search

jmeier64 / open-hardware-monitor

Automatically exported from code.google.com/p/open-hardware-monitor
0 stars 0 forks source link

Crossfire Support and Closing Program #111

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
I am using 2 x ATI Radeon HD5770 graphics cards in Crossfire mode but only one 
is being reported. Is it possible to include details of both cards? Also it 
would be helpful if the program could have an option that allows the program to 
be minimised to tray when the Red X is clicked ( I keep doing this instead of 
minimising) but the program to be closed from a right click of the tray icon ?

Original issue reported on code.google.com by robjenki...@gtempaccount.com on 2 Sep 2010 at 8:41

GoogleCodeExporter commented 9 years ago 
Minimize on close done from r230 onwards, will be in the next release. 
Alternatively you can use the build I've attached. It is an alpha build though, 
just so you know.

As for the Crossfire stuff, there's a couple of things I need to discuss with 
my colleague, but it is being looked into :)

Original comment by paulwere...@gmail.com on 23 Sep 2010 at 8:56

Attachments:

GoogleCodeExporter commented 9 years ago 
Can you post an Open Hardware Monitor report (File / Save Report) of your 
system? Otherwise it will be very difficult to goes what is going wrong.

Original comment by moel.mich on 24 Sep 2010 at 1:21

GoogleCodeExporter commented 9 years ago 
I attach the file as requested

Original comment by robjenki...@gtempaccount.com on 24 Sep 2010 at 7:00

Attachments:

GoogleCodeExporter commented 9 years ago 
Thank you. The AdapterID is zero for the second card. The ADL (AMD Display 
Library) docs say about this: "Zero means: The adapter is not AMD.".

Could it be that the second card is powered off or sleeping, because its 
currently not in use? Can you stress both cards with somthing that supports 
Crossfire and then start the Open Hardware Monitor?

Original comment by moel.mich on 24 Sep 2010 at 7:39

GoogleCodeExporter commented 9 years ago 
I just checked and the Open Hardware Monitor can handle Crossfire on other 
systems (at least I know of one using two HD 4800 series cards). So I think we 
must make sure here that Crossfire is really running as it should on your 
system. For example if only one card is running and the second one is disabled 
/ powered off, then the output of the Open Hardware Monitor would be right.

Original comment by moel.mich on 24 Sep 2010 at 8:56

GoogleCodeExporter commented 9 years ago 
I have ran GPU-Z 0.4.6 and it reports that ATI Crossfire is enabled with 2 
GPUs. I'll see if I can find a way to stress both cards to see if that makes 
any difference.

Original comment by robjenki...@gtempaccount.com on 25 Sep 2010 at 12:09

GoogleCodeExporter commented 9 years ago 
I ran Furmark and checked OpenHardwareMonitor and it showed 2 GPUs - I attach 
the report generated. I closed OpenHardwareMonitor and reopened it and it then 
showed 1 GPU. It looks like OpenHardwareMonitor only reports the second GPU 
when it is active. It would be nice if your program could show both GPUs even 
if one is not active

Original comment by robjenki...@gtempaccount.com on 25 Sep 2010 at 2:58

Attachments:

GoogleCodeExporter commented 9 years ago 
We'll look into this, but I've got another question for you.

Are these two identical cards? Because there's an uneven number of adapters, 
while it should show an even number. Do both cards have the same number of 
outputs? Because from what I can tell, you're using one with 2 outputs 
(adapters 0 and 1) and one with 3 outputs (2, 3 and 4).

Doesn't really matter, but it's just something I'd like to know :)

Original comment by paulwere...@gmail.com on 25 Sep 2010 at 6:02

GoogleCodeExporter commented 9 years ago 
The cards are supposed to be the same - both are ATI Radeon HD5770. GPU-Z says 
the cards both have the same device ID (1002 - 68B8) but the subvendors are 
listed as Sapphire/PCPartner (1748) for one card and ATI(1002) for the other. 
Perhaps this means that the cards are not identical after all !

Original comment by robjenki...@gtempaccount.com on 25 Sep 2010 at 9:24

GoogleCodeExporter commented 9 years ago 
Well could you check the outputs on the back? Because I know there are versions 
of the 5770 with just 2 outputs (these might be low-profile ones, not sure), 
with 3 outputs (the way they came out a year ago) and with 4 outputs (later 
revision).

Did you buy them together? Both Sapphire?

Original comment by paulwere...@gmail.com on 27 Sep 2010 at 2:24

GoogleCodeExporter commented 9 years ago 
What does the Catalyst Control Center report for the temperature of your second 
card when it has been idle? If the temperature drops there to 0°C as well a 
few seconds after becoming idle, then it is a problem (or feature) of the 
driver.

Looking at the report you created after starting the Open Hardware Monitor 
while stressing both cards it seems that one of the card is switched off (at 
least in the driver) when it is idle again. That's why the driver (via ADL) 
reports a 0°C temperature in the Open Hardware Monitor.

Have you installed the latest Catalyst driver?

If this is really a feature of the driver, than I am not sure it is a good idea 
to try to keep the second card awake for monitoring only.

There have been problems with the drivers some time ago:

http://hardforum.com/showthread.php?t=1458683
http://forums.amd.com/game/messageview.cfm?FTVAR_FORUMVIEWTMP=Linear&catid=260&t
hreadid=120325

I am not sure if they are resolved now (I don't have a crossfire system to 
test).

Original comment by moel.mich on 27 Sep 2010 at 3:16

GoogleCodeExporter commented 9 years ago 
In response to Comment 10 both cards have 4 outputs - they were bought 1 week 
apart from the same supplier and both for the same price. Regarding comment 11 
I cannot find in CCC where the temperatures are reported. I looked under 
CrossFireX Diagnostics which is blank. CCC reports Driver Packaging Version as 
8.771-100825a-104557C-ATI, Catalyst Version 10.9 and Catalyst Control Centre 
Version 2010.0825.2146.37182

Original comment by robjenki...@gtempaccount.com on 27 Sep 2010 at 6:49

GoogleCodeExporter commented 9 years ago 
The temperature should be shown in the "ATI Overdrive" section (cf. 
http://www.legitreviews.com/images/reviews/1105/5750_stock.jpg)

Original comment by moel.mich on 27 Sep 2010 at 6:55

GoogleCodeExporter commented 9 years ago 
I have just installed version 0.2.1 beta but it seems that minimize on close is 
not working - clicking on the red X still closes the program.

Original comment by robjenki...@gtempaccount.com on 7 Oct 2010 at 8:41

GoogleCodeExporter commented 9 years ago 
You have to set 'Minimize On Close' in the Options menu.

Original comment by morgan...@gmail.com on 7 Oct 2010 at 9:17

GoogleCodeExporter commented 9 years ago 
Sorry - my silly mistake

Original comment by robjenki...@gtempaccount.com on 8 Oct 2010 at 10:31

GoogleCodeExporter commented 9 years ago 
The same one I made :p

Original comment by morgan...@gmail.com on 8 Oct 2010 at 12:06

GoogleCodeExporter commented 9 years ago 
If Windows starts the programm automaticly most times there is only one card 
detected, but if I start it, when both cards are working (for one furmark), 
both cards are detected. 

I have added two reports. One on Windows idle and one running Furmark. Maybe it 
helps.

Original comment by ha...@gmx.net on 14 Apr 2011 at 4:35

Attachments:

GoogleCodeExporter commented 9 years ago 
What happens if you start the Open Hardware Monitor with Furmark running (so 
that both cards are detected), then close Furmark and wait a bit for the system 
to become really idle again? Do any readings on the GPUs in Open Hardware 
Monitor drop to zero or disappear? What does the temperature and fan display in 
the Catalyst Control Center show for both GPUs after this test?

If any readings disappear when you are back to idle, can you test if the 
readings return when you stress both GPUs with Furmark again (or if the 
readings remain dead once a GPU has been turned off)?

I think the problem is caused by the fact that the driver powers down one of 
the GPUs when they are not in use. Theoretically it should be possible to wake 
the GPUs for detection, I am currently not sure if this is a good idea. That's 
why answers to the above questions would be interesting.

Original comment by moel.mich on 15 Apr 2011 at 2:38

GoogleCodeExporter commented 9 years ago 
You're rigth. One card is going to sleep, this is a feature of ati/amd.

A while after Furmark was running, the temperature drops to 0°C and the other 
values are standing still (Core Voltage higher on 1V, the FAN fix on 1200RPM). 
This is ok, its not a bug, its a feature ;)

If I start Furmark again, the temperatur and the other values are detected 
automatically and the plot goes on.

It would be nice, if on windows startup the idle card could be displayed, so I 
can set the plot active. This is my mainproblem. I always use the plot to check 
the temperatur after gaming, but actually I have to "restart" Open Hardware 
Monitor after starting a game. Then everything works fine (gaming, idle, 
gaming, idle) till the reboot.

Original comment by ha...@gmx.net on 15 Apr 2011 at 3:19

Attachments:

GoogleCodeExporter commented 9 years ago 
Thx for the report. At the moment I do not really know how to wake up all GPUs 
(forced) without too much trouble. Of course I could create a rendering context 
and add try to run shaders on all hardware, but the overhead to do that is just 
huge. So I tried to enumerate all GPUs including those that are sleeping. Can 
you try once more with this new alpha version:

http://openhardwaremonitor.org/openhardwaremonitor-v0.2.1.18-alpha.zip

and post again reports for the above three situations.

Original comment by moel.mich on 15 Apr 2011 at 8:50

GoogleCodeExporter commented 9 years ago 
Now the sleeping card is available on startup. The clocks and the voltage of 
this card arent displayed. Starting Furmark the clocks and the voltage are 
automatically displayed. After going into idle again, now the clocks and the 
voltage are displayed as "-".

Nice one! Thanks a lot!

Original comment by ha...@gmx.net on 15 Apr 2011 at 9:58

Attachments:

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Thank you for the feedback and providing all the reports.

This issue was closed by revision r304.

Original comment by moel.mich on 15 Apr 2011 at 11:53