Well at pricewatch the prices for combos have finally dipped under $300.
Are any of these looking good?
I'm in need of more heat.
Are any of these looking good?
I'm in need of more heat.
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Sorry, don't see anything under $310 there...
I'd say it's a gamble, since I don't know of my 945 boards that can take G0 stepping.
[BRO]Alaskan;1031647030 said:Spend $300 and have a stock 2.4ghz quad set up right that can put out ~2800PPD or spend $420 (P35 MB, High end cooler, and a GO Q6600) and overclock to 3.4ghz (not much trouble) and get 4000PPD. Points per dollar is close. Points per watt will prob go to the overclocked unit.
Now if you could get a cheap set up and only clock it to 333mhz FSB for a 3.0ghz chip that would be a real winner
Bro, how are you getting 2800PPD out of a stock 2.4ghz Q6600? I've got 2, one at 2.7 and one at 2.8ghz and they only put out about 2450PPD.
[BRO]Alaskan;1031648487 said:Download VMware server and get their free serial number. This limited version of VMware is free and it allows you to to run a virtual O/S inside of your original O/S but you are limited to utilizing 2 cores. By running 2 linux O/S's through VMware under windows you get a higher CPU utilization (~25%). I usually get 2 WU's done every 21hrs- 24hrs(remember I am running 2 at a time)(~12:30Min/frame-14min/frame) for about 4000PPD on both of my 3.1ghz Quads with 2 gig of ram. You will nee 2 gig of ram for the besause you need to dedicate at least 512meg (I use 600meg)to each virtual Linux machine. Try it out. Its easy to set up. Free. and you might gain a few points. Just for fun I will set up a quad at stock speed and see how many points I can really get but it should be close to 2800PPD. Or see the above post that clicked in before me
Gotcha. I really really don't want to have to mess with Linux, so I'll just stick with my slower Windows version.
[BRO]Alaskan;1031648487 said:Download VMware server and get their free serial number. This limited version of VMware is free and it allows you to to run a virtual O/S inside of your original O/S but you are limited to utilizing 2 cores. By running 2 linux O/S's through VMware under windows you get a higher CPU utilization (~25%). I usually get 2 WU's done every 21hrs- 24hrs(remember I am running 2 at a time)(~12:30Min/frame-14min/frame) for about 4000PPD on both of my 3.1ghz Quads with 2 gig of ram. You will nee 2 gig of ram for the besause you need to dedicate at least 512meg (I use 600meg)to each virtual Linux machine. Try it out. Its easy to set up. Free. and you might gain a few points. Just for fun I will set up a quad at stock speed and see how many points I can really get but it should be close to 2800PPD. Or see the above post that clicked in before me
[BRO]Alaskan;1031650692 said:Ram is cheap. I just bought 2 gigs (2x1gig) for $50 with no rebates.
[BRO]Alaskan;1031650692 said:Do you assign affinity for your smp in windows? I set mine for each instance of Linux (I prob don't need to)
I don't lose as much time if the server won't send out a new WU or whatever.
When I upgrade my "desktop" to a quad I'll prob go the same route. R%ight now my quads are dedicated so responce is not an issue. They seem to open web pages fine though
[BRO]Alaskan;1031664934 said:When I upgrade my "desktop" to a quad I'll prob go the same route. R%ight now my quads are dedicated so responce is not an issue. They seem to open web pages fine though
Only when the coders realize that they need to reright their code to use more than 1 core
Noooooooooooo. Get a quad core. Any quad core. Do it now.OK you peeps - you hijacked my thread.
Because of this I ended slamming my $$ down on a 400mhz linux machine, which may not fold worth a darn.
http://www.nokiausa.com/A4626058
Sorry if this isn't the best place to ask this, but what's a reasonable PPD expectation for a single SMP instance on a Q6600 running natively on Linux?
Would it be better to run two instances?
What about a single instance on a C2D?
Sorry if this isn't the best place to ask this, but what's a reasonable PPD expectation for a single SMP instance on a Q6600 running natively on Linux? Would it be better to run two instances? What about a single instance on a C2D?
Sorry, haven't kept up with it lately...
[BRO]Alaskan;1031648487 said:Download VMware server and get their free serial number. This limited version of VMware is free and it allows you to to run a virtual O/S inside of your original O/S but you are limited to utilizing 2 cores. By running 2 linux O/S's through VMware under windows you get a higher CPU utilization (~25%). I usually get 2 WU's done every 21hrs- 24hrs(remember I am running 2 at a time)(~12:30Min/frame-14min/frame) for about 4000PPD on both of my 3.1ghz Quads with 2 gig of ram. You will nee 2 gig of ram for the besause you need to dedicate at least 512meg (I use 600meg)to each virtual Linux machine. Try it out. Its easy to set up. Free. and you might gain a few points. Just for fun I will set up a quad at stock speed and see how many points I can really get but it should be close to 2800PPD. Or see the above post that clicked in before me
If you're running a 32-bit OS, you won't have a full 4GB to work with, even if you have that much RAM installed. It's a situation that dates back a few decades, and boils down to the fact that all your computer's I/O is mapped out of the memory address space, and therefore those addresses aren't available as RAM. Dan's Data has a rather thorough article on the subject. So if you have 4GB of RAM and a 512MB video card in your machine, you'll have access to something less than 3.5GB of RAM.
If you're running a 32-bit OS, you won't have a full 4GB to work with, even if you have that much RAM installed. It's a situation that dates back a few decades, and boils down to the fact that all your computer's I/O is mapped out of the memory address space, and therefore those addresses aren't available as RAM. Dan's Data has a rather thorough article on the subject. So if you have 4GB of RAM and a 512MB video card in your machine, you'll have access to something less than 3.5GB of RAM.
The size of the video card effects what can be addressed, AFAIK. It cuts into the 3.5 or whatever is available in 32 bit.
I did not know that but I don't think that's the case here because it's a texture memory and not something an OS will really use?