• 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.

600 Pointer!!!!!!!!!!!!!!

May your boxen be graced by the presence of many more... and if they are, I feel it my responsibility to let you know that you have to change yout config so they fold for me! That's p, bracket, capital H, .... you get the idea! :D

 
I felt the same way when I got my first 600 pointer.
To bad it was on a 1.7 Celeron

I think all my machines, except the 500mhz HP have gotten at least one now.


 
I have to many computers to realize which ones have/had 600pointers...
 
darktiger said:
I have to many computers to realize which ones have/had 600pointers...
Click to expand...

Send them to me. I can "watch" them for you.
 
I am sad now.
My A64 used to just fold these 24/7 for 2 months. Now I am doing 239 to 241 pointers. :(
At least I am taking solace in the fact that hopefully these 241 tinkers possibly help cure a really bad disease. ;)
 
magnusvir said:
I am sad now.
My A64 used to just fold these 24/7 for 2 months. Now I am doing 239 to 241 pointers. :(
At least I am taking solace in the fact that hopefully these 241 tinkers possibly help cure a really bad disease. ;)
Click to expand...

At least you didn't just get a 46 point gromacs. This one is taking slightly less than half the time of a 241 point tinker, but pays about 1/5 the points. :mad:

 
I've been getting those little gromacs in about a 2 to 1 ratio to the tinkers. It's almost like spending more time up and downloading than it is folding... :rolleyes:

 
What's so great about these 600 pointers?
Or is it just on Intel?

I opened up EM when I got home from work today, and saw a 600 pointer on one of my CPUs.
I thought, "Awsome, this is what everyone has been talking about!"
Then I did a calculate points, and it says 82ppd.
It's a p1141_RIBO_FSpeptide_HEL_nospring
It's taking 1h 44m 59s per frame on a 1.6 Opteron.

Actually I was first excited to see all black dots on my other CPU, because I know they're good.
This one's a p1477_release31_0
I get 342ppd at 15m 18s per frame.

The 600 pointer is really slowing me down, as I get ~130ppd on the 241's.

Is this a new SLOW 600 pointer, or is something wrong?
 
I've a 1141 taking 51:40 per frame on a Sempron 2200+ stock.
EMIII report 167ppd, so that's about 100ppd/ghz.

At the same time I've a 1477 taking 27:22 per frame on a stock 1.13 p3 laptop (m?).
EMII reports 192 ppd, or about 180ppd/ghz.

just data points.

 
last time I had those 600 pointers they were giving me about 290 ppd.
Not as fast as those p14xx ones, when I had those they were close to 350 ppd. :eek:
 
lol, i used to have both of my boxen with 600 pointers for about a month.. My a64 did them in 45min each step.

i wish i had some more of these
 
My recommendation would be to use EMIII, it has a built in protein calculator that will determine your PPD/PPW on the fly, based one what you're currently folding.

 
I would love to get hold of some more of those 364 pointers on my mobile XP and Duron boxen. They tore through those things in no time.

It's actually kinda funny though since the Duron box will put out almost the same PPD with the 239 and 241 point Tinkers than it does with the bigpacket Gromacs. My mobile XP rig prefers the bigpacket Gromacs though. The Duron and XP2400@ 2.2 Ghz boxen will tear through Amber cores also. Since I don't have bigpackets enabled on the borged XP2400 I always hope it gets the Amber cores since it gets its best PPD from those.

Also, most the time, I'll get better PPD out of big Tinkers on the Preshott 2.8 at work than I will get from Gromacs other than bigpackets. I don't pull much over 100 PPD with Gromacs while I'll get between 115-125 PPD with Tinkers.

 
badasspenguin said:
How do you figure out PPD?
Click to expand...


You can also do it manually. Take the amount of time it takes to complete one frame: (For this example we will say 12 minutes)

60min/hour / 12min/Frame = 5 Frames/hour

Find out the number of frames and the total points of your Protein (We'll use P972 for this example):
http://vspx27.stanford.edu/psummary.html

P972: Double Gromacs, 122 Points/WU, 100 Frames/WU

122 Points/WU / 100 Frames/WU = 1.2 Points/Frame

5 Frames/Hour * 1.2 Points/Frame = 6 Points/Hour (PpH)

6 Points/Hour * 24 Hour/Day = 144 Points/Day (PpD)


 
This topic has been deemed way too ontopic by the topic police. :eek:

Need more pix plz!

and yea.. 64 pointers suck...
 
Hito Bahadur said:
You can also do it manually. Take the amount of time it takes to complete one frame: (For this example we will say 12 minutes)

60min/hour / 12min/Frame = 5 Frames/hour

Find out the number of frames and the total points of your Protein (We'll use P972 for this example):
http://vspx27.stanford.edu/psummary.html

P972: Double Gromacs, 122 Points/WU, 100 Frames/WU

122 Points/WU / 100 Frames/WU = 1.2 Points/Frame

5 Frames/Hour * 1.2 Points/Frame = 6 Points/Hour (PpH)

6 Points/Hour * 24 Hour/Day = 144 Points/Day (PpD)


Click to expand...

But please also put in the Ghz of what you are running it on, so we can see ppd/ghz.
Otherwise it reminds me of a story:
Back during the first gas crisis in 1973, this guy in a BIG Cadillac pulls up to a stoplight next to a kid on a Honda 50 motorcycle. He rolls down his window and asks the kid "How many miles per tank do you get?"
Kid replies "About 300"
Guy turns to his wife and says "Heck, this Caddie gets as many miles per tank as that scooter"
Click to expand...
 
I was assuming he wanted the calcs for internal use. Per Ghz is also a pain due to the difference in Ghz effectivenes based on the processor.
 
I remember another protein that my systems loved to chew on. That would be the little 20 point Tinkers I was getting several months back. Half the time its seemed like something was wrong considering it only took seconds to complete each frame. I was getting a great PPD average when running these things and it also allowed for instant gratification.

 
SmokeRngs said:
I remember another protein that my systems loved to chew on. That would be the little 20 point Tinkers I was getting several months back. Half the time its seemed like something was wrong considering it only took seconds to complete each frame. I was getting a great PPD average when running these things and it also allowed for instant gratification.

Click to expand...

I got one yesterday on a 400mhz pentium II. 3 hours per frame (only 20 frames though).
 
Now this is interesting.
After finishing the 364 pointer today my second CPU picked up a 600 pointer.
A p1140_RIBO_FSpeptide_EXT_nospring

opposed to the one still on my first CPU.
A p1141_RIBO_FSpeptide_HEL_nospring

Now only one frame has finished, so it might change a bit, but the second CPU is doing a frame in 42m 37s, opposed to 1h 45m 41s on the first CPU.

It started over 2 days later, and will finish 1.5 days earlier.
202 PPD opposed to 81 PPD.

Does anyone know what the HEL and EXT mean?
I'd like some more EXTs please.

I guess at least the 600 pointers are coming back, YAY!
 
I would suggest you make sure SSE boost is enabled on the protein that seems to be taking a long time. It would also probably be good to look in task manager and see if something else is eating up CPU cycles. Those would be the main things I can think of right now that would be a problem.

I have always had similar times on all of the 600 pointers I've gotten. Hell, I'd like to have a few of them right now.

 
Woke up this morning and found 2 600pointers in my systems that sit right next to each other.
 
SmokeRngs said:
I would suggest you make sure SSE boost is enabled on the protein that seems to be taking a long time. It would also probably be good to look in task manager and see if something else is eating up CPU cycles. Those would be the main things I can think of right now that would be a problem.
Click to expand...
How would I go about checking that SSE boost is enabled?
I did check the Task manager, folding is using 50% like it should be.
 
Check you FAH log. Look for the last time the program started. It should say something like:

[20:08:12] *------------------------------*
[20:08:12] Folding@Home Gromacs Core
[20:08:12] Version 1.80 (March 16, 2005)
[20:08:12]
[20:08:12] Preparing to commence simulation
[20:08:12] - Looking at optimizations...
[20:08:12] - Files status OK
[20:08:16] - Expanded 862122 -> 11078081 (decompressed 1284.9 percent)
[20:08:18]
[20:08:18] Project: 1322 (Run 0, Clone 404, Gen 6)
[20:08:18]
[20:08:22] Assembly optimizations on if available.
[20:08:22] Entering M.D.
[20:08:43] (Starting from checkpoint)
[20:08:43] Protein: p1322_1fnt_a20_9.1ps-1_uf
[20:08:43]
[20:08:43] Writing local files
[20:08:43] Completed 1300000 out of 2000000 steps (65)
[20:08:44] Extra SSE boost OK.


With Tinkers you probably won't see this. But when you load FAH it checks to see if there was a bad end the last time and will turn off SSE support if there was a problem. You can force it on with the -forceasm flag.
 
Lucky you. I want those. I have had a few high Pointers. And them some 2500 frame 55/s Frame.
 
3 of 'em showed up, including one to the bitchy A64, so we'll see what happens.

Aside from my production graph looking like the Von Trappes' backyard profile for the next few days... :D
 
Hito Bahadur said:
Check you FAH log. Look for the last time the program started. It should say something like:

<snip>

With Tinkers you probably won't see this. But when you load FAH it checks to see if there was a bad end the last time and will turn off SSE support if there was a problem. You can force it on with the -forceasm flag.
Click to expand...
Ok, I just got home, and CPU1 does not have SSE boost, but CPU2 does.
I have the -forceasm flag on both, so what else can I do?

actually, you can't get these protines without -forceasm on, can you?
 
I get about 27 min per frame on 600 pointers on a 64 939@2.5 gig. They were never meant for slower machines. If your in this for the points and have a slower machine then by all means disable big units. Your costing yourself production if you don't.



 
Ok, the power went out last night and now both CPUs are taking about 1h 45m per frame.
How do I re-enable SSE boost?
 
You'll need to restart the clients using the -forceasm switch. That'll force each client to use SSE upon every startup.

 
I guess it's like Hito said, if there was an error shutting down the core previously it turns SSE boost off.
Looking at the logs I see this is what matters:
Code: 
[23:30:49] Preparing to commence simulation
[23:30:49] - Ensuring status. Please wait.
[23:31:06] - Looking at optimizations...
[23:31:06] - Working with standard loops on this execution.
[23:31:06] - Previous termination of core was improper.
[23:31:06] - Going to use standard loops.
[23:31:06] - Files status OK
[23:31:09] - Expanded 2966018 -> 16166417 (decompressed 545.0 percent)
[23:31:11] 
[23:31:11] Project: 1141 (Run 45, Clone 22, Gen 8)
[23:31:11] 
[23:31:24] Entering M.D.
[23:31:48] (Starting from checkpoint)
[23:31:48] Protein: p1141_RIBO_FSpeptide_HEL_nospring
[23:31:48] 
[23:31:48] Writing local files
[23:31:48] Completed 118785 out of 250000 steps  (48)
[00:22:43] Writing local files
VS.
Code: 
[11:36:19] Preparing to commence simulation
[11:36:19] - Looking at optimizations...
[11:36:19] - Files status OK
[11:36:24] - Expanded 2966018 -> 16166417 (decompressed 545.0 percent)
[11:36:25] 
[11:36:25] Project: 1141 (Run 45, Clone 22, Gen 8)
[11:36:25] 
[11:36:30] Assembly optimizations on if available.
[11:36:30] Entering M.D.
[11:36:53] (Starting from checkpoint)
[11:36:53] Protein: p1141_RIBO_FSpeptide_HEL_nospring
[11:36:53] 
[11:36:54] Writing local files
[11:36:54] Completed 135000 out of 250000 steps  (54)
[11:36:59] Extra SSE boost OK.
The second one is the same thing Hito posted, I'm just showing the difference for anyone else who might have had this problem.
If it sees there was an error the last time it ran, it uses standard loops, which I guess turns off SSE boost.
So -forceasm doesn't actually force it to use SSE, only if there wasn't an error previously.
At least now I know, "And knowing is half the battle." :p
 
You must log in or register to reply here.
Back
Top