• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

PPD extremely low

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Suprchargd

Member
Joined
Feb 5, 2007
Location
Chicago, IL
jws2346 said:
The only thing I can say is , for me, the "proof is in the pudding" (like they have wont to say) I pretty much use my machines for folding using the Linux SMP client, my machines are mildly OC'ed
E6600 tpf 13 mins (2.7 MHz)
E6400 tpf >13 mins (2.8 MHz)
E6300 tpf about 15 mins even (2.4 MHz)
PD 805 tpf >29 mins (2.8 MHz)
AMD Athlon X@ 4200 (2.4 MHz) about 25 mins tpf, so by looking at these stats (time per frame,tpf) you can definitely see a large advantage in the C2D's. Besides the PD 805 is too hard to cool and the X2 is a 939 skt, on a MSI K8N Neo4 mobo and I like DDR2 RAM.

Edit: Come on April 22nd, a Q6600,yeah, yeah, yeah:drool:

Hope jws doesn't mind but I saw this post and wanted to use it here. I'm using XP so I don't know what kind of a difference there is with Linux, but I'm getting 43mins as my TPF under ElectronMicroscope with Rig 1 in my sig. I've got two instances running as services, is the SMP client that much faster than the normal client? Or is my 43tpf still a bit high?

To me it seems as though something is not right, theres no way he can get 13mins with a [email protected] and I'm getting 43tpf with a [email protected], even if Linux+SMP gives you a slight advantage.

What do you guys think?

PS I'm not currently folding for Team32 but I wanted to switch over. If I do so, will I lose all the points I've gotten up until now / leave those points with my old team?
 
Try running the Windows SMP client. What protein are you getting with the dual clients? You've probably just been getting huge WUs.
 
Your TPF's don't mean much without knowing which wu (project #) they are. There a quite a few low producing projects on the standard windows client.

You will get higher point production (ppd) with either windows smp or linux smp than you will get for 2 single core clients.

Part of the reason is they are both beta and have bonus points.

The other reason is that smp folding produces a longer duration similation in a shorter period than 2 single cleints (according to stanford) so more points are awarded.

There was a lot of debate earlier about the point system, but regardless of the merits of the debate, the current reality is that smp will produce more points assuming your rig can complete them within the short 4-5 day deadlines. If it can't meet the deadline (and some multicores can't) then run 2 standard clients.

The only other way to boost production is if you have an ati x19xx video card, you could use one core to frive the gpu cleint and use the other to run a standard client.

Rig #1 iin your sig will produce 1500-2000 ppd smp folding if dedicated, less if it does a LOT of other work.

Rig #2 will probably also be best folding smp but i can't quote #'s for it. I assume it isn't oc'ed much if at all on that mobo? A guess would be 600-800 ppd? But i have think i have heard of some folks haveing excessive frametimes. So, as inkfx suggested, I would try it out .... just check its progress to make sure frametimes are < 57? minutes and it should make the deadline if it runs 724 with no other load.

As to switching team, the points never move. So past points would stay under your id on that team.

But we would love to have you join us here anyway :)

And maybe we can even help you proiduce so much more that you won't miss the ones you left behind. :D

Beyond the points thing, we are all folding for the same cause.

And you wouldn't be the first person to have points on more than 1 team ... some have millions on more than 1 team :eek:
 
Last edited:
Suprchargd said:
Hope jws doesn't mind but I saw this post and wanted to use it here. I'm using XP so I don't know what kind of a difference there is with Linux, but I'm getting 43mins as my TPF under ElectronMicroscope with Rig 1 in my sig. I've got two instances running as services, is the SMP client that much faster than the normal client? Or is my 43tpf still a bit high?

To me it seems as though something is not right, theres no way he can get 13mins with a [email protected] and I'm getting 43tpf with a [email protected], even if Linux+SMP gives you a slight advantage.

What do you guys think?

PS I'm not currently folding for Team32 but I wanted to switch over. If I do so, will I lose all the points I've gotten up until now / leave those points with my old team?


To be completely honest with you Suprchargd I don't particularly like being called a BSer. Below is my folding progress file, I think it says about 13 mins for my E6600.

[02:08:35] Completed 300000 out of 500000 steps (60 percent)
[02:21:36] Writing local files
[02:21:36] Completed 305000 out of 500000 steps (61 percent)
[02:34:37] Writing local files
[02:34:37] Completed 310000 out of 500000 steps (62 percent)
[02:47:39] Writing local files
[02:47:40] Completed 315000 out of 500000 steps (63 percent)
[03:00:39] Writing local files
[03:00:39] Completed 320000 out of 500000 steps (64 percent)
[03:13:41] Writing local files
[03:13:41] Completed 325000 out of 500000 steps (65 percent)
[03:26:43] Writing local files
[03:26:43] Completed 330000 out of 500000 steps (66 percent)
[03:39:43] Writing local files
[03:39:43] Completed 335000 out of 500000 steps (67 percent)
[03:52:43] Writing local files
[03:52:43] Completed 340000 out of 500000 steps (68 percent)
[04:05:44] Writing local files
[04:05:44] Completed 345000 out of 500000 steps (69 percent)
[04:18:50] Writing local files
[04:18:50] Completed 350000 out of 500000 steps (70 percent)
[04:31:49] Writing local files
[04:31:49] Completed 355000 out of 500000 steps (71 percent)

Edit: In all honesty I think you should have called maybe I made a mistake in my math, that's what PM's are for

Edit again: I think some people on this team (Team 32) are getting a much lower ppd with a more OC'ed E6600
 
Last edited:
jws I don't think he meant any harm by what he said. He simply just didn't understand why your rig was doing better than his at lower frequency.
 
inkfx said:
jws I don't think he meant any harm by what he said. He simply just didn't understand why your rig was doing better than his at lower frequency.

If that's the case, I really don't care, I just read
" To me it seems as though something is not right, there's no way he can get 13 mins with a E6600 OC'ed to 2.7 GHz"and I thought Is there something wrong with my math, I thought I was figuring ppd right. Now to answer the question about points, I don't think you loose them switching teams, they're just not carried over, Me, I could give a sh*t less about my points if I have a shot at a team better than the one I'm on, I can always get more points.

Edit: I think a lot of people, myself included, will get more points when the April 22nd price cuts come into effect
 
Last edited:
inkfx said:
jws I don't think he meant any harm by what he said. He simply just didn't understand why your rig was doing better than his at lower frequency.

"theres no way he can get 13mins with a [email protected] and I'm getting 43tpf with a [email protected]"

jws, it wasn't a shot at you. You should have kept reading until the end of the sentence, I wasn't doubting your numbers, I was just wondering what was up with mine. The last thing I have time for is to e-thug all over forums. I just saw your post and realized that my ppd numbers seemed really low, so I wanted to use your numbers as a comparison.

I'll give the SMP a go and see how it works out.
 
I thought there were going to be price cuts April 22nd BUT the big cuts when the Quad will be ~260 dollars won't be until Q3?
 
pscout said:
Your TPF's don't mean much without knowing which wu (project #) they are. There a quite a few low producing projects on the standard windows client.

Project 2125, Protein 1258 currently.

Rig #1 iin your sig will produce 1500-2000 ppd smp folding if dedicated, less if it does a LOT of other work.
This one stays on 24/7 folding, doesn't really do anything less lately, no gaming, too busy with school so I just use it when I'm home to check e-mail and browse the forums I'm on.
Rig #2 will probably also be best folding smp but i can't quote #'s for it. I assume it isn't oc'ed much if at all on that mobo? A guess would be 600-800 ppd? But i have think i have heard of some folks haveing excessive frametimes. So, as inkfx suggested, I would try it out .... just check its progress to make sure frametimes are < 57? minutes and it should make the deadline if it runs 724 with no other load.
Not OC'd at all. I'm eagerly waiting to see the results on the 650i Ultra. Can't even POST with a 1 fsb bump. This one is barely on though, I'll have to see about getting the fam to agree to keep it on 24/7. Everyone complains about noise, which is why I've switched over to 120mm fans to try and alleviate that problem, but some people are just anal about it, claiming the PC disturbs them two rooms down.:rolleyes:

Beyond the points thing, we are all foling fot the cause.

Completely agree, but you have to admit its fun tracking your progress.
 
THEnumber337 said:
I thought there were going to be price cuts April 22nd BUT the big cuts when the Quad will be ~260 dollars won't be until Q3?

"The Q6600 currently costs $851 when sold in batches of 1,000 CPUs, slightly more if it's bought as a boxed part. On 22 April, its price will fall to $530, a reduction of 37.7 per cent, Chinese-language website HKEPC claims. It's been known that Intel is planning to cut the Q6600's price by that margin for some time, but only now do we have a date for the reduction."
 
Suprchargd said:
"The Q6600 currently costs $851 when sold in batches of 1,000 CPUs, slightly more if it's bought as a boxed part. On 22 April, its price will fall to $530, a reduction of 37.7 per cent, Chinese-language website HKEPC claims. It's been known that Intel is planning to cut the Q6600's price by that margin for some time, but only now do we have a date for the reduction."

Ah okay, I was just looking at that website. Pretty good price reduction and it looks like big reductions are in store for Q3 too.
 
THEnumber337 said:
I thought there were going to be price cuts April 22nd BUT the big cuts when the Quad will be ~260 dollars won't be until Q3?

Yeah, it's my understanding come Q3 the Q6600 goes to $266 :p . I think the Q6600 in Q2 goes from like $800+ US dollars to $530 dollars and, correct me somebody if I'm wrong, the server chip x3410 skt 775 (Q6400) goes for about $400 US dollars. I read a post on this forum from a very reliable Team 32 member that the x3410 is probably a better OC'er than the Q6600. I think even at $530 and $400 these are some pretty good deals. :shrug: (remember I'm only talking about folding)
 
Alright installed SMP. Can someone check how many PPD this would come out to:

[16:24:13] - Files status OK
[16:24:13] this execution.
[16:24:13] - Created dyn
[16:24:13] - Files status OK
[16:24:14] percent)
[16:24:14] - Starting from initial work packet
[16:24:14]
[16:24:14] Project: 2652 (Run 0, Clone 6, Gen 15)
[16:24:14]
[16:24:14] packet
[16:24:14]
[16:24:14] Project: 2652 (Run 0, Clone 6, Gen 15)
[16:24:14]
[16:24:14] ing from initial work packet
[16:24:14]
[16:24:14] Project: 2652 (Run 0, Clone 6, Gen 15)
[16:24:14]
[16:24:15] Entering M.D.
[16:24:21] Protein: Protein
[16:24:21] Writing local files
[16:24:21] Extra SSE boost OK.
[16:24:22] ercent)
[16:36:48] Writing local files
[16:36:48] Completed 10000 out of 1000000 steps (1 percent)
[16:49:12] Writing local files
[16:49:12] Completed 20000 out of 1000000 steps (2 percent)
[17:01:45] Writing local files
[17:01:45] Completed 30000 out of 1000000 steps (3 percent)
[17:13:53] Writing local files
[17:13:53] Completed 40000 out of 1000000 steps (4 percent)
[17:26:01] Writing local files
[17:26:01] Completed 50000 out of 1000000 steps (5 percent)
[17:38:12] Writing local files
[17:38:12] Completed 60000 out of 1000000 steps (6 percent)
[17:50:19] Writing local files
[17:50:19] Completed 70000 out of 1000000 steps (7 percent)
[18:02:28] Writing local files
[18:02:28] Completed 80000 out of 1000000 steps (8 percent)
[18:14:36] Writing local files
[18:14:36] Completed 90000 out of 1000000 steps (9 percent)
[18:26:59] Writing local files
[18:26:59] Completed 100000 out of 1000000 steps (10 percent)
[18:39:28] Writing local files
[18:39:28] Completed 110000 out of 1000000 steps (11 percent)
[18:51:53] Writing local files
[18:51:53] Completed 120000 out of 1000000 steps (12 percent)

And could someone give us the formula they're using so we can calculate on our own?
 
Suprchargd said:
Alright installed SMP. Can someone check how many PPD this would come out to:



And could someone give us the formula they're using so we can calculate on our own?

It's pretty easy.

(WU points)*[(1440 minutes in a day)/(time for 1% in minutes * 100)]

For the WU you quoted, it would be roughly 12.5 minutes per frame, so:

(1148)*[(1440)/12.5 * 100)] = (1148)*[(1440)/(1250)] = 1148 * 1.152 = 1322.5 PPD
 
Thanks. Unfortunately it stopped folding at 23%. Saw this post but no solution
I'm going to have to agree with this assessment of the windows smp beta client. As I posted above, the client 'paused' on me the first time I tried it. I reinstalled it before I went to bed and let it run. I wake up to find it sitting there doing nothing at 3% done in a new work unit.

I will likely wait until there's a smp client that isn't flakey before trying again.
 
jws2346 said:
Yeah, it's my understanding come Q3 the Q6600 goes to $266 :p . I think the Q6600 in Q2 goes from like $800+ US dollars to $530 dollars and, correct me somebody if I'm wrong, the server chip x3410 skt 775 (Q6400) goes for about $400 US dollars. I read a post on this forum from a very reliable Team 32 member that the x3410 is probably a better OC'er than the Q6600. I think even at $530 and $400 these are some pretty good deals. :shrug: (remember I'm only talking about folding)


You have to remember when hkepc lists those prices, they are usually based on intel lot sizes of 1000? or are they really msrp? If they are pricing to oem's, retail will be higher if only due to 3 vs 1 yr waranty and packaging differences.

Not sure how they really translate into real prices at retail ... but retailers are going to charge what the market will bear. Only time will tell what the real prices will be.
 
Suprchargd said:
Thanks. Unfortunately it stopped folding at 23%. Saw this post but no solution

Losing the lan connection will cause the client to go idle like this. When it happened to all my win smp clients one day, i stopped the client (ctrl-c), manaully killed the still running mpiexec.exe task with task mgr and restarted the cleints - they all restarted fine from where they were when they went idle.

I have read that if you start the cleint with the lan disconnected, then it will keep going even if the lan connects and disconnects after it is engaged. Folks with laptops and intermittent wireless lan connections were recommending this.

Not sure if this is a useful workaround .... ie will this carry over to the next wu if the client is not restarted?
 
Last edited:
If its any help jws I saw the first post and thought it was completly pointing towards suprcharged as having the problem. Not indicating your numbers were wrong.

JWS=Honest Abe :beer:

Anybody with e6600 including myself is getting 13 min or less per frame. Thats pretty much common knowledge.

I get 12 to 11.75 min frames on mine.

PS I'm not currently folding for Team32 but I wanted to switch over. If I do so, will I lose all the points I've gotten up until now / leave those points with my old team?
Just keep the same name and if you change teams the points will always be carried at stanfords site under the old team and then under the new team but still under your name.

What name are you folding under?

My solution for the win smp client is to use vmware and Linux SMP but thats a whole 'nother can of worms.
 
Last edited:
pscout said:
I have read that if you start the cleint with the lan disconnected, then it will keep going even if the lan connects and disconnects after it is engaged. Folks with laptops and intermittent wireless lan connections were recommending this.

Not sure if this is a useful workaround .... ie will this carry over to the next wu if the client is not restarted?

I'll give it a try. My internet goes out on me 4+ times per day so theres no way I can be around to keep restarting the client. I'll update you if/when I find out if the client will stall after completing a WU or not.
 
yeah, flaky internet is keepin me from win smp right now. im running a d-link gaming router and have found that if i disable "gamefuel" it stops dropping the net connection randomly. G/L
 
Back