Hi!

Testing version 2.6.4 on Ubuntu 12.04, found a (minor) bug, but I think it could be an older one. Yesterday I configured my system to use the onboard GPU (3300) for display and the PCIe card (5770) for mining only. Unfortunately, ADL and CL GPU mapping does not seem to work.

Even when using gpu-map, I cannot see the GPU temp, fan speed etccorrectly. Cgminer gets the values from onboard card and displays them for PCIe card. If I configure gpu-map, the initial *setting* of GPUfreq, voltage etc seems to work,but I still cannot see the measurements. And of course, auto adjustments does not work.
Debug log shows that cgminer sets (or tries to set) the initial values for *both* cards.

Here are my configurations and log files:
cgminer.conf (partially):
-------------
"device" : "0",
"gpu-memclock" : "450",
"gpu-engine" : "600-900",
"gpu-vddc" : "1.025",
"gpu-fan" : "80",
"gpu-dyninterval" : "21",
"failover-only" : true,
"intensity" : "12",
"gpu-map" : "0:1"
-------------

Corresponding debug log:
--------------------------
[2012-08-09 11:49:09] Started cgminer 2.6.4
 [2012-08-09 11:49:09] Loaded configuration filehome/xxxx/.cgminer/cgminer.conf
 [2012-08-09 11:49:09] CL Platform 0 vendor: Advanced Micro Devices, Inc.
 [2012-08-09 11:49:09] CL Platform 0 name: AMD Accelerated Parallel Processing
 [2012-08-09 11:49:09] CL Platform 0 version: OpenCL 1.2 AMD-APP (937.2)
 [2012-08-09 11:49:09] Platform 0 devices: 1
 [2012-08-09 11:49:09]    0   Juniper
 [2012-08-09 11:49:09] GPU 0 iAdapterIndex 0 strUDID 296:38420:4098:33613:4163 iBusNumber 1 iDeviceNumber 5 iFunctionNumber 0 iVendorID 40
98 strAdapterName  ATI Radeon HD 3300 Graphics
 [2012-08-09 11:49:09] GPU 1 iAdapterIndex 2 strUDID 512:26808:4098:9539:4098 iBusNumber 2 iDeviceNumber 0 iFunctionNumber 0 iVendorID 409
8 strAdapterName  ATI Radeon HD 5700 Series
 [2012-08-09 11:49:09] ADL found more devices than opencl!
 [2012-08-09 11:49:09] There is possibly at least one GPU that doesn't support OpenCL
 [2012-08-09 11:49:09] Use the gpu map feature to reliably map OpenCL to ADL
 [2012-08-09 11:49:09] Mapping OpenCL device 0 to ADL device 1
 [2012-08-09 11:49:09] WARNING: Number of OpenCL and ADL devices did not match!
 [2012-08-09 11:49:09] Hardware monitoring may NOT match up with devices!
 [2012-08-09 11:49:09] GPU 0 ATI Radeon HD 3300 Graphics hardware monitoring enabled
 [2012-08-09 11:49:09] Failed to ADL_Overdrive5_ODPerformanceLevels_Get
 [2012-08-09 11:49:09] Setting GPU 0 engine clock to 900
 [2012-08-09 11:49:09] Setting GPU 0 memory clock to 450
 [2012-08-09 11:49:09] Setting GPU 0 voltage to 1.025
 [2012-08-09 11:49:09] Failed to ADL_Overdrive5_FanSpeedInfo_Get
 [2012-08-09 11:49:09] GPU 0 doesn't support rpm or percent write
 [2012-08-09 11:49:09] GPU 1 ATI Radeon HD 5700 Series hardware monitoring enabled
 [2012-08-09 11:49:09] Setting GPU 1 engine clock to 900
 [2012-08-09 11:49:09] Setting GPU 1 memory clock to 450
 [2012-08-09 11:49:09] Setting GPU 1 voltage to 1.025
------------------------
:::These values are from the on-board card (temp,voltage and rpm not valid), should be PCIe GPU:
 [2012-08-09 11:49:54] [thread 0: 805306368 hashes, 92701.7 khash/sec]
 [2012-08-09 11:49:54] -1.0 C  F: -1%(-1RPM)  E: 497MHz  M: 667Mhz  V: 0.000V  A: 7%  P: 0%
 [2012-08-09 11:49:55] [thread 1: 536870912 hashes, 92702.8 khash/sec]
 [2012-08-09 11:49:55] (5s):203.9 (avg):197.0 Mh/s | Q:3  A:0  R:0  HW:0  E:0%  U:0.0/m
 [2012-08-09 11:49:57] -1.0 C  F: -1%(-1RPM)  E: 497MHz  M: 667Mhz  V: 0.000V  A: 7%  P: 0%
 [2012-08-09 11:50:00] [thread 0: 536870912 hashes, 92704.3 khash/sec]
 [2012-08-09 11:50:00] -1.0 C  F: -1%(-1RPM)  E: 497MHz  M: 667Mhz  V: 0.000V  A: 7%  P: 0%
 [2012-08-09 11:50:01] [thread 1: 536870912 hashes, 92708.9 khash/sec]
 [2012-08-09 11:50:01] (5s):197.1 (avg):195.7 Mh/s | Q:3  A:0  R:0  HW:0  E:0%  U:0.0/m
 [2012-08-09 11:50:03] -1.0 C  F: -1%(-1RPM)  E: 497MHz  M: 667Mhz  V: 0.000V  A: 7%  P: 0%
 [2012-08-09 11:50:06] [thread 0: 536870912 hashes, 92693.9 khash/sec]
 [2012-08-09 11:50:06] -1.0 C  F: -1%(-1RPM)  E: 497MHz  M: 667Mhz  V: 0.000V  A: 7%  P: 0%
 [2012-08-09 11:50:07] [thread 1: 536870912 hashes, 92697.1 khash/sec]
 [2012-08-09 11:50:07] (5s):192.8 (avg):194.7 Mh/s | Q:3  A:0  R:0  HW:0  E:0%  U:0.0/m

Otherwise 2.6.4 seems to work ok. One note thought; I think fan speed setting should have a tolerance for 'target' value also. Optimally, the algorihtm should find a stable speed for the fan RPM and keep it. Currently the temperature and fan speed varies during normal operation (auto-adjustments worked before I started to use onboard card also). Ie. It would be better to have 77 degrees and 61% speed all the time vs temp varying 74-76 and RPM varying 58-62.

If someone is interested, I can make a more detailed instructions on how to configure the system to use onboard card for display and PCIe GPU for mining.

Edit: Versions:
-Ubuntu 12.04
-Some 4 core AMD64 cpu
- Catalyst 12.6 Legacy (not the beta version, which did not support 3300..)
- AMD SDK 2.7
- cgminer 2.6.4

I do not seem to get all the power from 5770, but have not yet been able to optimize it. Actually, even window manager seems to affect the speed, ie Unity vs LXDE.

Get over it Luke.

Simple problems are going be solved in the same simple way. There's nothing novel or unique about the code you write. It's all straight forward and intuitive.

Neither of these two examples have the same, copied solution. They were both solved independently in similar ways as they are both soving the same problem. It's no suprise the timing is also similar as these issue are being found at the same time.

You are not a coding god. Anyone can come up with the same ideas as you.

Maybe Kano is just better than you then?

There were a number of bug reports posted here in this thread pointing to libcurl being the source of locking up. Shortly after that, the fix was made. It's not a giant leap to assume that, in fact, they solved the problem.

Aren't most of what people are arguing about code-wise GPL or similar open licenses? If so, copying is a non-issue and is to be applauded. It benefits the greater good.

A bit like taking some header files and removing the author/copywrite info?

Ah, an accrediting issue. As I know nothing about who did what I'll bow out.

At the end of the day it's just your paranoid (delusional?) opinon that they took your code, changed it then submitted it themselves.

Since all this code is GPL, would the easier solution for them not simply be submit your commit, which would involve almost zero effort, rather then re-arrange your code? What is to gain? What's their motive?

Isn't BFG a fork of cgminer?  So he came in, added a few lines, and then now HE'S complaining about stolen code?

Detraction and trolling me is their apparent motive.

Maybe also related to this: