Projects for ATI HD5670

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
So I got to the 2.5 million milestone I was aiming for in the Moo! Wrapper project with my HD5670 so now I want to see what other projects it can play with for a while. It's a somewhat older video card but has still been averaging 30-35K points per day on the Moo! project so I want to keep it busy. :)

This chart says that Collatz Conjecture, Einstein, PrimeGrid, and SETI@home should work. But I also already have the 2.5 million milestone in PrimeGrid thanks to the recent race, and none of the other projects will give me any GPU tasks. I did make sure that the project preferences are set to use the ATI GPU.

Are there any other projects that would work with this video card, or should I just let it keep pounding away at Moo! and PrimeGrid?
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Looks like you've got the projects covered.

I couldn't even find a GPU support list for SETI :rolleyes:, but it does list ATI apps now.
You've definitely got 'use ATI GPU' enabled in your online preferences? What does the event log say about it?
 

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
Yes. It's enabled. The client requests work and then says "Scheduler request completed: got 0 new tasks. Project has no tasks available."

It seems like the projects either don't have any GPU tasks, or they aren't recognizing my video card. Seti finally gave me some work today after not sending anything for about a week, but I'll have to wait to see if I get more once these tasks are completed.
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Yea they just didn't have any WUs, it would of said something about the GPU if it'd been that.
 

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
Since the "drive by" poster bumped this back up, I'll add a related question..

Why are the points so horribly crappy in the Seti project? My HD5670 got exactly 197 points for almost a full 24 hours of uninterrupted processing on the project with the tasks it manage to grab. In contrast, the same HD5670 just received 52,550 points for its first full 24 hours (and 21,446 points for the roughly 10 hours it ran before that) on Collatz Conjecture. And it was averaging just over 30K per day on Moo! Wrapper for a few weeks before that when I was only letting it run for about 12-14 hours per day.

Is Seti really that bad for points now, or did I get a set of 'inferior' work units, or does my older AMD GPU simply not work well with that project?
 

petrusbroder

Elite Member
Nov 28, 2004
13,343
1,138
126
Mark, it is the older AMD GPU. My NVIDIA GTX670 produces 6500 points running 12 hours/day - which would make some 13000 points/day when running 24h. The 500-series did even better ...
And, to a smaller part, the relatively meager points/WU seti gives.
 

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
That's disappointing. I was going to have Seti be my next GPU 'milestone' project, but it's not worth the power bill for that.. :(
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
My HD5670 got exactly 197 points
Seti is bad, but not that bad! How many WUs are "pending"?

A 6870 does about 7500 ppd in Seti.

To get better ppd in Collatz:
There is a "*.config" file in this path: "ProgramData\BOINC\projects\boinc.thesonntags.com_collatz". If it is empty, paste the following into it:

verbose=0
items_per_kernel=18
kernels_per_reduction=9
threads=8
sleep=1

Some of the middle three numbers might need tweaking one digit downward for your card.
 

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
Nothing is pending for Seti. That 24 hour period was a few weeks ago. The GPU actually ran Seti for about 5 days and the 197 points were for the last 24 hours before I shut it off so it had time to get going and have work validated. I just thought I'd ask since someone else bumped this thread.

My results in Collatz are pretty good already, but I'll take a look at that config file to see if it helps even more. :)
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
Getting the config file set right helps quite a bit. I don't remember how much but now the 6870 will do ~240,000 ppd (the same card that does ~7,500 ppd in Seti).
 

kowalabearhugs

Senior member
Sep 19, 2010
204
8
81
www.mattkowal.net
Can you expand on the meaning of those Collatz settings?

I have a Llano A8-3850 w/ HD6550D (400 Unified shader processors, 20 Texture mapping units, and 8 Render output units.)
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
Can you expand on the meaning of those Collatz settings?

http://boinc.thesonntags.com/collatz/forum_thread.php?id=1009#16503

Each Collatz 4.07 application is distributed with an empty config file. The config file has the same name as the executable but with the extension ".config".

There are a number of parameters that can be altered to improve speed or video response or to aid in solving issues. They are:

verbose=[0|1]
A value of 1 causes more information about the GPU, OpenCL version, etc. to be written to the log file. If enabled, this should be the first line of the config file so that it will report the other settings in the log file.

items_per_kernel=[10..22]
The number is the power of two 256-bit numbers (e.g. 2^N) that will be calculated per kernel call. Setting this number higher places a larger load on the GPU. Setting the number too high WILL cause the driver to crash and the application to hang. The default is 14, or 2^14, or 16384 items.

kernels_per_reduction=[2..9]The number (2^N once again) of kernels to run before doing a reduction. The default is 8 or 2^8 = 256. A lower number can improve video response. A larger number may result in a higher GPU load. Too high a number will result in CPU as well as GPU utilization.

threads=[5..10]
This contains the number of work groups to run in parallel. Higher is not necessarily faster. This number is device dependent. If set too high, the application will automatically reduce it to a value compatible with the device.
Most AMD GPUs allow up to 256 (a setting of 8). NVidia GPUs may allow 512 or even 1024 (a setting of 9 or 10). OpenCL requires a minimum of 32 (a setting of 5) according to the Khronos specifications.

build_options=[string containing any optional OpenCL build options]
This was added strictly for debugging in order to be able to use "-cl-opt-disable -Werror". If the OpenCL application crashes within 1-2 seconds of starting, you may want to use "build_options=-cl-opt-disable -Werror" and see if that fixes the problem.

sleep=[1..1000]
This controls the number of milliseconds that the application goes into a sleep state while waiting for the asynchronous kernel calls to complete. The default is 1. Setting this higher (e.g. 2-5) will result in better video response but will slow down the application considerably.

The config file will be renamed to collatz.config when it is copied to the BOINC slot folder when an application starts running. Exiting BOINC and editing the version in the project folder will not change the settings of the applications in progress as their config is taken from the slot folder.

A sample collatz.config file looks like:


verbose=1
items_per_kernel=20
kernels_per_reduction=9
threads=8
sleep=1
build_options=-Werror


Since the workunits very somewhat in the number of total steps they produce, I would suggest that you run several and take the average runtime to determine whether one set of values in the config works better than another set.

Note: The values in the sample above work quite well on my HD 6970 and HD 7970 without making either too sluggish.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
That's disappointing. I was going to have Seti be my next GPU 'milestone' project, but it's not worth the power bill for that.. :(
Well I was going to give you SETI stats for my old GTX 260 c216, but the SETI servers are down atm :rolleyes:, I do remember though that each WU took about 550s, well at least the cuda32 WUs did.
I see the client has a stash of cuda42 WUs now (whatever they are), so I guess that'll change everything.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Ok I've been able to look at my stats now, for the cuda32 WUs that took ~540-560s I got ~38-47 credits.
Of the cuda42 WUs that took ~620-660s I got 50-56 credits, 2 of the cuda42s took 1123 & 1121s & I got 92.3 & 91.1 credits.
Taking the middle of the 1st 2 sets of figures that's ~12.94 & ~12.07 credits/s & ~12.25 credits/s for the last set.

Maybe SETI just doesn't do well on ATI cards? (something about that rings a bell). Or it doesn't like 5670s much, I've no idea what sort of crunching power the 5670s have though.
 
Last edited:

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
I let Collatz run for about a week with those config.xml settings (lowering the middle three settings by 1 point each as you recommended) and my daily average has dropped from the roughly 50Kppd it was getting to about 22Kppd since I made the change so I'm thinking that at least for my old HD5670 it's better not to mess with the settings. I'm going to clear out that file (it was completely empty at first) to see if the points to back to the way they were.

Of course, it is possible that the drop in points is at least partially due to the fact that my video card is trying to die, but switching back will give me a better idea if that is the case or not.
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
my daily average has dropped from the roughly 50Kppd it was getting to about 22Kppd since I made the change so I'm thinking that at least for my old HD5670 it's better not to mess with the settings.
Bummer, sorry about that!
Without the changing the settings it defaults to where a minimum card would run best. I don't remember what the minimum card was, maybe yours is close to that?

The default "items_per_kernel" is 14, maybe your card would do better at 15 instead of 17. Or maybe the default is as good as it can do?
 

Fardringle

Diamond Member
Oct 23, 2000
9,188
753
126
It was worth a try since changing settings has a potential to help. It's probably a combination of my video card dying, and it being rather old.