TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
First off Welcome to Team EVGA! Looking to maximize your PPD follow these instructions. This is for people with a 5xx or newer video cards. You'll want to get the client-type beta flag in place for your video card. With that flag you'll get work more suited for your video card and more PPD=points per day. The core 17 work units also run cooler than the 8018 core 15 WU's=work units. You'll need to open Configure from the v7 Client Control, pick the Slot tab then the GPU slot be sure the GPU is marked before clicking add button or double click the GPU slot, Then click add button below and up will pop a smaller screen, top line client-type, next line beta. See picture one: To add to this you will want to free up 2 cores of the CPU slot to help the GPU fold. Not sure how many cores+HT your CPU has, but lets say 8 cores with HT enabled. Do same as above but this time double click the CPU slot, under the SMP CPU you will see a -1 as stated assuming the CPU has 8 cores you'll want to replace the -1 with the number 6. Doing as I say here should put the GPU in the 100k+ PPD instead of the low 34k PPD its doing now on core 15. See picture two: This also works with 4xx cards but needs to be a 470GTX or better to see any real benefit to PPD. The example 100K+ above is with a 780GTX, so don't expect 100k PPD on anything less than a 770GTX. Best Driver is 326:80 for Max PPD.At the time of writing 326:80 or 327:23 is the best Nvidia driver for max PPD on any Nvidia cards below a 780GTX. See the type of problems others are having here. Also be nice if you would post your findings there as well, so they can get this fixed asap, I hope. Attached Image(s) Click to enlarge  Note from the v7 Client SMP slot control: This value should be a multiple of 2.If you have 3 video cards in one Computer folding, 4 CPU cores should be free.  You'll need a PassKey and 10 completed bonus work units to get the QRB=Quck Return Bonus points.F @ H has introduced a new Quick Return Bonus (QRB) points program, which is tied to passkeys. To earn this bonus, you must use a passkey.Use the same passkey for all clients Thanks for your interest in folding. Edit: More Updates QRB info added,driver info added, format change, pictures: Last updated: 12/9/2013 a updated version from help I gave in another thread.
post edited by TheWolf - 2013/12/09 16:53:07
Attached Image(s)
|
bowlinra
SSC Member
- Total Posts : 883
- Reward points : 0
- Joined: 2011/12/05 20:58:14
- Location: Virginia, USA
- Status: offline
- Ribbons : 5
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/25 21:45:38
(permalink)
Good Reference.. Give that man a Blue Ribbon!
|
T-Wolf
Superclocked Member
- Total Posts : 101
- Reward points : 0
- Joined: 2010/10/09 06:17:10
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/26 03:59:14
(permalink)
So you would use 2 cores per 780?
|
bcavnaugh
The Crunchinator
- Total Posts : 38977
- Reward points : 0
- Joined: 2012/09/18 17:31:18
- Location: USA Affiliate E5L3CTGE12 Associate 9E88QK5L7811G3H
- Status: offline
- Ribbons : 282
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/26 06:21:51
(permalink)
T-Wolf
So you would use 2 cores per 780?
Somewhere I read the same, I make sure I have 2 Cores Per GPU with 780 (GK110) and above and 770 (GK104) and below 1 Core per GPU. Sometime this Winter I will move my 690's into a X79 MB with 12 Cores I have them now in a 8 Core Rig but want to give each 690 one CPU per GPU and turn off Hyper Threading. I is more for testing and playing around to get the most PPD I can but still Fold for the cause. I follow TheWolf and have yet had a problem, all the Old Team Members are very good at what they do I am only a newbie at Folding.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/26 07:24:48
(permalink)
T-Wolf
So you would use 2 cores per 780?
No it only needs one, but if your SMP folding odd number cores it can be unstable and loose work. Not all numbers of core are prime candidates for SMP folding especially odd numbers, plus it doesn't hurt to have a little left over free core for other things you may need to do. If you have dual 780's or dual what ever type cards then only free up two cores. One per card but always even number cores for SMP folding is best. Some SMP work units will fold on odd number cores its up to you if you want to test to see if 7 cores works for SMP and a single GPU card. I was just trying to give the most trouble free way to setup in my post above for first time users. First time users are probably not ready for all the trouble or the extra's some of us try for max PPD.  With a odd number core setup you would want to keep a close eye on it to make sure its not dropping WU's because of the odd number. Some even numbers are not prime stable such as 22 if you have a 24 core Bigadv rig you can't use 21 22 or 23 it auto drops to 20 if anything between 20 and 24 is used. They are blacklisted in the client core. The word escapes me I should be using here, maybe someone else will chime in with the correct wording.
post edited by TheWolf - 2013/11/26 08:22:50
|
notfordman
Omnipotent Enthusiast
- Total Posts : 10347
- Reward points : 0
- Joined: 2007/08/09 23:52:23
- Location: In a van, down by the
- Status: offline
- Ribbons : 28
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/26 10:51:04
(permalink)
+1 nice write up Wolf!  bowlinra
Good Reference.. Give that man a Blue Ribbon!
|
ArtyD42
Superclocked Member
- Total Posts : 212
- Reward points : 0
- Joined: 2013/11/18 18:20:26
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/26 13:06:41
(permalink)
This is a thing I needed back when I joined this team. Thanks for the write up Wolf. Also for those unaware of what their cpu's are capable of: http://www.piriform.com/speccy/download I hope I can post links now. If not the link I just tried to post was to speccy.
post edited by ArtyD42 - 2013/11/26 13:08:22
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/11/27 03:43:23
(permalink)
Updated OP format, clean up write up. Note from the v7 Client SMP slot control: This value should be in multiple of 2.
Attached Image(s)
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/07 03:30:43
(permalink)
Updated OP with QRB setup info. Will update or add to it as needed. Check back often.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/08 23:11:38
(permalink)
Update OP with driver info to use for max PPD on Nvidia video cards below 780GTX.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/09 16:50:29
(permalink)
Updated opening post with PassKey info for QRB. A must have.
|
Equitum
Superclocked Member
- Total Posts : 140
- Reward points : 0
- Joined: 2013/11/20 14:23:23
- Location: DFW, Texas
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/20 12:57:15
(permalink)
Your original post deserves a blue ribbon, TheWolf - so many people have been helped by it. Unfortunately, I haven't had any luck with my PPD. I'm running a i5-4670k @ 4.5GHz with two cores folding, leaving the other two cores to help my 780 Ti Classified w/ ACX (overclocked with +42 MHz on gpu clock and ~+30 MHz memory clock) fold, and I'm only getting 45-75k PPD, which is really disappointing. My current PPD shows just about 48k on HFM, and about 55k on the FAH v7 client. I'll attach some pictures of my configuration, but I have the beta flag set on both my cpu and gpu, and I have a passkey, and have completed over 10 WUs with the passkey, so I should be getting QRB as well.    Anything look awry?
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/22 01:37:09
(permalink)
Equitum
Your original post deserves a blue ribbon, TheWolf - so many people have been helped by it.
Unfortunately, I haven't had any luck with my PPD. I'm running a i5-4670k @ 4.5GHz with two cores folding, leaving the other two cores to help my 780 Ti Classified w/ ACX (overclocked with +42 MHz on gpu clock and ~+30 MHz memory clock) fold, and I'm only getting 45-75k PPD, which is really disappointing. My current PPD shows just about 48k on HFM, and about 55k on the FAH v7 client. I'll attach some pictures of my configuration, but I have the beta flag set on both my cpu and gpu, and I have a passkey, and have completed over 10 WUs with the passkey, so I should be getting QRB as well.
 
Anything look awry?
Sorry for getting back to you late, been into other things and forgot to check here to see if anyone needed any help. Everything is set correct by your screen shots, but looking at your stats page it must be coming down to the bad luck of the draw on work units as I see you are getting core 15 WU's and not core 17. There was I guess a shortage of core 17 "I think" for a short while, but I thought that was over. Only thing I can think of is to set your clients to finish "GPU", let it finish the work its doing, then close/exit v7 might even try a reboot then re-open v7 and maybe with luck it will connect you to a core 17 server. Good luck and hope your next WU is a core 17.
post edited by TheWolf - 2013/12/22 01:39:23
|
Equitum
Superclocked Member
- Total Posts : 140
- Reward points : 0
- Joined: 2013/11/20 14:23:23
- Location: DFW, Texas
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/22 09:15:20
(permalink)
TheWolf
Equitum
--snip--
Sorry for getting back to you late, been into other things and forgot to check here to see if anyone needed any help.
Everything is set correct by your screen shots, but looking at your stats page it must be coming down to
the bad luck of the draw on work units as I see you are getting core 15 WU's and not core 17.
There was I guess a shortage of core 17 "I think" for a short while, but I thought that was over.
Only thing I can think of is to set your clients to finish "GPU", let it finish the work its doing, then close/exit v7
might even try a reboot then re-open v7 and maybe with luck it will connect you to a core 17 server.
Good luck and hope your next WU is a core 17.
I thought it might have to do with only getting core 15 WUs... I'll try and close/reset/reinstall the client and see if I can get a core 17 WU after this WU is finished. Thanks!
|
bcavnaugh
The Crunchinator
- Total Posts : 38977
- Reward points : 0
- Joined: 2012/09/18 17:31:18
- Location: USA Affiliate E5L3CTGE12 Associate 9E88QK5L7811G3H
- Status: offline
- Ribbons : 282
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/22 10:13:24
(permalink)
Even with your settings on the 19 all 9 of my GPU Rigs received Core 15s. I removed the Core 15 Files and all the Work Data, waited 24 hours (I was Crunching) then went back to Folding last night and back now on Core 17's. I did not un-install and re-install FAH.
post edited by bcavnaugh - 2013/12/22 10:15:42
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/22 19:31:44
(permalink)
Equitum
I thought it might have to do with only getting core 15 WUs... I'll try and close/reset/reinstall the client and see if I can get a core 17 WU after this WU is finished.
Thanks!
Let us know how it works out? I'm pulling for ya to get some core 17 love.
|
Equitum
Superclocked Member
- Total Posts : 140
- Reward points : 0
- Joined: 2013/11/20 14:23:23
- Location: DFW, Texas
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/22 19:46:46
(permalink)
I waited for my last core 15 WU to finish, closed the client completely, and let it sit for a few hours. I just started it back up a few hours ago, and it gave me another core 15 WU. After this WU finishes, I might try doing what bcavnaugh did and delete some of the work data and core 15 files and let it sit longer. I'll keep y'all updated - it could just be blind dumb luck on getting core 15 WUs and no core 17s.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/23 07:57:35
(permalink)
You could always look at the log real quick when you 1st start the GPU client
to note what IP its connecting to, but 1st note what IP it is now getting core 15
from as core 17 will come from a different IP address. I have tested and caught
it before downloading from a certain IP did a dump or delete of the GPU client
re-setup and it grabbed a different IP address on next connection.
Then you could also have what ever IP core 15 is coming from block by
some type of software or ban that IP from connection to your system this
should force it to try to connect to a different IP and hopefully a core 17 IP.
Only fall back I see from doing a IP block is if there are no core17 WU to be gotten
it will sit there trying to get work until some become available. Sitting idle/no PPD.
|
Equitum
Superclocked Member
- Total Posts : 140
- Reward points : 0
- Joined: 2013/11/20 14:23:23
- Location: DFW, Texas
- Status: offline
- Ribbons : 0
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/24 14:39:15
(permalink)
A little update... While I was mining last night, FAH must have detected idle or something, so it kicked up and started folding on my GPU. I happened to look at it a few minutes ago, and, lo and behold, it's a Core 17 WU. Just my luck, I suppose, that right when I start mining so I can get some money for a dedicated folding rig, I get a Core 17 WU for the first time!
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/24 16:05:00
(permalink)
Equitum
A little update... While I was mining last night, FAH must have detected idle or something, so it kicked up and started folding on my GPU. I happened to look at it a few minutes ago, and, lo and behold, it's a Core 17 WU. Just my luck, I suppose, that right when I start mining so I can get some money for a dedicated folding rig, I get a Core 17 WU for the first time!
Well as always it seem it dammed if you do dammed if you don't. Glad you were able to grab a core 17, but you'll have to decide what to do now I can't help with that.
|
Panther-X
New Member
- Total Posts : 54
- Reward points : 0
- Joined: 2013/05/08 07:19:03
- Status: offline
- Ribbons : 1
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/25 04:38:38
(permalink)
FYI, there are plenty of FahCore_17 WUs ready to be assigned ( post). Thus, many donors should be getting FahCore_17 WUs. If not, then you may try to use a default configuration to ensure that there isn't a configuration conflict. If you don't get FahCore_17 WUs even with a default configuration, you may want to report it in that thread to see if there is a pattern and if so, help find an obscure bug. TheWolf -> Please note that in Linux (64-bit only for now), you need to have client-type set to advanced to get FahCore_17 WUs. If there isn't any client-type set, you will not get any FahCore_17 WUs (Empty Work Server Message will be printed in the log file). Also, I am not sure what happens if the client-type is set to beta under Linux.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/25 07:32:43
(permalink)
Panther-X
FYI, there are plenty of FahCore_17 WUs ready to be assigned (post). Thus, many donors should be getting FahCore_17 WUs. If not, then you may try to use a default configuration to ensure that there isn't a configuration conflict.
If you don't get FahCore_17 WUs even with a default configuration, you may want to report it in that thread to see if there is a pattern and if so, help find an obscure bug.
TheWolf -> Please note that in Linux (64-bit only for now), you need to have client-type set to advanced to get FahCore_17 WUs. If there isn't any client-type set, you will not get any FahCore_17 WUs (Empty Work Server Message will be printed in the log file). Also, I am not sure what happens if the client-type is set to beta under Linux.
Last I ran v7 under Linux beta flag worked for getting core 17 work, since core17 was released to Linux I've never used the advanced flag.
|
Panther-X
New Member
- Total Posts : 54
- Reward points : 0
- Joined: 2013/05/08 07:19:03
- Status: offline
- Ribbons : 1
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/25 13:34:00
(permalink)
Okay, thanks for informing me about that, TheWolf.
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/27 02:47:22
(permalink)
Here are some Expert extra flags that could be helpful in certain cases. "pause-on-start" "true" Say you might like to change your config setup for a certain client before getting new work, "pause-on-start" "true" is very useful. This way if a WU has been set to finished and has finished and lets say you need to reboot for some reason. It will not pickup a new work unit on start of v7 until you manually start each clients/GPU/SMP. Remove clients or re-config flags as needed, then manual start workers. Workers will always have to be manually started with this flag in place if v7 is closed then restarted. next-unit-percentage 100 This tells v7 to not download new work until the work its doing has completely finished. Default I think was 96% so it downloads the next WU before the other work is completed so this next unit sets idle for some time while the other unit completes costing you some of your QRB. max-units x max units is useful say if you would like to stop folding after a certain amount of completed WU. Replace x with the amount of WU's you would like to work before v7 pauses. This is useful if your getting a string of lets say 3 hour WU's and you only want to allow folding for 9 to 12 hours, set x to 3 and it will pause after 3 units. This is not a for sure thing as WU can change and so time can change that it takes to finish those, but still somewhat useful.
post edited by TheWolf - 2013/12/27 11:13:25
Attached Image(s)
|
Panther-X
New Member
- Total Posts : 54
- Reward points : 0
- Joined: 2013/05/08 07:19:03
- Status: offline
- Ribbons : 1
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/27 04:50:08
(permalink)
TheWolf...next-unit-percentage 100
This tells v7 to not download new work until the work its doing has completely finished.
Default I think was 96% so it downloads the next WU before the other work is completed
so this next unit sets idle for some time while the other unit completes costing you some of your QRB... The default is 99% However, in V7.3.6, there is a bug which causes the WU to be downloaded 1% before the configured value. Thus, if you have set it to 100%, it would download the WU at 99% V7.2.9 doesn't have this bug. Moreover, here are two additional settings which can be useful if you are trying to verify the stability of overclock: max-slot-errors Default is 10. The number of slot errors allowed to happen before the Slot is in failed state. I haven't yet figured out a way to reliably trigger this. max-unit-errors Default is 5. The number of WU errors allowed to happen before the Slot enters a failed state. Most of the time, if you see a slot in failed state, this is what triggered it. Screen shot which shows that I have set it on a per slot basis (can be configured as an extra client option too). With a value of 1 for each, I don't need to check the logs for error, just look at the Slot status.
Attached Image(s)
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/27 11:33:02
(permalink)
Panther-X
TheWolf...next-unit-percentage 100
This tells v7 to not download new work until the work its doing has completely finished.
Default I think was 96% so it downloads the next WU before the other work is completed
so this next unit sets idle for some time while the other unit completes costing you some of your QRB...
The default is 99% However, in V7.3.6, there is a bug which causes the WU to be downloaded 1% before the configured value. Thus, if you have set it to 100%, it would download the WU at 99% V7.2.9 doesn't have this bug.
Thanks for the correction on the % count, this is one reason I'm still using V7.2.9. Plus I like the button layout at the top of the older version, where you can stop/start all workers at once. If they ever get that % bug fixed I may move forward one day.
Attached Image(s)
|
Panther-X
New Member
- Total Posts : 54
- Reward points : 0
- Joined: 2013/05/08 07:19:03
- Status: offline
- Ribbons : 1
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/27 14:59:11
(permalink)
|
TheWolf
CLASSIFIED Member
- Total Posts : 3800
- Reward points : 0
- Joined: 2007/11/14 16:05:23
- Location: Moss Point, Ms
- Status: offline
- Ribbons : 9
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/27 19:58:36
(permalink)
|
Panther-X
New Member
- Total Posts : 54
- Reward points : 0
- Joined: 2013/05/08 07:19:03
- Status: offline
- Ribbons : 1
Re: Max PPD for GPU/SMP folding with Core 17.
2013/12/28 14:12:58
(permalink)
|
Zagen30
SSC Member
- Total Posts : 528
- Reward points : 0
- Joined: 2009/11/27 12:21:01
- Location: Laguna Niguel, CA
- Status: offline
- Ribbons : 3
Re: Max PPD for GPU/SMP folding with Core 17.
2014/02/06 14:13:44
(permalink)
 Core i7 3770K @4.3 GHz, 8 GB DDR3 @2133 MHz, 2xGTX 970 SSC @1379 MHz, Windows 7 Pro 64-bit Affiliate code: K4BJ4FHUX5
|