Have you seen these guys?

N

Nathan_P

[H]ard DCOTM x3
Joined
Mar 2, 2010
Messages
3,490
Looks like someone just bought half of god knows what online:eek:

WTF?

Thoughts over at EVGA are that they are HP testing their new cloud servers.
 
6.5k points per WU average so they are probably just burning though SMP units on about 1000 multicore cpus?
 
Zink said:
6k points per WU average so they are probably just burning though SMP units on about 1000 multicore cpus?
Click to expand...

Sounds about right. Here's another scary thought. They are currently on everyone's threat list apart from our's, EVGA's and OCN's:eek:
 
Looks like it is strait up SMP too.

Wow

They are going to be getting their mow on.
 
They're more than doubling [H]'s PPD. If those points are from 2/4/8p servers, I wonder how many points that one user could drop in a day running bigadv. They put out 40m points yesterday so with bigadv do you think they could hit 100 million? That much PPD on SMP puts that farm on the TOP500 supercomputer listing. It also means they are drawing a couple hundred kilowatts.
 
Nathan_P said:
Sounds about right. Here's another scary thought. They are currently on everyone's threat list apart from our's, EVGA's and OCN's:eek:
Click to expand...
Only because their average hasn't caught up yet
 
Zink said:
That farm it too wimpy to hit 40m ppd. You would need maybe four of those or more. I do get the idea though.
Click to expand...

He said it was an older one.

What if they are testing the Gordon SuperPC thats using straight up SSD's.

Never know!
 
Let's take bets on how long it runs F@H?




...unrelated note what happened with the [H]orde on the 14th?

12.14.11 - ppd 71,217,947 - wu 26,390
 
Kranium said:
Let's take bets on how long it runs F@H?




...unrelated note what happened with the [H]orde on the 14th?

12.14.11 - ppd 71,217,947 - wu 26,390
Click to expand...

productionday.png


Look at what happened a few days before. Some major hardware issues at standford, but they have been resolved and all the missing points have been re-credited.
 
Holy Science Batman!

They will mow us all if they can keep it up. Nothing wrong with being mowed by HP. If they used fah to burn in all of the servers they produce for a few days (one bigadv unit) they would surely mow us all.
 
If you look at the projections, they will reach 3rd place in a year, iirc. Unable to pass 2nd (us) or 1st, not that it matters. I hope they keep folding for a while.
 
Grandpa_01 said:
They appear to be a medical research facility. http://2011.hpcs.ca/
Click to expand...
I'm familiar with that conference. The researchers at that conference do, indeed, have access to some of Canada's top public-sector supercomputing facilities, but those facilities are all well-established so there is no reason for burn in tests at this point. The ones capable of a 50+ Teraflop workload are heavily scheduled as it is, so even if they were burning in some of the available nodes, it wouldn't be enough to cover off this kind of PPD. Besides, that conference was in June.
 
This has inspired me to fire up my X6 while I am out of town over the holidays. Mower blades are being sharpened and will be in use after work today. I should see ~18K PPD out of it if the last time I had it running is still accurate. Whee sooo close to 25K PPD total.
 
Zink said:
http://www.hpcloud.com/
These guys were the one theory, I haven't seen anything for sure though.
Click to expand...

This would definitely be plausible given the numbers they are churning out. Would also make sense about the burn-in idea. Maybe they will use extra processing time to fold also. That would definitely be a huge help to the cause.
 
M0rph3us said:
This would definitely be plausible given the numbers they are churning out. Would also make sense about the burn-in idea. Maybe they will use extra processing time to fold also. That would definitely be a huge help to the cause.
Click to expand...

This is very true, they will help clear the large backlog of SMP work units thar are out there.

Here's something else, there 24hr average is now higher than ours on EOC!
 
nwrtarget said:
Holy Science Batman!

They will mow us all if they can keep it up. Nothing wrong with being mowed by HP. If they used fah to burn in all of the servers they produce for a few days (one bigadv unit) they would surely mow us all.
Click to expand...

You took the words right out of my mouth!
Very impressive numbers, hope they keep it up!
 
As far as the science is concerned I hope they keep it up. I'm ok with dropping down to 3rd place with the amount they are getting done. :)
 
Yeah especially since they're chewing through standard smp units
 
ya im sure stanford is loving the smp. They may not have changed bigadv if this many smp units were getting done lol.
 
Update:

I sent an email out and have confirmed that HP Cloud is indeed burning in some gear.

We're working hard building HP Cloud Services, in fact we are in a limited Private Beta right now. We have been burning in some additional gear and thought we'd lend a helping hand! We will try to continue to contribute as we can. You can find more information about HP Cloud Services at http://hpcloud.com or by contacting our sales team at [email protected]. Thanks for the notice and interest!
Click to expand...
 
That is absolutely astounding!

Glad they have contributed to folding, instead of using some other break in method.
 
Cloud folding, hmmmm. If somebody just had access to their master VM for a quick team number change.... <muhahaha>

Dana
 
These are HP ARM servers I'm pretty sure. My dad works for HP and says these could be Project Moonshot servers.
 
aVaLaNcHe said:
These are HP ARM servers I'm pretty sure. My dad works for HP and says these could be Project Moonshot servers.
Click to expand...

I don't think there is an ARM-based F@H client. And using a non-optimized client wouldn't produce this kind of PPD regardless how powerful the chips are.
 
I just noticed the HP guys and searched the forum to see if anyone else did. Man! They will be #1 in 1.5 years if they keep folding.

Wouldn't it be cool if Amazon jumped in?
 
You must log in or register to reply here.
Back
Top