Lean how to use --gpu-map

My cross-compiling setup suddenly fails on the newest git master version. I get multiple errors about sys/socket.h not being found. I traced the problem back to commit 31aa4f6cebc51e26b349606fd78d71954bda87da. Commit 657e64477b75603bc9b08eed425bc47f606814cb and everything before that compiles correctly. Is there a new dependency that I'm now missing, or is this a bug? The only sys/socket.h I have (/usr/include/x86_64-linux-gnu/sys/socket.h) is not under my mingw toolchain, but that wasn't a problem when cross-compiling until now. For reference, here are the files in my toolchain: http://pastebin.com/sSEqFL63

Good to know, I guess I shouldn't expect every commit to work perfectly in marginal cases like cross-compiling.

EDIT: I saw some new commits and tried again. Now it works.

My config :

"intensity" : "20,20",
"vectors" : "1,1",
"worksize" : "256,256",
"kernel" : "scrypt,scrypt",
"lookup-gap" : "2,2",
"thread-concurrency" : "21712,21712",
"shaders" : "1792,1792",
"gpu-fan" : "100,100",
"gpu-engine" : "800-930,800-930",
"gpu-map" : "0:1,1:0",
"gpu-powertune" : "-11,-11",
"temp-cutoff" : "92,92",
"temp-overheat" : "88,88",
"temp-target" : "85,85",
"api-port" : "4028",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "60",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"auto-gpu" : true

Well it looks like your using scrypt which I know nothing about.

Your using auto gpu which is good.  I don't know if your engine range is OK?

Your not using auto fan so I would enable that and set the range for the fans to something like 60-85 instead of running at 100.  Still seems odd that your GPU is even reaching 92.  Are you running with your case on?

Since your not using auto fan and running them at 100% the only variable to control the temperature is to lower the engine clock and the floor is 800.

Something is missing.

That part is by design.

From the top post

But I'm saying that your GPU's shouldn't even be reaching the temp cutoff, so you need to work on that somehow.

cgminer can mine any double sha256 coin with the ASICs it supports (i.e. not scrypt coins like litecoin) and avalons have been used to terrorise some of those altcoins already.

I have issue about efficiency ...

U:9:02/m and U:5.02/m can i ignore it or worry?



thx

i can confirm this since i have a 580 too (slighty OC).
if your interested which miner it is, the miner is called cudaminer
also with cudaminer you only use RAM on your GPU, you dont need alot of sysRAM to mine it my cudamminer uses 35kB RAM and allocates 1252 MB on my GPU (this depends on the launch-config of cudaminer)

U: or Utility is how many shares per minute.  Why would you ignore or worry about it?  It's good information.