PDA

View Full Version : 05.19.03 it all changes




britboy
May 14, 2003, 01:02 PM
http://www.stanford.edu/~pande/5.19.03/5.19.03.html

Sounds like there's something new on the way ;) Is this the gromacs core?



cubist
May 14, 2003, 01:39 PM
Looks like more "Matrix" hype. Or maybe they're just copying it. Whatever it is, it's pompous and pretentious.

Tomasz
May 14, 2003, 01:42 PM
It looks like gromacs is the streaming word in the film.... but i could just be seeing things.

britboy
May 14, 2003, 02:17 PM
Originally posted by cubist
Looks like more "Matrix" hype. Or maybe they're just copying it. Whatever it is, it's pompous and pretentious.


I somehow doubt that it's anything directly linked to the matrix. 'Reloaded' comes out in the states tomorrow (the 15th), whereas the date from stanford's banner is the 19th.

Someone's probably just having some fun and was allowed a bit of leeway when designing this.

maradong
May 14, 2003, 02:36 PM
I somehow doubt that it's anything directly linked to the matrix. 'Reloaded' comes out in the states tomorrow (the 15th), whereas the date from stanford's banner is the 19th.

Someone's probably just having some fun and was allowed a bit of leeway when designing this.

I hope you are right, i really am not interesting in making a trip through the country to install everywhere a new client.:( But sooner or later it s what s gonna happen.

noel4r
May 14, 2003, 02:57 PM
what a waste of time

bennetsaysargh
May 14, 2003, 03:10 PM
sorry if i sound stupid by saying this, but i don't get it. i understand it has something to do with the matrix, but i don't get the rest. why is 19th?

zimv20
May 14, 2003, 03:20 PM
it's for folding. a similar banner is on my individual folding stats page.

plus, you can see folding@home in the flash thingy linked to in the first post.

so i'm guessing it's a folding software release for that date.

bennetsaysargh
May 14, 2003, 03:24 PM
Originally posted by zimv20
it's for folding. a similar banner is on my individual folding stats page.

plus, you can see folding@home in the flash thingy linked to in the first post.

so i'm guessing it's a folding software release for that date.

ok thanks for clearing that up.:D

Rower_CPU
May 14, 2003, 03:30 PM
We already have gromacs via -advmethods, so I'm not sure this is gromacs related...

MrMacMan
May 14, 2003, 03:58 PM
Originally posted by Rower_CPU
We already have gromacs via -advmethods, so I'm not sure this is gromacs related...

All of them are gromac?

Or maybe they acually made the code better, maybe folding update?

I doubt its folding v3...

BTW, bigger= better.
You can see gromac in there. (http://www.stanford.edu/~pande/5.19.03/teaserfinal.swf)

Rower_CPU
May 14, 2003, 04:03 PM
True, they are not all gromacs. They are randomly assigned.

Actually, it would be v4, wouldn't it?

zimv20
May 14, 2003, 04:17 PM
Originally posted by Rower_CPU
True, they are not all gromacs. They are randomly assigned.


what are gromacs?

Rower_CPU
May 14, 2003, 04:21 PM
Originally posted by zimv20
what are gromacs?

Gromacs is a diiferent core from the standard (Tinker) one that is used by the F@H clients.

It is supposed to enable Altivec processing on Mac clients.

There's quite a bit of discussion about it here (http://forum.folding-community.org/viewforum.php?f=8&sid=d764f11f7b90c80af3077e60893e74ff).

zimv20
May 14, 2003, 04:24 PM
Originally posted by Rower_CPU

It is supposed to enable Altivec processing on Mac clients.


oi. i should get that for my pmac. thanks for the link.

britboy
May 14, 2003, 04:35 PM
Originally posted by Rower_CPU
We already have gromacs via -advmethods, so I'm not sure this is gromacs related...


True, you *can* access gromacs, but it requires a little twiddling. Perhaps the 19.5 release will bring it to everyone else, who perhaps isn't quite as comfortable in terminal.

Or perhaps what you currently get is a beta test, and we're about to see the final released? Whatever though, it'll be nice if this really does bring more power :)

Rower_CPU
May 14, 2003, 04:41 PM
Originally posted by britboy
True, you *can* access gromacs, but it requires a little twiddling. Perhaps the 19.5 release will bring it to everyone else, who perhaps isn't quite as comfortable in terminal.

Or perhaps what you currently get is a beta test, and we're about to see the final released? Whatever though, it'll be nice if this really does bring more power :)

Absolutely. My interest is definitely piqued...

Mr. Anderson
May 14, 2003, 05:53 PM
I'm thinking its the gromac 2

key words in the animation

foldiing@homecoregromaccyou!

2coregromac

Looks like an update using the Matrix reference due to timeliness of the new movie release.

D

scem0
May 14, 2003, 06:22 PM
So basically folding will get Altivec support?

Well that's good. :)

szark
May 15, 2003, 09:58 PM
Originally posted by Mr. Anderson
I'm thinking its the gromac 2

key words in the animation

foldiing@homecoregromaccyou!

2coregromac

Looks like an update using the Matrix reference due to timeliness of the new movie release.

D


You missed a couple of important letters:

sse2coregromac

So definitely optimized for PCs (at least)...

zimv20
May 19, 2003, 11:57 PM
okay, it's 5.19.03. where's my new happy?

Rower_CPU
May 20, 2003, 12:05 AM
Originally posted by zimv20
okay, it's 5.19.03. where's my new happy?

http://www.stanford.edu/group/pandegroup/folding/reloaded/

;)

zimv20
May 20, 2003, 12:20 AM
Originally posted by Rower_CPU
http://www.stanford.edu/group/pandegroup/folding/reloaded/

;)

"the long awaited sequel is finally here."

uhhhh... is it?

i found this faq (http://www.stanford.edu/group/pandegroup/folding/gromacs.html)


As a participant of Folding@Home, is there anything I need to do? No. For the participants of Folding@Home, nothing new need be done, the update is automatic.


that's good, but i'm not sure how to check to see if i've got it, or if i'm even supposed to have it now. hm.

Rower_CPU
May 20, 2003, 12:42 AM
Check your FAHlog.txt and see if you've got "Folding@home Gromacs Core" at the top of the current WU.

zimv20
May 20, 2003, 12:59 AM
Originally posted by Rower_CPU
Check your FAHlog.txt and see if you've got "Folding@home Gromacs Core" at the top of the current WU.

[04:50:30] Folding@Home Client Core Version 2.47 (June 14, 2002)

doesn't look like it. but it d/loaded version 3.25 about half an hour ago. it was 3.24. dunno if that makes a difference.

oh, and i'm running GUI client. don't know if _that_ makes a difference.

MacBandit
May 20, 2003, 01:57 AM
Originally posted by zimv20
[04:50:30] Folding@Home Client Core Version 2.47 (June 14, 2002)

doesn't look like it. but it d/loaded version 3.25 about half an hour ago. it was 3.24. dunno if that makes a difference.

oh, and i'm running GUI client. don't know if _that_ makes a difference.


Hahahahaha. That would be funny if all this hoopla was over the 3.25 version. Maybe it's ther big public announcement over there small update that most of us have had for a few weeks already.;)

britboy
May 20, 2003, 07:02 AM
Originally posted by MacBandit
Hahahahaha. That would be funny if all this hoopla was over the 3.25 version. Maybe it's ther big public announcement over there small update that most of us have had for a few weeks already.;)


I doubt it's something as minor as that. The hype was about a new core, as opposed to a client update. That's why it's an automatic update, and you don't have to do anything about it. Besides, they were talking about SSE etc. optimisations, and the windows client hasn't been updated.

Let's see if our daily figures start rising, now that altivec is apparently supported.....

FelixDerKater
May 20, 2003, 10:46 AM
Got my first Gromacs unit today thanks to mc68k's scripts. Hopefully the AltiVec optimization will speed things up quite a bit.

MacBandit
May 20, 2003, 12:52 PM
Originally posted by britboy
I doubt it's something as minor as that. The hype was about a new core, as opposed to a client update. That's why it's an automatic update, and you don't have to do anything about it. Besides, they were talking about SSE etc. optimisations, and the windows client hasn't been updated.

Let's see if our daily figures start rising, now that altivec is apparently supported.....

Yup, there are now 10 gromacs units to work on instead of just 1. Hopefully this helps.

MacBandit
May 20, 2003, 01:07 PM
I was just trying to comprehend the projects page for Folding@Home.

http://folding.stanford.edu/psummary.html

Here's what I think. I don't think we will see any speed increase out of the Gromacs units because they are more complex units. Just look at the number of atoms in them. The biggest number of atoms in the non-Gromacs units is about 675. Now the Gromacs units start with around 12000 atoms. I think they must be using a G4 Mac to determine the points awarded and time required for these units and therefore we will not see a boost to our scores because of these units because the scale has been changed accordingly. Kind of sucks if you ask me.

britboy
May 20, 2003, 03:51 PM
Originally posted by MacBandit
I was just trying to comprehend the projects page for Folding@Home.

http://folding.stanford.edu/psummary.html

Here's what I think. I don't think we will see any speed increase out of the Gromacs units because they are more complex units. Just look at the number of atoms in them. The biggest number of atoms in the non-Gromacs units is about 675. Now the Gromacs units start with around 12000 atoms. I think they must be using a G4 Mac to determine the points awarded and time required for these units and therefore we will not see a boost to our scores because of these units because the scale has been changed accordingly. Kind of sucks if you ask me.


I very much doubt that's true. All benchmarks are done on a 500MHz Celeron, which has none of the optimisations stanford have recently implemented. I still believe we will start to see an increase in the next week or so.....

MacBandit
May 20, 2003, 04:21 PM
Originally posted by britboy
I very much doubt that's true. All benchmarks are done on a 500MHz Celeron, which has none of the optimisations stanford have recently implemented. I still believe we will start to see an increase in the next week or so.....

I don't know about that. Have you taken a look at the number of atoms in these projects? It seems like they definitely use a different standard for the Gromacs units.

MrMacMan
May 20, 2003, 05:25 PM
Yes they *obviously* didn't use a 500 MHZ celeron because it aint optomized for, anything!

With such an old model... they have to use something new.

MacBandit
May 20, 2003, 06:25 PM
Originally posted by MrMacman
Yes they *obviously* didn't use a 500 MHZ celeron because it aint optomized for, anything!

With such an old model... they have to use something new.

Yeah, if you go to Gromacs.org you will see that they recommend AMD systems. I think that's a hint to what they must use for benchmarking.

MrMacMan
May 20, 2003, 07:21 PM
Originally posted by MacBandit
Yeah, if you go to Gromacs.org you will see that they recommend AMD systems. I think that's a hint to what they must use for benchmarking.


More importantly if you go there you see Dual G4's kick the snot out of everything else on the list, execpt I think it was dual xeons or something...Itanium? I dunno.

MacBandit
May 20, 2003, 09:04 PM
Originally posted by MrMacman
More importantly if you go there you see Dual G4's kick the snot out of everything else on the list, execpt I think it was dual xeons or something...Itanium? I dunno.

Dual Xeons running two processors at twice the MHz of the G4. Also the G4 outpeforms all the AMD on a MHz to MHz basis in those tests too.:D

MacBandit
May 20, 2003, 09:24 PM
Well I have been processing the gromacs unit P909 for most of today. I just ran the Folding@Home Progress program and it says it will be completed by tomorrow before midnight. This is a 38 point unit. That's approximately 1.1 points/hour. Now I just got done doing p638 which is a TINKER unit and it took nearly 8 full 24 hour days. It's a 70.9 point unit which factors out to be about .37 points/hour. So I definitely think there is a speed advantage to the Gromacs units. I still think that we're doing way more work with the Gromac units then we are with the other units though. Also I know that not all units will process at the same rate but I'll take all of the 909s I can get it will definitely help our overall score.

digitalgiant
May 20, 2003, 09:51 PM
the unit hat I am working on now is not a Gromac unit? So I wont see any speed inprovment until my next unit?

MrMacMan
May 20, 2003, 11:09 PM
Originally posted by digitalgiant
the unit hat I am working on now is not a Gromac unit? So I wont see any speed inprovment until my next unit?
Probably not. You need to check here:
http://folding.stanford.edu/psummary.html

BTW, i just wrote 20 min of stuff saying macbandit it wrong, and then safari desided to make this window disappear.
:mad: :mad: :mad:

MacBandit
May 21, 2003, 12:19 AM
Originally posted by MrMacman
Probably not. You need to check here:
http://folding.stanford.edu/psummary.html

BTW, i just wrote 20 min of stuff saying macbandit it wrong, and then safari desided to make this window disappear.
:mad: :mad: :mad:


See even Apple is on my side:D

Hey no really I'll be the first to admit I know absolutely nothing about the processes taken in this whole folding thingamajig.

So what is it I have wrong now?:rolleyes: :)

MacBandit
May 21, 2003, 12:21 AM
Originally posted by MrMacman
Probably not. You need to check here:
http://folding.stanford.edu/psummary.html

What he is trying to say before he went on the tirade is that you may get any one of those units each time the folding program downloads a new one. There is no way to choose which one you get.

britboy
May 21, 2003, 04:04 AM
Originally posted by MrMacman
Yes they *obviously* didn't use a 500 MHZ celeron because it aint optomized for, anything!

With such an old model... they have to use something new.


We'll continue to benchmark the deadline and credits on our 500 MHz Celeron's (w/o SSE).

He who laughs last, laughs best! :p
Link (http://www.stanford.edu/group/pandegroup/folding/gromacs.html)

MacBandit
May 21, 2003, 01:02 PM
Originally posted by britboy
He who laughs last, laughs best! :p
Link (http://www.stanford.edu/group/pandegroup/folding/gromacs.html)

So I guess you didn't expect someone to actually read through all that did you? Are you planning on being a politician? I say this because you cut out a line that appears to support your statements but when read in complete context doesn't really.:p I'm just giving you a bunch of **** I'm not trying to pick a fight.

So after reading the whole page what I got from it is the Gromac units do process a lot more data and because of that we will not be processing points quicker. We will be processing a lot more data but the points given will still be based on the overall processor time given to it. Second is that the only thing they are using the Celeron for it to determine the amount of time given to each computer to finish the processing. That's all it has nothing to do with the points alloted for each unit.

britboy
May 21, 2003, 01:45 PM
Originally posted by MacBandit
So I guess you didn't expect someone to actually read through all that did you? Are you planning on being a politician? I say this because you cut out a line that appears to support your statements but when read in complete context doesn't really.:p

So after reading the whole page what I got from it is the Gromac units do process a lot more data and because of that we will not be processing points quicker. We will be processing a lot more data but the points given will still be based on the overall processor time given to it. Second is that the only thing they are using the Celeron for it to determine the amount of time given to each computer to finish the processing. That's all it has nothing to do with the points alloted for each unit.


I'm still fairly sure that those of us running G4's will see an increase in our stats. Whilst it's true that stats will continue to be based on the CPU time they require, the benchmarks are set on a non-optimised celeron. That means that if you have a celeron, with none of the optimisations (SSE, 3D-Now, Altivec), then your stats will not increase. However, for those of us using G4's, the inclusion of altivec support will mean that units are processed faster, when compared to the non-altivec tinker core.

Also, they clearly state that benchmarks and credits will be set by the celeron.

I'm just giving you a bunch of **** I'm not trying to pick a fight.

Such a shame ;) :D

MrMacMan
May 21, 2003, 05:03 PM
God damn, I wish safari didn't close my damn window... I could have nailed that last night! :(

Small proformance gain, but for ALL.

MacBandit
May 21, 2003, 06:56 PM
Originally posted by britboy
I'm still fairly sure that those of us running G4's will see an increase in our stats. Whilst it's true that stats will continue to be based on the CPU time they require, the benchmarks are set on a non-optimised celeron. That means that if you have a celeron, with none of the optimisations (SSE, 3D-Now, Altivec), then your stats will not increase. However, for those of us using G4's, the inclusion of altivec support will mean that units are processed faster, when compared to the non-altivec tinker core.

Also, they clearly state that benchmarks and credits will be set by the celeron.



Such a shame ;) :D

So I gues that means time will only tell. For my sake I hope I'm wrong we can use all the help we can get.