• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Project 6702

M

musky

[H]ard|DCer of the Year 2012
Joined
Dec 14, 2009
Messages
3,154
Hadn't seen this one before.

Project ID: 6702
Core: GRO-A3
Credit: 921
Frames: 100


Name: believer
Path: \\192.168.1.140\fah\
Number of Frames Observed: 57

Min. Time / Frame : 00:05:58 - 17,126.7 PPD
Avg. Time / Frame : 00:06:10 - 16,300.3 PPD
Cur. Time / Frame : 00:06:10 - 16,234.0 PPD
R3F. Time / Frame : 00:06:10 - 16,234.0 PPD
All Time / Frame : 00:06:10 - 16,234.0 PPD
Eff. Time / Frame : 00:06:15 - 16,017.6 PPD

This is on my dual E5530 machine. About the same (horrible) ppd as 6701's. This is not a good sign for A3 folding. I wish they would get bigadv fixed for Linux.
 
musky said:
This is on my dual E5530 machine. About the same (horrible) ppd as 6701's. This is not a good sign for A3 folding. I wish they would get bigadv fixed for Linux.
Click to expand...
I wish they would too. Strange as it may seem, I was getting approximately 3-4 min lower TPFs with slightly slower frequencies than I am in Windows. I don't believe it's necessarily true that all A3 WUs or all system architectures perform better in Windows from my experience.
 
APOLLO said:
I wish they would too. Strange as it may seem, I was getting approximately 3-4 min lower TPFs with slightly slower frequencies than I am in Windows. I don't believe it's necessarily true that all A3 WUs or all system architectures perform better in Windows from my experience.
Click to expand...


that was all the hype that everyone believed in.. the only benefit to the winsmp and linux smp is the stability which we all knew was going to happen.. i was the only one saying there probably wasnt going to be a benefit point wise when the bigadv moved to windows due to the windows overhead on the client.. but oh well..
 
Yea this is brutal. I've gotten nothing but a steady stream of 6701 and 6702s for days now. Is the windows bigadv server still having issues?

My 6702 ppd results so far:

Q6600 at 3ghz: about 4100 ppd

1055T at 3.8ghz: about 10000 ppd
 
On other than my Big Adv rigs, I am getting only 6701 & 6702. 6702 seems to be slightly (and I mean slightly) higher PpD than 6701, but yes the PpD is STILL in the crapper.

 
I seem to get them in cycles...I've had 2-3 machines finish their WU's within a 30 minute span and all of them end up picking up one of them.

They can make my total ppd drop by probably 25k give or take. The only consolation though is that other teams are getting them too, so we aren't the only ones that get the point swings.
 
my AMD X6 just got a 6024 and it's slicing through it like butter at 15,166 ppd. I wish I could get those all day long.
 
You can keep from getting the 67xx units if you turn off -advmethods, at least for now from what I've found. Only machines with -advmethods on get them so far. As soon as I shut it off on my other machines, they stopped getting them.

I'm not 100% sure about this, but it's worth a shot.
 
10e said:
You can keep from getting the 67xx units if you turn off -advmethods, at least for now from what I've found. Only machines with -advmethods on get them so far. As soon as I shut it off on my other machines, they stopped getting them.

I'm not 100% sure about this, but it's worth a shot.
Click to expand...
ooohhh....and what unit did you end up getting?
 
10e said:
You can keep from getting the 67xx units if you turn off -advmethods, at least for now from what I've found. Only machines with -advmethods on get them so far. As soon as I shut it off on my other machines, they stopped getting them..
Click to expand...
Do you mean in the config, because I don't use the -advmethods flag and I still receive them.
 
APOLLO said:
Do you mean in the config, because I don't use the -advmethods flag and I still receive them.
Click to expand...

Whether it's in the config (ie. when using -configonly or -config flag when running) or as an added flag in either the shortcut or in the "additional switches" option (again during -config or -configonly), it amounts to the same thing if I'm remembering correctly.

I've changed up my setup so that I have it in the "additional switches" area during configuration ONLY, because at least then I can see in HFM.NET if it's in the command line.

Lemme know what happens, because so far, as I said I'm not 100% sure about this, as well as it can change at any time if the units are going from "beta/advanced" to general distribution. I've changed up all my clients and removed the switch and I'll be watching for any P670x units.
 
10e said:
Lemme know what happens, because so far, as I said I'm not 100% sure about this, as well as it can change at any time if the units are going from "beta/advanced" to general distribution. I've changed up all my clients and removed the switch and I'll be watching for any P670x units.
Click to expand...
I eliminated it entirely from at least two clients, but did this less than a day ago. I'll see what happens. However, if Stanford decides on a general release, it won't make a difference how are clients are configured assuming they've been releasing these to -adv clients only.
 
jebo_4jc said:
ooohhh....and what unit did you end up getting?
Click to expand...

Mostly 601x units, and a couple of 604x and 605x units mixed in there.

My last P6701 is finishing tonight or tomorrow morning on my last remaining -advmethods running client, so we'll see what happens next.
 
10e said:
Mostly 601x units, and a couple of 604x and 605x units mixed in there.

My last P6701 is finishing tonight or tomorrow morning on my last remaining -advmethods running client, so we'll see what happens next.
Click to expand...

Have you run across a P6041 yet? I'm curious to see your PPD...I am crunching the 3rd WU and I can barley get 10800K PPD out of it..

When I built this rig I was seeing between 14-16K PPD consistently just out of the SMP client..The last three weeks or so I can barley average 16-18K PPD with my 8800GT picking up almost 6K of that total:mad:...

 
ccityinstaller said:
Have you run across a P6041 yet? I'm curious to see your PPD...I am crunching the 3rd WU and I can barley get 10800K PPD out of it..

When I built this rig I was seeing between 14-16K PPD consistently just out of the SMP client..The last three weeks or so I can barley average 16-18K PPD with my 8800GT picking up almost 6K of that total:mad:...

Click to expand...

Don't feel bad as my folding rig always saw a big hit when crunching a 6040/6041 A3 WU. I got quite a few of those, and it bought my PPD average down to about 10-11k ppd, which is about what I saw when I started getting a bunch of those pesky 6701's in a row. At least for those of us running Windows bigadv right now, we pretty much know what we will get since windows bigadv is rock stable and there is only one WU right now(the 2684).
 
Oh yeah, I've seen quite a few.

Name: UGLY_1_SMP
Core i7 87K @ 4.00Ghz (160*25) and two nVidia clients running concurrently
Number of Frames Observed: 43
Min. Time / Frame : 00:09:44 - 15,278 PPD
Avg. Time / Frame : 00:10:11 - 14,276 PPD

Name: Ugly_0_SMP
Core I7 920 @ 3.85 Ghz (203*x19) with NO GPU clients running concurrently
Number of Frames Observed: 102
Min. Time / Frame : 00:09:17 - 16,402 PPD
Avg. Time / Frame : 00:09:23 - 16,141 PPD

These two boxes average around 19K on regular P601x, P602x units. I haven't gotten these on my 1055T @ 3.5Ghz yet, but a p605x gets about 11.8K PPD with 4xGPU2 clients running on it at the same time. I am finding the X6 CPUs handle GPU clients with less PPD drain/loss than the Core I7s though.

Yay real cores.

ccityinstaller said:
Have you run across a P6041 yet? I'm curious to see your PPD...I am crunching the 3rd WU and I can barley get 10800K PPD out of it..

When I built this rig I was seeing between 14-16K PPD consistently just out of the SMP client..The last three weeks or so I can barley average 16-18K PPD with my 8800GT picking up almost 6K of that total:mad:...

Click to expand...
 
10e said:
Oh yeah, I've seen quite a few.

Name: UGLY_1_SMP
Core i7 87K @ 4.00Ghz (160*25) and two nVidia clients running concurrently
Number of Frames Observed: 43
Min. Time / Frame : 00:09:44 - 15,278 PPD
Avg. Time / Frame : 00:10:11 - 14,276 PPD

Name: Ugly_0_SMP
Core I7 920 @ 3.85 Ghz (203*x19) with NO GPU clients running concurrently
Number of Frames Observed: 102
Min. Time / Frame : 00:09:17 - 16,402 PPD
Avg. Time / Frame : 00:09:23 - 16,141 PPD

These two boxes average around 19K on regular P601x, P602x units. I haven't gotten these on my 1055T @ 3.5Ghz yet, but a p605x gets about 11.8K PPD with 4xGPU2 clients running on it at the same time. I am finding the X6 CPUs handle GPU clients with less PPD drain/loss than the Core I7s though.

Yay real cores.
Click to expand...

Those are some nice numbers on your i7 rigs..I am curious to see what your X6 does..

I just paused my single GPU client for about 20 mins to see it's affect on PPD...With GPU running, I was seeing 13230K PPD...With it paused it only jumped to 14400K PPD...This is on a P6015...I am curious to see what happens if I run the -smp5 flag vs the -smp6..

I am really starting to think that I am being severely hampered by running my ram in "single channel." I didn't think it would matter much since the PII has independent MC's, but I guess it does..I have tried everything possible to get this board to boot in dual channel, even replacing my ram..I am running Corsair DDR1600 CAS 8 at the moment, which is supposed to be uber friendly..:rolleyes:

 
Yep, the memory bandwidth helps for sure with A3 units and BigAdv.

I am running these modules stock 1333mhz at 9-9-9-24-2T on an MSI 790FX-GD70. I saw your other thread about the dual channel issues. That is teh sux.

So I shut off the -advmethods flags on my clients for the last few days and no P670x units. Today I ran the client in regular cmd line mode (ie. not service mode) and added -advmethods, and guess what? P6701.
 
10e said:
Yep, the memory bandwidth helps for sure with A3 units and BigAdv.
Click to expand...
Yep, depending on the board I would agree even with older systems like Core2 architecture, going up to quad channel support substantially decreases TPF from what I noticed.

So I shut off the -advmethods flags on my clients for the last few days and no P670x units. Today I ran the client in regular cmd line mode (ie. not service mode) and added -advmethods, and guess what? P6701.
Click to expand...
Yep, all but one of my systems has no admethods set. The one remaining is running a P6702...
 
Shutting off the -advmethods does not work for me. I do not have it set on any of my boxen and I still get 670x series WU
 
capreppy said:
Shutting off the -advmethods does not work for me. I do not have it set on any of my boxen and I still get 670x series WU
Click to expand...
What do you have as the client type in your config file?
 
Sure it's not in the config still? I don't use config in the client, I use the shortcut so I can change them quick
Posted via [H] Mobile Device
 
I reconfigured a couple of clients removing the advmethods flag and changing the WU size to small. We'll see if that makes a difference or not.
 
mplee Looks like you will be passing me in a few days. ill throw you a beer as you pass bye
 
Well I checked all of my clients and I guess I had -advmethods in a few places. I've removed all existence of it and with the exception of one client (it is still working thru a 6702 and should finish tonight), all of my SMP clients are OFF OF 670x series WUs. Whew!!!
 
mplee said:
I reconfigured a couple of clients removing the advmethods flag and changing the WU size to small. We'll see if that makes a difference or not.
Click to expand...

Changing WU size in the SMP client won't help i'm afraid - it is hard coded to accept big WU
 
lehmann said:
mplee Looks like you will be passing me in a few days. ill throw you a beer as you pass bye
Click to expand...

Thanks!

capreppy said:
Well I checked all of my clients and I guess I had -advmethods in a few places. I've removed all existence of it and with the exception of one client (it is still working thru a 6702 and should finish tonight), all of my SMP clients are OFF OF 670x series WUs. Whew!!!
Click to expand...

It looks like its worked for me as well, since packet size doesn't do anything.

I've got a few machines that still need to finish their current 670x WU's though.
 
Last edited:
Well, I guess I lied. I just picked up 2 670x series WU. This is on boxes that no longer have the -advmethods flag set. :(
 
Yeah I removed advmethods a few days ago, I still have picked up a few (though not as many). They must not be as reliant on the flag anymore to be doled out or something. :confused: Wouldn't really surprise me in the least. :rolleyes:
 
I picked up a couple this evening too. I hadn't gotten any since removing the advmethods flag.
 
Maybe a few runs made it to general release?
Posted via [H] Mobile Device
 
6701 and 6702 are out for the general public. Removing the -advmethods flag will not make any difference to your chances of getting one.
 
Damn.... That would explain the point drop...
Posted via [H] Mobile Device
 
I knew it. Oh well, I guess this means more D... :eek:
 
some people were hoping they woul up the value to reflect the level of work being done
 
You must log in or register to reply here.
Back
Top