOpenCLBridge::opencl_error_code *] = -54
[struct LibOpenCLBridge::opencl_error_name *] = CL_INVALID_WORK_GROUP_SIZE
[struct LibOpenCLBridge::opencl_kernel_name *] = phong_shade_points
[struct LibFundament::tag_robvis_message *] = Failed to enqueue run kernel comma
nd.

e:\hudson\workspace\ras_reconstructme\Platforms\waldohreule.profactor.local\sour
ce\robvis\lib_openclbridge\inc\openclbridge/kernel.h(202): Throw in function cla
ss cl::Event __cdecl LibOpenCLBridge::kernel::start(class cl::CommandQueue &,cla
ss cl::Kernel &,const class cl::NDRange &,const class cl::NDRange &,const class
cl::NDRange &)
Dynamic exception type: class boost::exception_detail::clone_impl<class LibOpenC
LBridge::opencl_error>
std::exception::what: Failed to enqueue run kernel command.
[struct LibOpenCLBridge::opencl_error_code *] = -54
[struct LibOpenCLBridge::opencl_error_name *] = CL_INVALID_WORK_GROUP_SIZE
[struct LibOpenCLBridge::opencl_kernel_name *] = phong_shade_points
[struct LibFundament::tag_robvis_message *] = Failed to enqueue run kernel comma
nd.

[2012/03/07 18:51:33.692780][info   ]
[2012/03/07 18:51:33.692780][info   ] Thanks using ReconstructMe!
[2012/03/07 18:51:33.692780][info   ] Visit us http://www.reconstructme.net

C:\ReconstructMe199\bin>

[Christoph Heindl]

Am Donnerstag, 8. März 2012 00:54:56 UTC+1 schrieb fyrebug:

[struct LibOpenCLBridge::opencl_error_name *] = CL_INVALID_WORK_GROUP_SIZE
[struct LibOpenCLBridge::opencl_kernel_name *] = phong_shade_points

This indicates that your gpu doesn't offer our minimum workgroup size requirement. Could you open the log file (in the logs subdirectory) and paste its content from the last run here? 

[fyrebug]

here's the log. from what I read the 4800 series supported the draft spec of openCL but when the draft was finalized the workgroup size was changed.
I caved and bought a 6850 and that's running great, but if occasional 4870 tests will help I will gladly do it to help out. there's still lots of 4800 series cards out there becuase they run great. compared to the 6850 I just played deus ex and saw no graphics speed difference.
clearly openCL is a different beast, especially if the GPU can't buffer the data fast enough, I'm just saying I'm sure there's lots of people that might not want to upgrade their cards.
for me though, $129 is a very small price to pay to have a 3d scanner. it is going to be great friends with my makerbot printer!


[2012/03/08 19:37:05.904141][info   ] Welcome to ReconstructMe
[2012/03/08 19:37:05.904141][info   ] Running 0.4.0-199
[2012/03/08 19:37:05.919741][info   ]
[2012/03/08 19:37:05.919741][info   ] Licensed for non-commercial purposes only.
[2012/03/08 19:37:05.919741][info   ] See 'License of ReconstructMe.txt'
[2012/03/08 19:37:05.935341][info   ]
[2012/03/08 19:37:07.900944][info   ] Starting sensor
[2012/03/08 19:37:07.900944][info   ] Testing OpenNI driver backend
[2012/03/08 19:37:14.686956][info   ] Preparing environment for Device ATI RV770
[2012/03/08 19:37:14.686956][debug  ] Preparing environment for Device ATI RV770
  CL_DEVICE_VENDOR Advanced Micro Devices, Inc.
  CL_DEVICE_VERSION OpenCL 1.0 AMD-APP (851.4)
  CL_DEVICE_PROFILE FULL_PROFILE
  CL_DRIVER_VERSION CAL 1.4.1664
  CL_DEVICE_AVAILABLE 1
  CL_DEVICE_GLOBAL_MEM_CACHE_SIZE 0
  CL_DEVICE_GLOBAL_MEM_CACHE_TYPE 0
  CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE 0
  CL_DEVICE_GLOBAL_MEM_SIZE 268435456
  CL_DEVICE_LOCAL_MEM_SIZE 16384
  CL_DEVICE_LOCAL_MEM_TYPE 2
  CL_DEVICE_IMAGE_SUPPORT 0
  CL_DEVICE_MAX_CLOCK_FREQUENCY 750
  CL_DEVICE_MAX_COMPUTE_UNITS 10
  CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE 65536
  CL_DEVICE_MAX_PARAMETER_SIZE 1024
  CL_DEVICE_MAX_WORK_GROUP_SIZE 256
  CL_DEVICE_MAX_WORK_ITEM_SIZES [256, 256, 256]
  CL_DEVICE_EXTENSIONS cl_amd_fp64 cl_khr_gl_sharing cl_amd_device_attribute_query cl_khr_d3d10_sharing

[Christoph Heindl]

Am Freitag, 9. März 2012 06:20:58 UTC+1 schrieb fyrebug:

  CL_DEVICE_IMAGE_SUPPORT 0

That's the problem. Images are not supported, result however are drawn to images... So here you are really out of luck, can't help with that one :)

Great it works now with a new display adapter :) Are you happy with the results you get?

[fyrebug]

so it's not something you can get running? or soon anyways? if so I feel less guilty about buying the new card.

the results are great, it's a little hard looking over my shoulder to scan though so I may use VNC over my phone so I have a small portable screen attached to the kinect. I'm thnking wireless usb and the phone screen, or a usb 7" lcd would make this even better...
I saw the 12V cordless drill battery hack to power it, which was smart.

I'll post up some results soon. you guys are doing amazing stuff!

[Christoph Heindl]

Am Freitag, 9. März 2012 15:07:08 UTC+1 schrieb fyrebug:

so it's not something you can get running? or soon anyways? if so I feel less guilty about buying the new card.

no, cannot fix that soon :) I think this is a so called won't fix.

I'll post up some results soon. you guys are doing amazing stuff!

Great! Looking forward to it.

 

[fyrebug]

well then. I don't feel guilty about buying the new card and I guess I'll enjoy the ability to scan AND have a 3rd monitor!

I feel like I'm living in the future. ;)

[Joel Clemens]

Thanks for all your hard work Christoph, I also have a 4850 and have the exact same log file fyrebug, I hope there will be a more specific list of compatible display adapters in the future.

If you say that the 4850 is simply not going to work with openCL, then I guess we can drop it and get an upgrade like fyrebug did.

[hieveryone]

Well, I tried the latest build and it worked. I will do more tests but so far it seems you solved the problem with my gpu. Thanks.

[Christoph Heindl]

Am Samstag, 17. März 2012 00:48:21 UTC+1 schrieb hieveryone:

Well, I tried the latest build and it worked. I will do more tests but so far it seems you solved the problem with my gpu. Thanks.

Hah! Great :) 

[hieveryone]

I've done some scans with realtime and reconstruction on XP 32bit and it works well.I still can't get some details like lips and eyes but all in all the result is quite fast. Only thing I noticed is that if I record for, say, a minute or longer the reconstruction stops before it finishes the process. Maybe it's due to memory limit?
I tried to scan on 7 64bit but I get various kind of errors depending on the command I use, is it normal? I can post the logs if it can be useful.

[Christoph Heindl]

Am Sonntag, 18. März 2012 21:59:38 UTC+1 schrieb hieveryone:

I've done some scans with realtime and reconstruction on XP 32bit and it works well.I still can't get some details like lips and eyes but all in all the result is quite fast. Only thing I noticed is that if I record for, say, a minute or longer the reconstruction stops before it finishes the process. Maybe it's due to memory limit?

I don't think so. It might have lost .. ? I will have a look.

 

I tried to scan on 7 64bit but I get various kind of errors depending on the command I use, is it normal? I can post the logs if it can be useful.

Yes please post the details (maybe in a new thread). 

[hieveryone]

I used version 231 and it seems that highres reconstruction now works. Thanks.