For the people with enough experience. Would the Q9450 really be that good for the extra money? It won't overclock as good as the Q6600 and I wonder if the 12MB over 6MB cache would make the extra price it's worth.
Printable View
For the people with enough experience. Would the Q9450 really be that good for the extra money? It won't overclock as good as the Q6600 and I wonder if the 12MB over 6MB cache would make the extra price it's worth.
I would like to know the numbers on this one to, I know the cache plays a big role in crunching but just how much?Quote:
Originally Posted by Stijndp
It plays a role, but it's like ~10% if the analysis of average time for WU's is right. For that money you can buy three Q6600's and that would make a big difference. Of course, you might want such a chip for other reasons besides folding :)
10% isn't worth the price premium it comes with :) Thanks for clearing that up.
Is MPI based on Fortran? My history doesn't go back that far. MPI is heavily used in the C, C++, etc, modern cluster space. It's what I use for writing small cluster-based apps.Quote:
Originally Posted by Ugly n Grey
It's not a bad interface, but it is meant for generality, rather than taking the best advantage of any particular OS. The real nut of it all is that it is built for distributed memory systems, not SMP systems like ours. Programming using MPI is usually easier to get your head around than using SMP constructs like threads, locks, etc, so maybe that's why they picked it.
Totally not surprised that the windows MPI port is not ultra optimized. There is an open-source implementation of MPI, so if someone wanted to really put the effort in, that person could optimize the open-source MPI for windows and get a F@H performance boost.
So when you'll be ready to publish it ;) :D Wasn't we talking about that already :shakes:Quote:
Originally Posted by angra
There it should be downloaded :p:
Yeah they've moved it to be fully compatible with everything up to and including Java, so there is lots of it around. I see it in modern clusters though there other effective approaches, what we're really suffering from is a lack of new methodology. This is changing though as multi core systems become a way of life. It will take time, but we'll see it happen. New tools are needed that's for certain!
I agree with you fully on why they made the choice for MPI, I think coding resources at Stanford are quite limited, they needed something they could leverage across platforms and this was it. MPI is HIGH on the list of compatibility items MS would love to drop, trust me, so I'm not sure how many optimizations we'll see for it.
LOL@ distributed memory, I know what you're saying and yeah, if you think back to the 70's when MPI was put into play ... all memory was distributed and non allocated to a processor. You had to code your app to use a quantity of that resource. Processor were virtual for the most part, just based on time slices etc.. an entirely different model but here is MPI still plugging away, pretty wild when you think of it.
As a side note, this is what I do. I do research trying to improve this situation. The idea is to contain the complexity (i.e. cost) of programming high performance computers so that more scientists have access to the resources. Parallelism is everywhere now and parallel programming is inherently anethema to the way that we normally are taught to think about computers. Most of the ways we program parallel systems are band-aids on a broken abstraction model. As you note, there has been progress, but it is still very much an open research question.Quote:
Originally Posted by Ugly n Grey
Well Angra, you Sir are leading the way then. I've been reading up on early approaches to programming for quantum computing which I think it pretty much the ultimate in distributed computing when you think about.Quote:
Originally Posted by angra
The problem is I think the best work is being done at IBM and they are playing their cards close to their chest. I haven't been able to glean much data from their patents lately either. Are you touching on that topic much, even from an observer perspective?
sadly, I don't work in the realm of really basic science. I mostly work with systems that are either in use now or could be in use in the next ~5-8 years. IMO Quantum still has a way to go before it will be in that window.
I do work with some IBM folks, but, as you say, they keep close to the vest :).
I've been getting 3060's lately. Currently @ 3gigs using 1024mb memory it takes 14 min per frame.
Name: p4120_ntl9_folded_Amber03_Native
Tag: P4120R108C19F3
anyone else working on this one right now. it's running on my :banana::banana::banana::banana:ty old laptop that usually does 20-25 minute frames, but for whatever reason this one is taking almost 45 minutes a frame and has been running fr 48 hours already. sounds like something went wrong along the way, no?
Amber is just very slow :) But when I had one I checked it up on their wiki and it seems they want to add SSE2 support in the future, right now it has none.Quote:
Originally Posted by mike8913
Well I got my quads going solid again, stupid :banana::banana::banana::banana:ing weather! For Valentines day I got my mom a E6600 which replaced a P-D 930, I think. Now she wants a computer in her bedroom so I'm going to build a rig around the 930 soon too. I'm coming for you guys!:)
But my family has the flu and I'm going back to bed now.:(
More CPU power always better! Of course, that means you'll be after me too :eek:Quote:
Originally Posted by twiggy
Same here :( Get better!Quote:
But my family has the flu and I'm going back to bed now.:(
Had to convince her the computer would heat her room?:p:
If you have any problems convincing her i would be glad to provide a testimonial.Quote:
Originally Posted by Kingcarcas
heater:)
Hey, you can say that you can turn the thermostat down now :DQuote:
Originally Posted by Kingcarcas
Have had intermitent internet problems today and it looks like I lost 3 wu's :dammit: That is the only bad thing about the smp client, it does not like a bad internet connection.
You could try manually setting your computer's IP. I ended up doing that here at school since school was renewing IP addresses every 4 hours which naturally screwed the SMP up :mad: I let the DHCP assign me the IP/gateway/DNS/etc first, then just manually set my PC to those settings. No more lost or stalled WUs :yepp:Quote:
Originally Posted by [XC] 4X4N
Sorry to hear 4X4N... I lost too 3 wus last weekend :(...
...WE'RE gaining Mr. Pande... time to make visit to The Lab :D
Hey Barnacle whats up? 0 points yesterday. I thought you were catching up to me. :rofl:
And I see your avg ppd is falling too. :(
And one computer has been stuck on an amber core for a couple of days.:cussing:
Hey Team, I'm gonna take a break from F@H for awhile & put my crunching power back onto WCG...I'll be back though... PM me if you need an extra push for some big conquest or something.:up: :up: :up:
Thanks for all you help :up:Quote:
Originally Posted by DAK1640
OK Thanks for the continued support :toast:Quote:
Originally Posted by DAK1640