Supercomputer Week: May 1-7 2012

[Otis11]

4.4ghz on a H100

I just...ah... don't get it. My 2600k at 5.2 wasn't putting out numbers like that!!!

It's gotta be the projects... but that doesn't make sense either.

Consider me stumped.

[masterg]

I just...ah... don't get it. My 2600k at 5.2 wasn't putting out numbers like that!!!

It's gotta be the projects... but that doesn't make sense either.

Consider me stumped.

beats me

huh.jpg

left is return tasks. avg was like 1hr. right is what i woke up to this morning on the 3770k. 100 results out of my total of 161. my HCC seem to be all running high priority

[Bun-Bun]

IVY having a bigger affect on WCG than we though?

Here is my 2600K at 5.0GHz
http://stats.free-dc.org/stats.php?p...hostid=1678266


And on that note... no one made a list totaling everyones score for the week like last year?

[Otis11]

IVY having a bigger affect on WCG than we though?

Here is my 2600K at 5.0GHz
http://stats.free-dc.org/stats.php?p...hostid=1678266

And on that note... no one made a list totaling everyones score for the week like last year?

We've got a list of the top 20, but sadly you're like me... somewhere between 20 and 30.


And yeah... I'm thinking there might be more to Ivy... gonna wait for some more numbers (actually I can't afford to add an ivy atm, so yeah...)

[Bun-Bun]

I meant for just this past SC week.

[masterg]

keep asking me. ill give you what ever info i can. I dont want to shut it down tho

[masterg]

GFAM takes 5 hours =\ really slow and really slowing me down

[Movieman]

GFAM takes 5 hours =\ really slow and really slowing me down

then go to your WCG page, uncheck that project, back to BOINC manager and abort them and send them back..Problem solved.

[masterg]

i did

[Movieman]

i did

Ok, then your golden..

[masterg]

stuff.png

162 tasks, mainly cause GFAM slowed me down. but points also down. same reason?

[D_A]

Some project score better or worse than others, it's not consistent across all WCG projects. I found HCC scores well compared to most others but Say No to Schistosoma if off the charts. This is for Linux of course and may not apply to Windows or Mac. Of the two projects shown above GFAM is returning more per hour (39pph) than HCC (35pph). A bigger sample of GFAM units would help confirm or refute that. So far it looks to me like GFAM is returning BETTER than HCC so if you're just after a higher points output then killing all your GFAM units just shot you in the foot. If your goals are different then it would help to know what you're after.

[masterg]

i suppose. never tried schistosoma

[masterg]

should have replied before movieman. hes feeding bad advice in the loop ;p

[masterg]

Some project score better or worse than others, it's not consistent across all WCG projects. I found HCC scores well compared to most others but Say No to Schistosoma if off the charts. This is for Linux of course and may not apply to Windows or Mac. Of the two projects shown above GFAM is returning more per hour (39pph) than HCC (35pph). A bigger sample of GFAM units would help confirm or refute that. So far it looks to me like GFAM is returning BETTER than HCC so if you're just after a higher points output then killing all your GFAM units just shot you in the foot. If your goals are different then it would help to know what you're after.

hows HFCC and C4CW compare?

[D_A]

If he is it's not deliberate. Like I said, it would help to know exactly what you're looking for in the numbers and what you goals are. If you're trying to get the highest number of completed results then HCC is the way to go. If you after points then possibly not. If you're after the highest return on a specific project then that's different again.

[retro77]

To get steady numbers, crunch one project at a time. But to dedicate your time to more than one project at a time will always very your daily numbers because you don't get the same amount of tasks each different project.

[pfm3136]

My 3770k @ stock (3.7GHz turbo) this past week got an average of 44.5k wcg points with Computing for Sustanaible Water using 85-88 Watts.
No overclock numbers tough, it's on an Asus P8H77-I mini-itx that will be used as a media server on a Fractal Design Array R2 case.
It has a Ocz Agility 3 120GB SSD and 8GB G.Skill DDR3 1600 Ripjaw X kit, so the power numbers at full cpu use, seem very good.

[anubis]

88w? nice!

[Movieman]

88w? nice!

Remember, he's on 220v not 120v like I am..Still impressive numbers..
88W..LOL..Like Sammy that does 8000PPD..

[anubis]

As am I.

pfm3136 - what PSU You're using? Seems like You could use picopsu..

[Movieman]

HMMM.. Say it's 100W here in the USA on 120v..
Damned impressive.
Can someone here in the USA with a 3770K stick a kill-a-watt on it at defaults and also list your PSU and anything else that draws electrical attached?
Also list idle and 100% WCG loading..Thanks

[pfm3136]

I'm using the 300W SFX PSU that comes with the case, it's 80 plus certified.
Also using a Noctua NF-F12 fan on the heatsink (0.6 Watts), and remember that it's a mini-itx H77 board, it will obviously use less power than other boards.

Anyone here knows what's the turbo speed on the i7-3770 non K for the full processor use? (Not the 3.9GHz single thread, but full processor working) Here it's about 40€ between the normal 3770 and the K version.

[anubis]

3770 vs 3770k only differ by open multiplier on "K" model. AFAIK.

[pfm3136]

No, the 3770 has a 3.4GHz speed base, while the 3770K has a 3.5GHz speed base.
That's why i asked if anyone knows the turbo speed for the 3770 full processor use. I wanted to know if it's 3.6GHz or 3.7GHz like the K version.