new bigadv/bigbeta projects

The most demanding Projects are only available to a small percentage of very high-end servers.

Moderators: Site Moderators, PandeGroup

Re: new bigadv/bigbeta projects

Postby 7im » Fri Aug 10, 2012 1:25 pm

I don't know if all these new ones are going out yet, but some are. I've seen the PPD reports. And if on the PSum page, they are being assigned.
How to provide enough information to get helpful support
Tell me and I forget. Teach me and I remember. Involve me and I learn.
User avatar
7im
 
Posts: 14648
Joined: Thu Nov 29, 2007 4:30 pm
Location: Arizona

Re: new bigadv/bigbeta projects

Postby Grandpa_01 » Fri Aug 10, 2012 5:21 pm

The only one that appears to be released so far is the 8102 and it has been released in a limited numbers, Run 0 is the only string that anybody has reported so far.
Image
2 - SM H8QGi-F AMD 6xxx=112 cores @ 3.2 & 3.9Ghz
5 - SM X9QRI-f+ Intel 4650 = 320 cores @ 3.15Ghz
2 - I7 980X 4.4Ghz 2-GTX680
1 - 2700k 4.4Ghz GTX680
Total = 464 cores folding
User avatar
Grandpa_01
 
Posts: 1757
Joined: Wed Mar 04, 2009 7:36 am

Re: new bigadv/bigbeta projects

Postby 7im » Fri Aug 10, 2012 5:59 pm

All three of the new BA projects are listed on the PSum page, so getting assigned, or soon will be. Although no way to tell in what kind of quantity, as those 2 servers also share existing BA projects. Originally, only p8102 was listed, as noted previously.
Last edited by 7im on Mon Aug 13, 2012 2:39 am, edited 1 time in total.
User avatar
7im
 
Posts: 14648
Joined: Thu Nov 29, 2007 4:30 pm
Location: Arizona

Re: new bigadv/bigbeta projects

Postby Grandpa_01 » Fri Aug 10, 2012 6:51 pm

yeah they have been listed for a while but there have not been any reported yet.
User avatar
Grandpa_01
 
Posts: 1757
Joined: Wed Mar 04, 2009 7:36 am

Re: new bigadv/bigbeta projects

Postby cordis » Sun Aug 12, 2012 8:17 pm

Well, somebody over there likes me, just got an 8102! It's pretty awesome, big points! It's running on my dual xeon, it gets 270k-290k ppd on average 8101s, and it's getting around 430kppd on the 8102! Good stuff, thanks a bunch, pande group!
Image
cordis
 
Posts: 68
Joined: Wed May 27, 2009 6:35 am

Re: new bigadv/bigbeta projects

Postby kasson » Tue Aug 14, 2012 12:27 am

Enjoy! Please do keep in mind as we have posted above: 8102 does yield more points than our target, and we do plan to adjust the points yield in the near future. In the meantime...
User avatar
kasson
Pande Group Member
 
Posts: 1906
Joined: Thu Nov 29, 2007 9:37 pm

Re: new bigadv/bigbeta projects

Postby Patriot » Tue Aug 14, 2012 10:48 pm

Could always just fix the target to make up for not benching the units...
Image
Patriot
 
Posts: 89
Joined: Sat Aug 16, 2008 2:04 pm

Re: new bigadv/bigbeta projects

Postby 7im » Wed Aug 15, 2012 1:11 am

If not benched, they'd be worth 0 points, or some other small default value.
User avatar
7im
 
Posts: 14648
Joined: Thu Nov 29, 2007 4:30 pm
Location: Arizona

Re: new bigadv/bigbeta projects

Postby PinHead » Wed Aug 15, 2012 2:00 am

Kasson,

Why do you benchmark a Work Unit, assign points based on the benchmark machine and then reduced the points to a level based upon the fastest machines in Beta instead of sticking with the benchmark machine results? Isn't that the whole and sole purpose of the benchmark machine?
PinHead
 
Posts: 351
Joined: Tue Jan 24, 2012 3:43 am

Re: new bigadv/bigbeta projects

Postby k1wi » Wed Aug 15, 2012 2:39 am

8102 has the same points profile as 8101 so therefore I imagine they brought these out without benching them specifically, expecting them to be relatively similar in performance to 8101. The beta testing (this) has shown that 8102 is actually resulting in higher than anticipated ppd and therefore they are to make adjustments.
k1wi
 
Posts: 1152
Joined: Tue Sep 22, 2009 10:48 pm

Re: new bigadv/bigbeta projects

Postby PinHead » Wed Aug 15, 2012 2:51 am

k1wi wrote:8102 has the same points profile as 8101 so therefore I imagine they brought these out without benching them specifically, expecting them to be relatively similar in performance to 8101. The beta testing (this) has shown that 8102 is actually resulting in higher than anticipated ppd and therefore they are to make adjustments.


Higher than expected results on the benchmark machine/configuration? Or higher than expected results on faster or O.C.'d machines?
PinHead
 
Posts: 351
Joined: Tue Jan 24, 2012 3:43 am

Re: new bigadv/bigbeta projects

Postby Patriot » Wed Aug 15, 2012 3:13 am

k1wi wrote:8102 has the same points profile as 8101 so therefore I imagine they brought these out without benching them specifically, expecting them to be relatively similar in performance to 8101. The beta testing (this) has shown that 8102 is actually resulting in higher than anticipated ppd and therefore they are to make adjustments.


A Didn't benchmark ^...
B SWAG (still didn't benchmark)
C benchmark and see who whines the most and adjust to please the loudest voices... which by clients running /user-base is the minority :)

Random note...these latest units weren't beta tested...and its for the best as they ignore the beta testers anyways...
At-least there is no longer the appearance that they care about releasing wu's that hit "the mark."
Patriot
 
Posts: 89
Joined: Sat Aug 16, 2008 2:04 pm

Re: new bigadv/bigbeta projects

Postby Grandpa_01 » Wed Aug 15, 2012 3:20 am

The only way that Stanford has to bench these that has ever been announced is with a machine that they are not intended to run on. The beta team is where they should be benched if they want a more accurate bench / representation of how the machines that they were meant to be run will run them.

I do not believe any of the bigadv should have been benched on Stanford's bench machine a error of the acceptable +- 15% on Stanford's benchmark can be huge + on a machine that greatly surpasses the minimum requirements. The value given to the WU is correct for Stanford's benchmark machine, but given that there machine is nowhere close to real world capabilities and the (H) factor, it fails at its job of benchmarking these WU's. :wink:
User avatar
Grandpa_01
 
Posts: 1757
Joined: Wed Mar 04, 2009 7:36 am

Re: new bigadv/bigbeta projects

Postby Patriot » Wed Aug 15, 2012 3:23 am

Grandpa_01 wrote:The only way that Stanford has to bench these that has ever been announced is with a machine that they are not intended to run on. The beta team is where they should be benched if they want a more accurate bench / representation of how the machines that they were meant to be run will run them.

I do not believe any of the bigadv should have been benched on Stanford's bench machine a error of the acceptable +- 15% on Stanford's benchmark can be huge + on a machine that greatly surpasses the minimum requirements. The value given to the WU is correct for Stanford's benchmark machine, but given that there machine is nowhere close to real world capabilities and the (H) factor, it fails at its job of benchmarking these WU's. :wink:


And because the smp and bigadv QRB curves are misaligned the benchmark machine puts the gap at the 20% mark which then grows as machines are multiples faster...
Patriot
 
Posts: 89
Joined: Sat Aug 16, 2008 2:04 pm

Re: new bigadv/bigbeta projects

Postby quickz » Wed Aug 15, 2012 6:46 am

Hi, kasson. One thing should be noted that I've only seen Opteron 61xx's benchmark compasion between 8101 & 8102 in this forum, but as we know that Opteron 61xx series were playing relatively bad in folding p8101 (see foldingforum.org/viewtopic.php?f=38&t=21512 for example), however, their performance with p8102 would be much better. So, it would not be a good idea if adjusting points just according to Opteron 61xx's benchmark difference.
quickz
 
Posts: 2
Joined: Tue Aug 14, 2012 7:40 am

PreviousNext

Return to SMP with bigadv

Who is online

Users browsing this forum: No registered users and 1 guest

cron