G@H

[WaR-ped]

is it possible to run two instances of G@H at the same time so that i can get twice as much done? I can then set one processor to run one instance, and the other, to run the second.

------------------
.:880Mhz Deathtrap:.

• Lite-On FS020 enclosure w/4 120mm Panaflos and soon a 172mm Nidec
• MSI 694D Pro w/ BIOS 1.6
• 2x800E cC0 Pentium 3 w/ 2xVolcanoII
• SyncMAX(NEC) PC166 VCM SDRAM 4x128mb w/ CAS = 1
• Matrox G400MAX 176/219
• And other non-important stuff like hard drives and a dvd drive
• Pineapples

[Jorden]

is it possible your ever went into the FAQs for one fleeting moment and tried to get your answers from there?

[Pace]

Crazy idea! That's up there with survival guide, rules and policies for "stuff to read" by all those Mini's that have been lurking for a while And general manners/intelligence too!

What are FAQs though?

F*ck off
Assholes with repetitive
Questions?

Sorry to be harsh guys, that's the only joke I could think of on FAQ

P.

[WaR-ped]

wow boy... yup, im being repetitive... considering that this forum only has about 1000 posts anyway.

sheesh jorden. lighten up. besides, your two thread FAQ didnt help much anyway. that was not an FAQ about running multiple instances. it was just y'all yammering about what you have tried and a list of links.

so i ask ONE question that could be answered from the FAQ. big deal. its not like this forum is overrun by tons and tons of posts anyway and doesnt warrant a perturbed attitude from you.

[Hecubus]

Way to generate some team spirit guys!

Tim

[rocketmanx]

Yeah guys lighten-up a little.
A simple answer wouldn't have been that difficult. Besides, that's a great way to recruit and -keep- members on our team.
BTW WaR-ped- I just install G@H normally then rename the installed folder Genome@home2 then install G@H again (which installs it to the default directory again) and add a new folder and shortcuts to my programs folder for G@H2. It's really simple actually. Oh- if you run GenomeStealth you will have to install it in both folders and another shortcut for Stealth in G@H2.

[This message has been edited by rocketmanx (edited 10 May 2001).]

[DentyCracker]

I'm interested in seeing if your output (almost) doubles, should show how efficiently genome uses the processor(s)

[Guyver]

There is (IMHO) a better way...

After installing the client the first time, and getting it running.

#1 - Stop the client.
#2 - Copy the client directory (I use Genome1, Genome2, etc...) to another name.
#3 - Uninstall the client (this get's rid of the stuff in the Program Menu area, as well as the installed products list), you can skip this step.
#4 - Copy the directory again (if you removed the original)
#5 - Remove the following files, input, input.inp, and id.
#6 - Place my batch file (the wrapper) into the directory (ies)
#7 - Run the wrapper(s).

I also tend to rename the ghclient.exe to ghclient1.exe, ghclient2.exe, etc...
This allows me to set things like background color, and priorities under Win2K and know for sure which client I'm adjusting.

Hope this helps someone (somewhat)

Guyver

[rocketmanx]

Exactly Guver, that's actually how I do it I just thought the way I explained it was a little more "fool-proof" if ya know what I mean. The production should be better than 1 and a half times as great as the dual p3 750 I'm on now runs a gene in about 12-15 hours roughly and 2 clients will produce 3-4 genes per day (depending of course on AA length)so that's almost double. Don't forget to set priority to normal for each client every time they're started or they will be very slow if you do anything else on your comp and anything higher than normal will probably lock it as it does mine.
Pete

[Jorden]

Alright, I barked too fast too loud. Sorry

Jord.

[Liquid Snake]

On my system, I do what Guyver does, but I don't use his batch file. Does a second instance of Genome automatically use the second CPU?

[Rattledagger]

Liquid Snake, on a idle smp-capable system with two cpus, one instance of genome@home will use both cpus, but totalling 50% of the cpu-power. Two and more instances will use 100% of the cpu-power. Genome@home has no cpu-locking-switch as seti@home has.


WaR-ped, the best is to lock the affinity to the genome@home-instances. This way, no time will get lost with genome jumping to the other cpu. If you're running as batch-file, start two cmd-instances, lock the affinity to these, and run the batch-files from these instances.

Another good thing is to install genome@home as services. If you're using FireDaemon, you can set the affinity here.

As for changing priority, this has negligible effect if you're not running another program like seti@home. Seti@home & genome@home works together, except they must have different priorities if cpu-locked to the same cpu.

[Guyver]

(EDIT: Sorry Rattledagger - posted at same time)

Not necessarily.

If you are running under 2k, go into the task manager, pick one of the ghclients and set the affinity to processor 2, then the other to processor 1.

Repeat if you're lucky/rich enough to have more than 2 processors!

I use my batch job to restart the process in case it dies for some reason other than a major crash.

I am tempted to write the output of ghclient to a file and parse for certain types of messages/warnings, and then have it delete the input, and input.inp files and then restart. This will prevent hours of "Cannot proceed" type problems. (Have to be very carefull with this though)

Guyver

[This message has been edited by Guyver (edited 11 May 2001).]

[Pace]

Yea, sorry again for being harsh, I know that sounded bad

But, I like to encourage FAQ usage, because it keeps Jord happy, and ever boosts my chances of getting a pic of Holly's tits

Either that or he twisted my arm on ICQ to get stuck into a mini and he'd show me half of them

P.