[White Runner]

Do I need to finish my current WU that my cpu is working on before I enable my gpu's?

[Janus67]

Shouldn't, mine didn't have any problems.

[Macaholic]

Originally Posted by Janus67

I cross-posted to the benching team and the AMD GPU section, hope that is alright.

Excellent. Just be sure all criteria are met and understood;

Usage:
This is a still a very new core, a lot of the features have yet to be fully tested. Thus, as is the beta policy, no official support is given. You must enable the -beta flag on FAHClient, ie. set client-type=beta. If you’re using client 7.2.x or earlier, there are two options:
1. Specify -gpu-vendor=XXX, where XXX is either nvidia or ati
2. If -gpu-vendor is not specified, the core will automatically guess the platformId.
Otherwise, 7.3.6 lets you specify the particular -gpu-vendor as an option
Supported NVIDIA cards: Fermi or better (Titan does not work atm, as NVIDIA needs to publish new OpenCL drivers)
Supported ATI cards: HD5000 or better
As always, please use the latest drivers (Win XP is NOT supported due to super-old AMD drivers).

[madhatter256]

Originally Posted by Janus67

just added my 7970 to the mix, will see how it does. I didn't add the 'beta' tag first though so I will have to wait for this current one to finish me thinks.

running the 7.2.9 client right now so the gpu-vendor / ati tag didn't work, hope it doesn't matter.
-- edit deleted the gpu and re-added, just jumped 40k ppd.

wow 40k ppd for the 7970?

now I'm considering AMD GPU for my GPU upgrade purchase.

[Glaze132]

How do you add the beta flag? Where?

[Humanoid1]

Fiddly new client to get running this 7.3.6
Got there in the end though and just completed my 1st 7661 project

I am running the OpenCL APP taken from the Cat 12.8 package, to see how it would do, having had many problems with the 12.11 making NaNs and 13.1 making REaLY Low PPD on the normal projects.

So:
7661
on 7950 @ 1,050 / 1312(5248)
OpenCL APP from 12.8 package
GPU usage : 96 - 98%
GPU Temp : 62c @ 25% fan speed
CPU usage : 7.60% (Xeon X5650 6core/12thread 12MB cache @ 3.6GHz)
31,820 PPD


Cannot say I am not happy with this
Tonight after work I will try updating to 13.1 (useful for me as an app I use needs Exactly 13.1)

Thanks for the heads up on this project guys ^^


NOTE:
1) This setup runs error free on normal GPU WU's.
Last night at 68% NaNs were generated on 7661.
But completed the next 7661 just fine.

2) To get 7.3.6 working I had to:
- Express Install
- Set Folding Power to Full

At this point the client said it was folding, but GPUz said otherwise. So, after waiting Some time I :
- Right clicked the GPU folding slot and selected Fold
Then almost immediately GPUz show the GPU was working and the 7661 WU progressed a few mins later

3) Beta Tags... I followed these instructions:
http://www.overclock.net/t/1367557/core-17-beta-wu

[GoD_tattoo]

Liking this...My machine in my sig is looking at at least 110k ppd now.

[Humanoid1]

as a little update...

I have since updated from the Catalyst 12.8 OpenCL APP to the 13.1

Basically stopped the odd NaNs failed GPU WU's (had quite a few more of these) I was getting and also reduced my CPU load from 7.6% down to 0.76%

Also seems they have updated to WU to 7662 around just after the last 7661's I got after I upgraded my drivers. This Wu seems to give 2 - 3k PPD more than the original 7661 project for me ^^

NOTE:
My PPD look really low for the last few days as I lost internet. Seems some engineer accidentally unplugged my cable in the main junction box up the road -.-"
All fixed now though ^^

[Edward2]

I just checked and my GTX580's are getting 46K and 49K PPD on the 7662 WU.

[Humanoid1]

Oh very nice!
I would not have thought they would fare so well still. Those are surprisingly strong cards in many areas.

[Macaholic]

Originally Posted by Humanoid1

Oh very nice!
I would not have thought they would fare so well still. Those are surprisingly strong cards in many areas.

Read more about why, here.

[WhitehawkEQ]

I'm getting crap for these core 17 WU's

[Macaholic]

Originally Posted by WhitehawkEQ

I'm getting crap for these core 17 WU's

Read the Fermi vs. Kepler thread if you have access.

[GoD_tattoo]

I'm doing great with them. I went from 100k to almost 400k. Got to add in 6 7970's and 2 6950's. Getting around 40k from each 7970.

[Humanoid1]

LOL

That's Quite the Step Up!
my PPD have about doubled with my one 7950 and thought I was so happy!

Really nice to see AMD's GFx cards finally able to show their true colours.
I've known they are stronger than NVidia's offerings for a while in some cryptographic applications I use. So was being patient and Hoping Pande Labs would release an updated GPU core (since my 5850 lol )...

All Happy now

[Hicksimus]

It's nice that my 7850 is useful now.

Is it possible to stop the client from dropping back to using 1 less CPU core when I use the GPU? The CPU sits at 75% utilization almost all night (3/4 cores). Is it necessary to have the 4th core almost idle?

[Humanoid1]

if you want to try folding with all 4 cores...

In the "Advanced Control" click Configure

Select the Slots tab
Select CPU

click the Edit button
Change the CPUs from "-1" to "4"
Click OK
Click Save

You will likely have to then Pause Folding for the CPU.
Restart Folding on the CPU and the change should now be live.

Check out your PPD before and after.
But being a 4 physical core CPU I think you should get a nice improvement.

On my i7 6 core 12 thread CPU I find I get the same PPD running SMP 10 as I do when on 12.
Better perhaps on some projects with 10.

Good Luck

[Hicksimus]

Originally Posted by Humanoid1

if you want to try folding with all 4 cores...

Check out your PPD before and after.

Good Luck

That's awesome, thanks.

It went from ~35000 PPD with my 4.5GHz 3570k(3 cores) and 7850 to ~46000 with 4 cores. Much better.

[Humanoid1]

Now that's a satisfying improvement!

Good going

Did a new test on mine while running CPU Project 7808. Here are the results after settling for several hours:

SMP 11 - 26,130PPD TPF:7m22s
SMP 12 - 21,084PPD TPF:8m41s

Normally I run SMP10 which was recommended on earlier clients. But this new 7.3.6 client auto sets SMP11 on my CPU.
Seems the default setting for my 6 core i7 with Hyperthreading enabled is ideal.

[Janus67]

sigh, looks like one of my GPU WUs froze/got stuck trying to send at 99.9% and was like that all day.