climateprediction.net home page
Posts by old_user662966

Posts by old_user662966

1) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44416)
Posted 16 Jun 2012 by old_user662966
Post:
This week on the Mystery Twilight Zone Theater 4000 radio broadcast....

"Illusions Of Grandeur"
(A "morning-at-the-computer-with-a-cup-of-coffee" tale, narrated by its author.)

Prologue:

The fish tank and accessories found their way out to the curb on Tuesday for trash pickup. It's algae-stained glass and remnant few gravel bits and musky smell barely hinted at it's barely note-worthy past. Gone, too, a season prior, the wrapping paper and shiny bows in which it all arrived.

Segue:

No one to tell the tale of the rise and fall of little Billy's big birthday wish except for Billy himself, and then, only in passing, "Oh yeah, I had a fish tank once. But then all the fish died and we got rid of it."

And what of the mighty Beta who once and briefly ruled over this small and glorious kingdom? And its minions of lesser fish? And what of the tales of their glorious lives lived out amongst a few plastic sea plants and bubbling Pirate's treasure chest? Who would tell their tales? Lives once blessed with abundant resources of food, fresh air and tank cleanings? Lives which quickly fell into deprivation, starvation and, ultimately, suffocation?

Epilog:

The reflections of the young and mesmerized Billy would no longer be found upon the surface of the tank's glass, his now slightly-older face finding a new home in the glass of a shiny new computer monitor. And, Billy's parents, noticing his fleeting interest, abruptly and quietly stopped financing the upkeep of his 12th birthday gift. Unrecorded in the annals of fish history were the self-important and omnipotent bleatings of an anonymous Beta and its tank-mates of lesser fish.

The end

Music and credits


And now some scenes from next week's radio drama, "Grandeur II: Billy's Face Finds A New Home"...

Preview: Rated PG-13

The computer found it's way out to the curb on Tuesday...Billy's face found a new home on the shiny new screen of a smart phone...Recorded in the annals of research history...

Music

Be sure to join us, we'll "see" you all next week on the radio!

:)
2) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44375)
Posted 12 Jun 2012 by old_user662966
Post:
[Jimmy Gondek wrote:]So you and I want to see CPDN discover that indisputable truth about Climate Change that will finally wake up the world and get humanity to finally start moving towards a sustainable and clean future. And the Researchers are frantically looking for a steady stream of data and person-power and equipment to make those vital discoveries possible.
... it might be a mistake to assume that the academics have any view about climate change at all: for them it's just another interesting problem to solve, albeit a topical one: atmospheric science, oceanic science and who-knows-what other disciplines would be worth studying absent of climate change. Indeed, I suspect that one of the disjunctions between the volunteers and the academics is precisely the question of motive. Your first sentence is true for only some of us, and your second for only some of them. Maybe our teleological knickers are in a twist as well. Ouch!


Idle: Teleological knickers? I thought he was being farcical.
Cleese: No, he was being serious.
Gilliam: I thought he was practicing irony.
Chapman: That was never irony (chomping on pipe)...poofery, perhaps.
Palin: Dont' you mean, puffery?
Chapman: No, poofery. Clearly.
Jones: No such thing.
Chapman: Well there is now.

:)

3) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44374)
Posted 12 Jun 2012 by old_user662966
Post:
...snip...

That may well be true but those who commission the research pay the bills at the moment and for any project, that will always be the bottom line.


And here I've been thinking that CPDN, and any other Grid programs for that matter, were sitting in a quiet corner of the freeware section at the software trade show hoping that someone, anyone, just might pick up a copy of their floppy and give their wares a spin. And to get themselves a little more visibility they decided to huddle under the BOINC tent with the other Grid projects hoping to find a computer to play on. And it's not like this freeware is gonna balance the user's checkbook or straighten out their hard drive or even act as an egg timer, no, it doesn't perform any useful function for the user. It asks potential installers for a free lift down the road with the promise of a handshake for their troubles when the ride is over...hitchhikerware, if you will. And not a cent to help you fill your tank. Paraphrasing Wimpy for their motto, "I'll gladly pay you imaginary credits Tuesday for the coal I burn today." And CPDN is welching on their handshake.

Where you might be seeing fat cats in top hats with 50-dollar bills falling out of their pockets commandeering research-software yachts, I'm seeing earnest researchers on miniscule budgets who, in a million years, could never afford the upfront costs of the tens-of-thousands of computers they need to run their project's software, never-minding the person-hours and energy costs that go along with running something on those scales, standing on a wooden dock next to their rubber dinghy's with a tin cup plaintively sitting by their shoe whilst wearing a cardboard sign offering free handshakes for a lift.

And it's only gonna get tougher for Grid projects to find a ride. Laptop and desktop sales have been flat and are now falling with the advent of smart phones and tablets. The days of using a tethered contraption-in-a-box to check emails is quickly going the way of grandmama's rotary phone as folks migrate to mobile devices and texting. What's a researcher to do? I'm thinking a port to pocket devices running iOS and Android, that hibernate after 30-seconds of non-use, is not a viable future-proofing strategy for saving the planet, finding aliens and cracking encryption codes.

So, I should sleep well at night in comfort and with the confidence that the "fat cats" shouldering a few small percentages of the costs for running their projects will figure this out? I stand by my "upside down thinking" comments. :)


"I don't think these knickers are tight enough, give 'em a yank for me, will you? (hard pull) Oooh, that's about right, now. Thanks, mate!"

(Kindly read above paragraph with Eric Idle intonations)

:)
4) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44350)
Posted 10 Jun 2012 by old_user662966
Post:
Hi David,

ontological knickers


You got me chuckling with that one! Ha! Thanks, I need a good hosing down once and a while! :)

My apologies for not being clearer in my previous post as to whom I was speaking about when I said "management", I was referring to the University's management and not CPDN's.

To further clarify my analogy, and hopefully simplify discussion here...

Corporate Management = University Management
Corporate Project = University Project (CPDN)
Primary Corporate Project Clients = CPDN Clients (Crunchers)
Generated Primary Corporate Project Client Data = CPDN Cruncher Data Results
Secondary Corporate Project Clients (i.e. say, Ad Services, etc.) = Secondary CPDN Clients (i.e. Climate Researchers, etc.)

Now let me try this from a couple of different angles...

To say that the CPDN Researchers are CPDN's Primary Clients and the Crunchers are its Secondary Clients is to argue that the Restaurant's Primary Clients are it's Food Suppliers and it's Secondary Clients are it's Walk-in Customers. What I'm saying is that this is upside-down thinking, and it's the kind of thinking that sinks projects. I stand by my "Bottom Line" comments. :)

Methinks the horse has got one final cautionary tale to tell before it passes on...

"Heaven goes by favor. If it went by merit, you would stay out and your dog would go in." - Mark Twain

Over the 32-years at my own "nickel-and-dime-you-to-death" factory I've seen plenty of worthy projects fall by the wayside and an equal number of worthless projects get the green light. When it comes to getting resources, "worthiness" is not the primary or final deciding factor in loosening the purse strings.

The horse goes on...

So you and I want to see CPDN discover that indisputable truth about Climate Change that will finally wake up the world and get humanity to finally start moving towards a sustainable and clean future. And the Researchers are frantically looking for a steady stream of data and person-power and equipment to make those vital discoveries possible. And a lone professors sits by himself, cloistered in a lonely corner of a vast university campus fretting over how to make his project thrive and, thus, save the world from imminent disaster. And all of this, all of this, falls on the deaf ears of the elder and jaded University Lord Master who is looking out across his well-trimmed lawns on this perfect, sunny day, whose mind is only concerned with the latest cost-cutting measures for the upcoming semester. And no one's attention is ever turned to the marketing folk who have been showing up everyday banging their drum trying to get someone, anyone, assembled in this vast army and effort to listen to their plan on how to make it all work. All the while, many miles away, the ocean floor temperatures continue their incessant rise, degree-by-perilous-degree, and the trillions of tons of trapped methane lying there slowly creep to their liberating boiling point, threatening the future existence of every living thing on the planet save for a few microbes...

Tapping on the microphone. "Ahem, check, check." The room comes to order. "Can I have everyone's attention please. The Project's Marketing folks have something they'd like to share..."

"Yes, it all reads like a soft disaster thriller." gasped the horse, And with a final,"Gaaaah." it expired.

Ta-ta for now!
:)
5) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44341)
Posted 9 Jun 2012 by old_user662966
Post:
This project is something that one of the professors at the University of Oxford is trying to run.
As can be seen from a link here on our front page, he is only a small part of one department of the Uni, with no doubt, limited resources, and possibly limited interest from the board of governors.

The university's bottom line may well be: "We're here to teach. Where can we make budget cuts, as asked for by the government?"

But the credits matter has been raised yet again with the project, and will be discussed next week, taking time away from other matters. Such as the servers which have been shutting down a lot lately, and are currently not accepting data. No working servers, no data for the paying customers.



Hi Les,

"Do, or do not. There is no try." -Yoda

Looks like they've built and paid for their own in-house solution...

Grid computing | University of Oxford Department of Physics:
http://www2.physics.ox.ac.uk/it-services/grid-computing

..., maybe CPDN can invite a few folks from Grid over to that meeting next week so they can add BOINC-fluency to their accreditation's list! Grid's "client-services" can then lobby for more resources to resolve university-project inefficiencies (it's that whole, "nominal upfront investments will yield huge data-and-results dividends" spiel) while CPDN makes the case for the "critical importance" of their data and results and how they're being "under-served" by current university resource priorities.

Seriously, somebody needs to sell the heck out of this to management, though. You folks have any "cheerleaders" on your team or any "allies" over at "Budget" whom you can bring into that room?! Sharks don't get to eat by waiting for some fish to swim into their mouths.

Given the choice between "waiting for the "Server Fairy" to come and leave me a present under my pillow" or "hunting down the right person on staff to come over and show me how to get those money gears turning properly", well, you know what I'd be doing..."Oil can, Nurse!"... :)

Come on! This is the fun stuff!
:)
6) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44333)
Posted 8 Jun 2012 by old_user662966
Post:
Hi Les,

Thanks for taking the time and effort into putting that behind-the-scenes exposition together, it certainly makes clear what is (and is not) going on at CPDN. My sympathies go out to you in more ways than you know...please know that my comments below are not directed at you personally.

Time for a story...

When your IT department makes a call to my company because they are looking into upgrading their server racks, or your service has gone down, my face would be one of the first you would encounter. I know what it is to be the guy behind whom an entire fumbling army was hiding.

...end of story.

In the context of CPDN/BOINC your in-depth explanation is certainly appreciated. However, back in the real world, burdening my customers with the minutia, politics, workloads, incompetence, nepotism, in-fighting, and general clumsiness that lead them to their frustrations at my company's inability to meet their needs or deadlines was never an option. I got paid to solve problems for everything from making sure there were the right screws on the job site to attach your equipment to the racks with to making sure everything was up and running when we signed-off on your fiber loop. I got paid to make things happen, I got paid to make things right. So you'll have to forgive my lack of sympathy with the powers-that-be at CPDN when it comes to resolving user issues! :)

("Jimbo, you're not gonna bottom-line them, are you?" "Yeah, I'm gonna bottom-line them.") So, here's the bottom-line...it appears to me that CPDN management has not quite embraced wholly the idea that the key players on their project are their volunteer crunchers...not their programmers, not their scientists, not their professors, not even the folks who are funding the project...crazy, huh?! No crunchers = no data = no project. Kind of like my job, no customers = no business = no job for Jimbo. Yeah, that's the bottom line rule. I didn't come up with it, it's just the way things are...

So, as I see it, CPDN is in the hemorrhage-control-mode slope of an equalization cycle, whether they're aware of it or not. Too many data cycles to process leading to both a backlog in processing and a drop-off of user participation. When the load gets down to manageable levels they'll (possibly) be able (or willing) to free resources to solve user issues (if the project is still around). The project will find its own equilibrium.

Well, I feel better now, my vacation from crunching for them is, in fact, helping to alleviate stress on their system.

Best!
:)

P.S. I ran an all-volunteer non-profit some years ago, I know what it's like not to be "paid" for one's efforts! We've walked the same mile but in different shoes, Les. ;)

7) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44320)
Posted 8 Jun 2012 by old_user662966
Post:
Hi Dave and David, etal,

I appreciate your altruistic views here, truly and sincerely.

Now, I could go into a whole long drawn-out bit about how I voluntarily pledged to myself to give up both owning and using an internal-combustion-engine automobile for a whole year (and am now into my second year of doing so) and try to impress the world with my real-life acts of consumer greening, but I'll restrain myself here. In fact, it was because of my beliefs and views on the often detrimental impacts of human activity on the global environment that I joined CPDN...I thought it a positive way to help offset the impact of the coal I was burning to generate the electricity to run my machines.

All that aside, we're all participating in BOINC projects. IMHO, this should be a fun activity for one-and-all. In the course of participation user-issues arise. In this case, the accurate and timely reporting of accumulated imaginary credits. (I know, pathetic, really.) And here at CPDN it has been an ongoing issue for the several months I've been here, and the issue, clearly, predates me, as well.

Oh dear, what's a user to do?! :)

Some will say nothing and just keep crunching along hoping a fix will magically appear one day...this is the non-dev-feedback approach to systems problem-solving and usually results in years of stuff not working properly. Some will bring it to the attention of the powers-that-be (who may or may not know a problem exists) in the hopes of a resolution...this is the positive-dev-feedback approach to ironing out bugs, glitches and other system misbehaviors, and it usually results in things getting fixed. I just finished up 32 years in my telecom career and I've got to tell you, I'm a "second-kinda-guy" of the two because that's how things get done. :)

The forums are CPDNs storefront in my analogy above, and what the customers see is a project with data-handling issues. "Hmm, if they can't produce consistent and accurate low-level user-data needs how well are their high-level climate-data needs being handled?!" (Yeah, I know, I'm using my "If the tables at your local restaurant are dirty, one can be fairly sure that the kitchen is not any better." analogy here...sue me.) And for a project like CPDN, in this environment of "climate change poo-pooing", I should think they'd like to allay any concerns of their data's accuracy, consistency and handling. Of course, all of this is just my humble opinion.

Now, in the real world of telecoms with paying clients the average is about 5-out-of-every-100 customers with a service-affecting problem will generate a call-in complaint. And with the advent of customer feedback forums the telecoms discovered that many types of service-related complaints actually went down...a phenomenon of "when one sees that someone else has already filed an identical complaint they feel it's already been reported and the problem will be resolved"...it doesn't mean there are less over-all problems, necessarily, it just means it's harder to judge how wide-spread the problem is based on this smaller sampling...and so, companies (well, those who care about their customers) have had to learn to pay closer attention to their customer's feedback data. Mine is a long-form service complaint. :)

So, just what is my long-form post all about? Simple, I'd like to see CPDN be successful! As of today's stats CPDN has an installed user base of 260,440 folks with an active rate of 19,702, and it has a host-base of 529,508 machines with only 26,769 active. From here...

Overall project stats for Climateprediction.net:
http://stats.kwsn.net/project.php?proj=cpdn

...and one can easily determine for their own pleasure the long-term user participation trends from the graphs found here...

Boinc all Project Stats:
http://allprojectstats.com/po.php?projekt=21

...and I'll be crazy and go out on a limb here and speculate that some of the folks not participating or who have defected have chosen to do so because of the (wait for it...here it comes) credit issue. The only difference between them and me is, they chose not to tell anyone that's why they stopped!

So, I'll be down the street at the place with the clean forks if you need me. Oh yeah, I'll be taking a stroll to the grocery store later, can I bring you back anything?!

Best, as always!
:)
8) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44313)
Posted 6 Jun 2012 by old_user662966
Post:
Credits aren't high on the Too Do list. They're probably way beyond the local event horizon by now.



Hi Les,

I appreciate your candor. It's unfortunate that user credit-issues and accurate member-data reporting are not on management's radar, it doesn't make for a good storefront window-dressing for their project, imho. Oh well. I'll keep checking back to see how things are from time-to-time, in the meantime my machines will continue to be busy with other projects.

Best, as always,
:)
9) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 44296)
Posted 5 Jun 2012 by old_user662966
Post:
...3-month check-in/reminder...same numbers, no change...
10) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43921)
Posted 3 Mar 2012 by old_user662966
Post:
Hi etrecords,

FWIW, since my previous post back in January I tried a few things to see if the credits would catch up on the stats pages...all to no avail.

I checked with a stat board admin (as suggested earlier in this thread by CPDN folks) to see if the problem was on their end and, as my resulting post showed, the problem is with CPDN.

I temporarily stopped processing existing CPDN WUs for a week or so to see if there was some sort of reporting lag, and that didn't solve the discrepancy.

I tried completing all my CPDN WUs and then waiting to see if the credits caught up, but that didn't solve anything, either.

So, here I am (going on two months later) with all the stats sites reporting the same 231,504 for CPDN when the actual number is 243,945...a missing 12,441 credits or, roughly, 5% of my CPDN total...all accumulated in just 4 months of running their software... :(

And we're not alone on this missing credit thing. I've randomly checked other user stats and, while not universal, there are clearly other folks experiencing this problem.

Anyhoo, that's where things seem to be at this point from where I'm sitting. I wish I had more encouraging news.

Best to you,
:)

11) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43664)
Posted 11 Jan 2012 by old_user662966
Post:
To Whom It May Concern,

Perhaps someone at CPDN can drop the folks over at MWAH a line about how they went about solving their recent stat site credit reporting issue?...

Free DC does not report MW credit:
http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=2739

...Just trying to put some problem-solvers together.

:)
Jimmy G
12) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43652)
Posted 6 Jan 2012 by old_user662966
Post:
Hi Les, etal,

On your recommendation I inquired via email with an admin over at one of the stats sites. I described the issue with them and referred them to this thread and their take on the problem was this...


"It's an issue at CPDN for sure. (though not a major one)

There are two places for the stats in a BOINC project. One is the project database itself which is updated in (almost) real time as credit is awarded. On the project site this is what's displayed, usually with perhaps an hour of caching or so. The other place in the exported xml files. These are created periodically by the project, sometimes only once per day, though I try to get most projects to run the job 4 times per day if possible. These files are what my site and other external sites use to update our stats.

It would appear to me that CPDN have a problem in the creation of these xml files being out of date and if I'm reading that thread correctly, digging into my memory I seem to remember that CPDN used to export the xml from a mirrored database, perhaps that mirror is not synched correctly. Only speculation, but whatever the case I checked the xml for some of the reported data in that thread and it's certainly matching what I'm displaying and IS different to what the project itself reports. (I checked Warped's score and yours).

xxxx"


On Fri, Jan 6, 2012 at 9:18 AM, Jimmy Gondek wrote:
Hi xxxx,

Sorry to bother you again, but I thought, perhaps, you could help educate me on a current mystery...

I recently returned to crunching after a several year hiatus and I can see that the landscape has changed a bit...multi-processors, gpu processing, multi-project processing, etc...my question having to do with the latter...

I've been noticing a bit of growing "results lag" on the stats boards with one of my projects (CPDN, in particular) and decided to investigate. After some searching through the various boards and doing some googling for an answer it began to look like an ongoing issue with how CPDN sent out their results for posting with the various stats boards. There have been several thread about this issue over the past couple of years, so it appeared not to be an isolated occurrence. So, upon discovering a recent thread concerning this over at CPDN I posted my own observations/experiences with this issue. At first, it seemed to be an acknowledged problem on their end (as you'll read), but, then that thread was updated and the suggestion was proffered to inquire with the various stats pages about it...which brings me here. Here's the thread in question...

Credit not synchronized on stats-pages
http://climateapps2.oerc.ox.ac.uk/cpdnboinc/forum_thread.php?id=7347#43635


...Any thoughts, insights, avenues for me to investigate, etc. would be gratefully appreciated.

I know, minor stuff in the grand scheme of things, but squeezing what little joy there is to be had with the whole point system disincentivises that whole bit of competitive fun that goes along with performing mundane and repetitive tasks...stroke, boom, stroke, boom, stroke, boom..."Eh? Yer reckon the Cap'n knows where 'e's headin"?"..."'ey! Quiet down down there and keep rowin'! Y'want them "froggies" gettin' t'port with their goods before us?"...stroke, boom, stroke, boom...

Thanks so much for your time on this!

:)
Jimmy Gondek


...Anyhoo, hope this is of some help for you folks!

:)
Jimmy G
13) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43630)
Posted 4 Jan 2012 by old_user662966
Post:
Woo Hooo! I just cracked 200,000 on CPDN aaaand broke 1,000,000 combined all in one day! Yay, me! (Happy dance!)...I'm crunchin', I'm crunchin'...Hoo, hoo...

...huh?...what?...what do you mean you can't see that?...let me take a look...WTH?...only 188,542 CPDN and, what?, 987,474 Combined???...no, that's not right, come here and look at my computer...whadddayamean you can't?...no, really...mine says 200,983.33 CPDN and 1,005,257 Combined right here in BOINC Manager...really, you've got to believe me...no, wait, come back...why would I lie?...that's what it says...come on, happy dance, happy dance...party at Jimbo's...wait...

...sigh...

:\

14) Message boards : Number crunching : More FPU or Integer Power needed? (Message 43617)
Posted 28 Dec 2011 by old_user662966
Post:
your calculation is not correct i think.
You have to multipy results with the core-number:
Integer:
11954 (4core)= 11954x4=47816
6690 (8core) = 6690x8=53520 (some little increase)

FPU:
4360 (4core) = 4360x4=17440
3767 (8core) = 3767x8=30136 (giant increase, same TDP)

It seems Intel SMT could boost your calculating power, especially FPU-calculations got most profit of SMT.

But this is theoretical Benchmark, and how we know, HadCM3N-Models 'hate' Intel-SMT.


Maybe, the best Benchmark for CPDN is the 481.wrf at the SPEC CPU-Benchmark. http://www.spec.org/cpu2006/Docs/481.wrf.html

Intel's I7 3960x generates 233 wrf-Points! Best Value ever for signle-CPU.
http://www.spec.org/cpu2006/results/res2011q4/cpu2006-20111121-18924.html

Otherwise i think, CPDN-Executes doesn't use all Compiler-optimizations for modern and different CPU-types. Is there any outlook to run the models on nVidia's TESLA-Chip?


Hi ojum-le,

I think we're making the same point, though I didn't follow through with the core multiplications as you have to get the total core benchmark performance. But, methinks there's still one more comparative computation that needs to be made...

Had there been no hyperthread benchmark penalties on either Integer or FPU computation rates we would result with the following numbers...

Integer:
11954 (4core)= 11954x4=47816
6690 (8core) = 6690x8=53520 (some little increase)
11954 per core with no hyperthread penalty would = 11954x8=95656 (Again, 53520 = ~44% performance hit)

FPU:
4360 (4core) = 4360x4=17440
3767 (8core) = 3767x8=30136 (giant increase, same TDP)
4360 per core with no hyperthread penalty would = 4360x8=34880 (Again, 30136 = ~14% performance hit)

So two side-by-side i7 2600K @ 3.4GHz machines running 4 WUs each will outperform one i7 2600K @ 3.4GHz running 8 WUs based on these results. Though I should think that based on a cost ratio between the two setups that it would be less expensive per WU to do 8WUs across one machine vs. two... :)

Expositions of any mathematical errors or poor logic on my part are always welcomed!

:)
15) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43611)
Posted 26 Dec 2011 by old_user662966
Post:
Hi Iain,

Would you kindly post a link to that "relevant thread on the other message board" to which you refer? It'll help keep the issue from getting too scattershot across the forums...

:)
16) Questions and Answers : Macintosh : Why do jobs pause every 30-40 seconds? (Message 43610)
Posted 26 Dec 2011 by old_user662966
Post:
Hi Richard,

I noticed the same behavior in Activity Monitor on my iMac when I began running this project, too. It appears to me to be what happens everytime the CPDN/BOINC software performs a "checkpoint" of each of the Work Units. If you go into the Advanced menu in BOINC Manger and select "Event Log..." you'll be able to view what is going on in real time.

:)
17) Message boards : Number crunching : More FPU or Integer Power needed? (Message 43595)
Posted 22 Dec 2011 by old_user662966
Post:
Thank you, everyone, for your feedback here, it's certainly helped me with my understandings of a few things going on here. :)

However, my main question still remained unanswered, Does using CPDN using 4 actual cores vs. 4 actual cores + 4 virtual cores cause the Integer benchmarks to be halved, or otherwise affected? Thank you, Les, for having me consider that this might be a BOINC thing instead of a CPDN thing...you prompted me to do an experiment I was loathe to do earlier because of the fragility of BOINC Manager 6.12.35 (x86). (I found that messing with the preferences was causing random appearances of the dreaded Waiting for GPU Memory WU status glitch!)

So I decided to brave messing with my preferences to try and determine whether this Integer Benchmark-halving effect was being caused by actual and virtual core usage. Here are my results...

BOINC Manager set to 50% CPU usage, no projects running...

Thu Dec 22 08:05:37 2011 | | Number of CPUs: 4
Thu Dec 22 08:05:37 2011 | | 4367 floating point MIPS (Whetstone) per CPU
Thu Dec 22 08:05:37 2011 | | 11954 integer MIPS (Dhrystone) per CPU

BOINC Manager set to 50% CPU usage, 4 CPDN WUs running...

Thu Dec 22 08:06:38 2011 | | Number of CPUs: 4
Thu Dec 22 08:06:38 2011 | | 4360 floating point MIPS (Whetstone) per CPU
Thu Dec 22 08:06:38 2011 | | 11868 integer MIPS (Dhrystone) per CPU

BOINC Manager set to 100% CPU usgae, 4 CPDN WUs running...

Thu Dec 22 08:08:18 2011 | | Number of CPUs: 8
Thu Dec 22 08:08:18 2011 | | 3762 floating point MIPS (Whetstone) per CPU
Thu Dec 22 08:08:18 2011 | | 6683 integer MIPS (Dhrystone) per CPU

BOINC Manager set to 100% CPU usage, 8 CPDN WUs running...

Thu Dec 22 08:12:21 2011 | | Number of CPUs: 8
Thu Dec 22 08:12:21 2011 | | 3767 floating point MIPS (Whetstone) per CPU
Thu Dec 22 08:12:21 2011 | | 6697 integer MIPS (Dhrystone) per CPU

And, finally,
BOINC Manager set to 100% CPU usage, 0 CPDN WUs running...

Thu Dec 22 08:37:29 2011 | | Number of CPUs: 8
Thu Dec 22 08:37:29 2011 | | 3764 floating point MIPS (Whetstone) per CPU
Thu Dec 22 08:37:29 2011 | | 6696 integer MIPS (Dhrystone) per CPU


It appears to me from these results that both Floating Point and Integer performance are directly affected by BOINC Manager CPU settings regardless of project WU load or (as these numbers now match more closely those of mo.v's identical CPU) OS platform...

Floating Point went from 4-core @ 4367 to 8-core (4-core + 4-virtual core) @ 3767 (Whetstone) an ~16% performance hit per core.

Integer went from 4-core @ ~11,900 to 8-core (4-core + 4-virtual core) @ 6696 (Dhrystone), an ~44% performance hit per core.

Any and all �feedback welcomed, along with pointing out any errors in my methods or results!

Hope this all adds to the original discussion on FPU or Integer Power?

:)
18) Message boards : Number crunching : More FPU or Integer Power needed? (Message 43589)
Posted 21 Dec 2011 by old_user662966
Post:
Well, astroWX, therein lays my problem in understanding this...I would have assumed that since mo.v and I are using identical processors, and that my Integer benchmark is pretty much half of his, that CPDN was Integer-intensive and not Floating Point-intensive...or possibly that the reason my benchmarks were showing up as half was because the ALU had to split time, cycles and circuitry between the work of the actual core and the associated virtual core...that is, if he was only running 4 WUs and me, 8WUs. As it is both our Floating Point benchmarks are in the same ball park.

Thanks for your help, but I'm thinking that maybe I'm still missing some other obvious point here about the discrepancy in our Integer benchmark differences. Any thoughts anyone?
19) Message boards : Number crunching : More FPU or Integer Power needed? (Message 43583)
Posted 21 Dec 2011 by old_user662966
Post:
mo.v wrote:

I'm pleased so far with my new i7. I was expecting it to have four cores but was surprised to find eight.

<snip>



Hi mo.v, etal,

After looking at your setup something shocking was staring me in the face...I'm using the same processor in an iMac and my Integer benchmarks are nowhere near as high as yours. I'm curious whether you are using just the four actual cores only on CPDN or are you using the virtual cores as well...i.e. 4 WUs at a time or 8 WUs?

My iMac i7

Your i7
Measured floating point speed 3282.91 million ops/sec
Measured integer speed 12205.76 million ops/sec

My i7
Measured floating point speed 3760.46 million ops/sec
Measured integer speed 6692.85 million ops/sec

I run 8 simultaneous CPDN WUs. I'm thinking these measurements are based on per-core performance, perhaps someone can shed some light on this for me?

Thanks! :)
20) Message boards : Number crunching : Credit not synchronized on stats-pages (Message 43574)
Posted 19 Dec 2011 by old_user662966
Post:
...Just thought I'd chime in about the credit lag...don't know if it has been addressed by the project folks yet, but the issue continues on my end, FWIW...somewhere in the 5,000-10,000 range depending on the day...

...best wishes to all for a healthy and safe holiday season!... :)


Next 20

©2024 climateprediction.net