Genome @ Home

Genome@Home along with Folding@Home is run by PandeGroup at Stanford.

The goal of Genome@home is to design new genes that can form working proteins in the cell. The 4 main applications are: Engineering new proteins for medical therapy, designing new pharmaceuticals, assigning functions to the dozens of new genes being sequenced every day, & understanding protein evolution.

Genome@Home project and client download

Team Anandtech stats

Client-setup:
Download & install to whichever directory you want...
Run ghclient.exe
On the 1st run you'll have to input a user Name: Case sENsitiVe...choose whatever you want
Team account number - 604374405
Server - leave at default if you don't need a proxy-server. There's currently little support for this.
Port - probably leave at default. Note, Genome uses port 10100, and this must therefore be open somewhere.
Firewall- yes/no... if yes, you'll have to input some more info.
If there's no problem the client will now download a wu from Stanford and start crunching...

Make sure the info in ghclient.cfg is correct, if something is wrong you can edit it here. The important things are :-
email=YourUserName
account=604374405
Note in Genome, all results returned to a username/group will stay with this group & username forever. Mistakes in account-number will either be credited to another group or to no group at all.

Normal running
Just let it keep on crunching. Normally, it will upload/download when done with a WU. If for any reason it can't connect, it will retry a couple of times, and after about 6 minutes restart the current WU.

Genome will on every restart first upload/download. If you've got a partially done WU it will download an extra one if you've not already got a spare WU. Afterwards it will Filter. This is just a pre-process on the WU, where it counts up from 1 to aa. (see below for aa)
Now the actual prosessing starts. It will count 10%-20% ... 100% and afterwards,sequence nn completed. A WU contains 30 sequences. Genome only saves after a sequence is completed.

Smp or dublicating the same install to multiple machines...
If you've got multi-cpu, you'll have to run multiple instances of genome. Just put them in different directories. There's no special switch for smp.
If you're copying an already made install, MAKE SURE to DELETE id.dat on all new copies. Also delete the wu with genome -clear

Nonetting, sneakernetting
Genome's biggest advantage over other projects is that it can re-run the same WU. Genome uses a random number for every re-run, and should therefore not normally make duplicates. For this reason it's very easy to put genome on machines without net-connection and let them crunch away for days/weeks before uploading the results.If you're running without net-connection, use genome -nonet
When you need to upload, you should move the whole directory to a net-machine.Or if you want to keep it running, copy the directory (except lock.gah that is locked), delete csum.#####, output.chi.#####, str.##### (##### is a number, the same for one result) from the un-netteded machine. On the networked machine, run genome -upload

WU-score
aa stands for AminoAcid, and its how many there are in a WU. This number is shown when filtering, and it's the number of letters printed after every sequence is done.But the easiest way to find the number of AminoAcids is to look at the number on the 1st line in input.inp
The score to a WU is calculated by: WU-score = 0.05 * aa + 0.0025 * aa * aa
This is more or less following the cpu-time, but there are some discrepancies.
As a comparison, an Intel machine running Seti will in the same time manage to produce 10x the WU's in Genome.AMD is a little faster, about 11x-12x

Misc info
In Genome, it's the raw cpu-speed that counts.Cache size, memory speed, cpu multiplier is important in Seti, but not for Genome.
The only difference is that AMD Athlon/duron/palomino is faster than Intel for the same MHz. There's no performance-hit running multiple instances on smp-machines.
Stats are updated every 3 hours, but due to checking it can take a couple of updates before results show up.
Genome-switches, run genome -help
Service-install, use either FireDaemon or srvany.exe.
Genome in windows runs at priority 4, the same as Seti CLi. A combination with Seti will therefore give 50-50-split.
Currently the only clients are win32-cmd and linux.