Collatz 4.xx Windows OpenCL App for AMD 32 and 64 bit Released
log in

Advanced search

Message boards : News : Collatz 4.xx Windows OpenCL App for AMD 32 and 64 bit Released

Author Message
Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16190 - Posted: 5 Apr 2013, 14:42:22 UTC

Collatz 4.xx app for AMD OpenCL for both 32 and 64 bit Windows has been released. Please let me know of any problems with the new applications.

Corsair
Avatar
Send message
Joined: 16 Nov 09
Posts: 21
Credit: 824,334,256
RAC: 3,964,869
Message 16209 - Posted: 9 Apr 2013, 14:09:37 UTC - in response to Message 16190.
Last modified: 9 Apr 2013, 14:10:29 UTC

How could be downloaded this version?? if my computer is continuously downloaded 3.11 x64 openCL and all unit are giving error in the first second of run??

I'm using AMD HD777O & APU A10-5700 Cat. 13.1 with OpenGL 6.14.10.12002 in a computer with windows 8 pro x64.

if possible where could be downloaded and an app.xml applied for only crunch if previous version still fails in cat 13.1??
____________
over the sailor's grave doesn't grove grass.

Corsair

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16224 - Posted: 11 Apr 2013, 21:33:35 UTC - in response to Message 16209.

How could be downloaded this version?? if my computer is continuously downloaded 3.11 x64 openCL and all unit are giving error in the first second of run??

I'm using AMD HD777O & APU A10-5700 Cat. 13.1 with OpenGL 6.14.10.12002 in a computer with windows 8 pro x64.

if possible where could be downloaded and an app.xml applied for only crunch if previous version still fails in cat 13.1??


If you have an app_info.xml file, that is probably causing the problem. Rename it to app_info.xml.old or something and you should get the 4.xx app.

Pollux_P3D
Send message
Joined: 19 Oct 11
Posts: 6
Credit: 18,196,287
RAC: 0
Message 16226 - Posted: 12 Apr 2013, 6:34:03 UTC
Last modified: 12 Apr 2013, 6:54:05 UTC

Hi

Collatz 4.03 app for AMD OpenCL 64 bit double the time for the calculation!!!


Collatz Conjecture v3.11 x86_64 for OpenCL
AMD Radeon HD 7700 Series
OpenCL 1.2 AMD-APP (938.2)
2900 sec



Collatz Conjecture v4.03 x86_64 for OpenCL 1.0 or later
AMD Radeon HD 7700 Series
OpenCL 1.2 AMD-APP (938.2)
6700 sec

Pollux_P3D
Send message
Joined: 19 Oct 11
Posts: 6
Credit: 18,196,287
RAC: 0
Message 16227 - Posted: 12 Apr 2013, 8:20:31 UTC
Last modified: 12 Apr 2013, 8:23:21 UTC

Collatz Conjecture v4.02 x86_64
AMD Radeon HD 7700 series
driver: 1.4.1720
mini_collatz v4.03
23800 sec

Pollux_P3D
Send message
Joined: 19 Oct 11
Posts: 6
Credit: 18,196,287
RAC: 0
Message 16229 - Posted: 12 Apr 2013, 9:59:15 UTC - in response to Message 16227.

Collatz Conjecture v4.02 x86_64
AMD Radeon HD 7700 series
driver: 1.4.1720
mini_collatz v4.03
23800 sec



Sorry, an error was, indeed on the cpu

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16230 - Posted: 12 Apr 2013, 15:10:13 UTC

Slower depends upon your definition of slower. When run with the default settings, it is slower but only because the default settings are set so that video response is still decent with low end GPUs. Change the collatz.config to check more numbers per kernel and more kernels per reduction. Or, use an app_config.xml to run multiple WUs at a time. Either way, it is just as fast as the previous version.

What is the GPU load? (Use GPU-Z or similar) If < 98%, use an app_config.xml and use 0.5 GPUs (two at a time) or 0.33 GPUs (3 at a time) etc. until the load
is 98% or higher. For example:

<app_config>
<app>
<name>collatz</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.50</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_version>
</app>
</app_config>

Profile Peciak
Avatar
Send message
Joined: 4 Aug 09
Posts: 13
Credit: 669,955,177
RAC: 764
Message 16233 - Posted: 12 Apr 2013, 19:47:06 UTC

New application is tragic.
I'm going back to the old.

Luis R. Resto*
Send message
Joined: 5 Apr 13
Posts: 1
Credit: 1,042,665,875
RAC: 117,925
Message 16238 - Posted: 13 Apr 2013, 15:57:12 UTC

collatz 4.03 (opencl_amd) on 7970 causing computation errors.

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16242 - Posted: 14 Apr 2013, 13:50:37 UTC

Do you get upset with your automobile mechanic when he can't fix your car when your car is locked in your garage?

If you want to keep your computers hidden, please post the reasons you think it is tragic or the errors you are getting along with the driver version, operating system, which application (mini or collatz), whether you are using an app_info or app_config, etc.

Claggy
Send message
Joined: 27 Sep 09
Posts: 288
Credit: 14,320,498
RAC: 0
Message 16243 - Posted: 14 Apr 2013, 15:34:42 UTC

Urs Echternacht fould the reason the Setiathome AMD/ATI MB and AP apps were failing on Cat 13.2 and Cat 13.3 (APP runtime 1124.2) was because that APP runtime doesn't like the build option "-fno-bin-llvmir" from OpenCL compiler options,
removing that option when building the apps (or it might be when the apps build it's compilations) allowed the apps to work again, while i doubt it's the problem here, it's something to try:

http://devgurus.amd.com/message/1288791#1288791

Claggy

Ira Dei
Send message
Joined: 11 Jan 13
Posts: 1
Credit: 19,448,102
RAC: 0
Message 16408 - Posted: 30 Apr 2013, 15:41:59 UTC
Last modified: 30 Apr 2013, 15:45:36 UTC

I do not know why, but I get calculation errors all the time:
http://s7.directupload.net/file/d/3241/qyda3hqw_jpg.htm
('Berechnungsfehler' = error)

I own a FX-8350 and a VTX3D HD 7850 X-Edition on Catalyst 13.5. On earlier versions it was that way, too.

Juergen
Send message
Joined: 29 May 12
Posts: 4
Credit: 10,925,841
RAC: 39,085
Message 16429 - Posted: 5 May 2013, 14:47:58 UTC - in response to Message 16408.

I do not know why, but I get calculation errors all the time:
http://s7.directupload.net/file/d/3241/qyda3hqw_jpg.htm
('Berechnungsfehler' = error)

I own a FX-8350 and a VTX3D HD 7850 X-Edition on Catalyst 13.5. On earlier versions it was that way, too.



Same here with a 5870, catalyst 13.4 and w7 64

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16430 - Posted: 6 May 2013, 5:17:40 UTC - in response to Message 16243.

Urs Echternacht fould the reason the Setiathome AMD/ATI MB and AP apps were failing on Cat 13.2 and Cat 13.3 (APP runtime 1124.2) was because that APP runtime doesn't like the build option "-fno-bin-llvmir" from OpenCL compiler options,
removing that option when building the apps (or it might be when the apps build it's compilations) allowed the apps to work again, while i doubt it's the problem here, it's something to try:

http://devgurus.amd.com/message/1288791#1288791

Claggy


It uses no build options at all. Maybe it should. ;-)

When kernels are compiled, if compilation fails, there is supposed to be a build log that can be retrieved. In past AMD driver versions that worked fine. Starting with the 13.x drivers, it has a fatal crash during the compile which causes an unhandled exception. I've tried to add a try-catch but around it but that doesn't work.

There is something really screwed up in the AMD compiler. With the kernel which uses ulongs, if I comment out the mul_hi, it compiles OK. It won't run correctly, but it compiles OK. If I use uints instead of ulongs so that I can omit the mul_hi, it crashes at compile time as well.

Tony Stark
Send message
Joined: 30 Sep 11
Posts: 1
Credit: 43,654,400
RAC: 0
Message 16433 - Posted: 6 May 2013, 17:22:27 UTC

solo_collatz 4.05 (opencl_ati_100) computation errors

Radeon 5870
Radeon 6950
Catalyst 13.4 + Windows 7x64

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2525
Credit: 740,580,099
RAC: 1
Message 16434 - Posted: 7 May 2013, 4:15:12 UTC - in response to Message 16433.

solo_collatz 4.05 (opencl_ati_100) computation errors

Radeon 5870
Radeon 6950
Catalyst 13.4 + Windows 7x64



Pretty much anything after 12.10 causes errors on a large number of GPUs. On the other hand, I do plan on seeing Iron Man 3 in 3-D in the next week or two.

[AF>FAH-Addict.net]toTOW
Send message
Joined: 15 Oct 10
Posts: 15
Credit: 100,193,374
RAC: 0
Message 16435 - Posted: 7 May 2013, 19:38:31 UTC

Solo Collatz doesn't start here too :(

<core_client_version>7.0.64</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>
Collatz Conjecture v4.05 x86_64 for OpenCL 1.0 or later
Based on the AMD Brook+ kernels by Gipsel


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x000000007730E4B4 write attempt to address 0x00000024


This is the only OpenCL application I use that doesn't work on this system (HD 6970 + 7 64 bits + Catalyst 13.1) ... :(

[AF>FAH-Addict.net]toTOW
Send message
Joined: 15 Oct 10
Posts: 15
Credit: 100,193,374
RAC: 0
Message 16436 - Posted: 7 May 2013, 21:07:33 UTC
Last modified: 7 May 2013, 21:23:39 UTC

Same error after upgrading to Catalyst 13.4 ... :(

<message>
(unknown error) - exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>
Collatz Conjecture v4.03 x86_64 for OpenCL 1.0 or later
Based on the AMD Brook+ kernels by Gipsel
verbose=1
Device Vendor Advanced Micro Devices, Inc.
Name Cayman
Board Name AMD Radeon HD 6900 Series
Max Clock 840
Compute Units 24
Driver Version 1124.2 (VM)
OpenCL Version OpenCL 1.2 AMD-APP (1124.2)
Alignment 2048 bits
Workgroup Size Auto
Kernel bit shifting


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00000000770EE4B4 write attempt to address 0x00000024


Same error with Collatz v4.03 :(

<message>
(unknown error) - exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>
Collatz Conjecture v4.03 x86_64 for OpenCL 1.0 or later
Based on the AMD Brook+ kernels by Gipsel
verbose=1
Device Vendor Advanced Micro Devices, Inc.
Name Cayman
Board Name AMD Radeon HD 6900 Series
Max Clock 840
Compute Units 24
Driver Version 1124.2 (VM)
OpenCL Version OpenCL 1.2 AMD-APP (1124.2)
Alignment 2048 bits
Workgroup Size Auto
Kernel bit shifting


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00000000770EE4B4 write attempt to address 0x00000024


Post to thread

Message boards : News : Collatz 4.xx Windows OpenCL App for AMD 32 and 64 bit Released


Main page · Your account · Message boards


Copyright © 2018 Jon Sonntag; All rights reserved.