I use 64bit Windows and 10GB ram. And cards I use: 2*5970 + 2*6870 + 5870 Tried different kernels, same problem. There is no limit on the number of commands in a commandqueue. It is only restricted by the system memory.
If you exceed this limit you are likely to get a CL_OUT_OF_HOST_MEMORY.
I hope it is clear
If HOST_MEMORY means RAM then I'll try 16GB of it. EDIT: I realised I use 32bit version of cgminer. Does it make sense if I try to compile 64bit version of it? You have a lot of GPU devices. You will probably need to disable certain system devices in the BIOS such as serial/parallel port, USB ports, audio, PS/2 support, and so on to make room for the GPU's to be mapped.
|
|
|
Card Manufacturer: ATI Card Model Number: 5870 Card OEM: Asus MHash per Second: 504.3-504.5 MH/s Core Clock: 1090 Mhz Memory Clock: 205 Mhz Voltage: 1.200v Average Operating Temperature: Varies. Usually Ambient + 38c Ambient Temperature: Varies. Sits outside. Currently 8c as I type this. Fan Speed: Dedicated 60mm delta FFB0612EHE @ 100% Host OS: Win7 x64 Driver Version: 11.12 CAT SDK Runtime Version: 2.1 Mining Program: Phoenix 2.0 Beta RC1 Command Line Flags/GUIminer Settings: phatk2, AGGRESSION=14, VECTORS=true, BFI_INT=true, WORKSIZE=128 Other Information: Stable for 1 month so far. Will reduce speed/volts as spring temperatures increase.
|
|
|
Pictures don't mean anything either these days. Escrow or gtfo.
|
|
|
*Sigh* See, the problem is people just copy paste. They don't try to understand what a config file does or what you should put in there. Every time I see that autodetect line, I just facepalm.
Well maybe if there was some actual sample code or less assholeism from forum members itd be easier to get working..? You mean the example on the very first post of the very first page? It's an example. It's not meant for you to copy paste. Understand it. Learn from it. Make your own to best suit your needs. I don't know what hardware you have or what parameters best suit your needs, only you do.
|
|
|
AFAIK, there aren't any really successful miners with GCN-friendly kernels that are compatible with 7970 and similar cards that are also compatible with the GUIMiner frontend. GUIminer itself is not a miner. It's just a GUI so people that are scared of command lines have something more familiar to grasp onto, but GUIMiner has to know how to understand what the miner is outputting in order to show you basic stuff like shares, mhash, and so on, so it won't work with every miner, but only with the ones it's been programmed to understand. There was some work on DiaKGCN, a phoenix-compatible kernel that's got optimizations for the GCD architecture, but I'm not sure if its phoenix 1.7.5 compatible (I only saw phoenix 2 variants, but I could be wrong. Phoenix 2 currently isn't compatible with guiminer due to how it's argument are stored in a config file now). If the card is also dedicated, you should probably ramp up the aggression so it doesnt fluctuate so much because it sounds like you are using the computer and have it at a low aggression
|
|
|
I dont understand how people can pop the fan part off of the metal and electrical parts like nothing. I broke 2 blower fans at the part that joins the triangle plastic thing to the PCB (actually, just before it, it leaves a circle piece of plastic left in the PCB part)
|
|
|
*Sigh* See, the problem is people just copy paste. They don't try to understand what a config file does or what you should put in there. Every time I see that autodetect line, I just facepalm.
|
|
|
300 is too fast for worksize 128 (much slower). Close trixx and reopen it and you should be able to select lower memory speeds. You should get like a 10 mhash increase at the least going from 300 to 205 with 128 worksize.
|
|
|
stuff thanks
|
|
|
Hash speed seems to scale linearly per shader per MHz with proper memory speeds/settings/sdk/etc. With 735 core and 1440 VLIW5 shaders, you should get about 307 MHash, so you're pretty much close to the mark. I can suggest to run 205 MHz memory and 128 worksize (About 0.15% mhash loss, but the lower temperatures makes up for increased core speed headroom. I got about 4 watts less per 5830 going from 330 memory to 205 memory, about 1c less in temperature and allowing a small bump in core) and cranking up the core speed. If you're wondering about where I got the memory speed from, I was bored one day and did an exhausting test of many memory mhz vz worksize. you can see it here https://docs.google.com/spreadsheet/ccc?key=0AjXdY6gpvmJ4dEo4OXhwdTlyeS1Vc1hDWV94akJHZFE&hl=en_US#gid=0 or in this picture of a graph.
|
|
|
Thanks! I have tried. MEGA-BOOST!!! HD5830@925MHz - before 297Mh/s after 327Mh/s HD6930@950MHz - before 354Mh/s after 389Mh/s HD6770@925MHz - before 213Mh/s after 234Mh/s Bullshit. Screenshot/movie (fraps harder to fake) or gtfo
|
|
|
Found them newegg has them now. and I7- 3930K's too.
wait what? I missed 3930K? FUUUUUUUUUUUUUUUUUUUUUUUUUUCK
|
|
|
Feel free to make an offer...
Would have been interested, but I've been scammed twice on these forums from people that -just- got out of newbie forums. Going to pass.
|
|
|
If you want to save about 10 watts worth of heat, instead of running 300mem and 256 worksize, try ~205 (or less, higher may be unstable/slower) and 128 worksize. I was able to save 13 watts at the wall 1 5870 and 2 5830's and lost maybe 1 mhash between the three (used to run 330 mem)
|
|
|
watercooling them beasts will set you back couple thousand of $, lol.
How about oil-cooling (i.e. submerging the whole thing in a tank of mineral oil)? No. That is a gimmick and not 24/7 recommended. Heck its not even 2 hour recommended. Without something to swiftly move all the oil quickly across hotspots, it'll just overheat. Plus, all the oil will just soak up heat. It won't dissipate it. Thats what radiators are for in water cooling.
|
|
|
For 5000 series cards 2.1 is the best but only by 2% or 3% or so. Any hard evidence or just speculation. I am using 2.1 myself but I don't think it really is that much better than 2.4. I want to see actual 2.1 vs 2.4 results for a 5xxx card showing 2.1 is better if that exists anywhere and also core to memory ratio analysis. Thanks ! Well I have a good chunk of sample data at https://docs.google.com/spreadsheet/ccc?key=0AjXdY6gpvmJ4dEo4OXhwdTlyeS1Vc1hDWV94akJHZFE&hl=en_US#gid=0 but it's currently for 2.1 SDK. Would need to uninstall 2.6 and install 2.4 sdk instead (which is fine for me, dont use 2.6 anyways) and run all the combos the same way. Here's a pretty graph in case you don't want to read the spreadsheet. Note: My sample is very incomplete. It requires a lot of time to test all the combos. Environment was Win7 x64, phoenix 2.0 RC1, phatk 2.2 kernel, 12.1 driver, 2.1 SDK, AGGRESSION=14 BFI_INT with varying worksize and vectors. Card used is an Asus Radeon 5870 @ 1035mhz core.
|
|
|
[0-0]barts is a 6870 (with 2.5 sdk) and [1-1]juniper is a 5770 with 2.1 sdk. Both sdks working great together. Shouldnt be a problem to throw in mix 2.6 and others together. It only works like that because of your situation. You can install 2.1 plus ONE of the following: 2.4, 2.5, 2.6 (havent bothered with 2.2/2.3, people say its cpu bug) because 2.1 installs into an entirely different path than 2.4-2.6. 2.4, 2.5, and 2.6 all share the same install path and, afaik, you can't install 2 or more of them since the new one will overwrite the old one (also the installer refuses to downgrade, have to 2.4 -> 2.5 -> 2.6, can't go backwards without manually uninstalling)
|
|
|
Its not the motherboard. I know that for sure. The oven method I have not tried. I am not experienced with that. I've read about people doing it for different things. It creates a "re-flow" of something so that the electrical condition is better. I'll google it but please tell me more about this and what it might fix Edit: I did google it. I am considering this. Bake it at 385 for 8 minutes maybe longer. I didn't see a consensus online but 8-10 minutes seemed to be it. I might just try this. Remember this is only temporary. [/quote] Reflow means melting the joints that join whatever is on the PCB to the PCB itself. In many heat/cool routines, the joints can become stressed and crack. Liquifying, or reflowing, them removes these cracks and completes the circuit.
|
|
|
Everyone with high power costs hates overlocking. Personally every hash I can squeeze out of the cards just means more profits for me. 11.11 driver works well, if 11.12 you need to delete the following files after uninstalling 11.12 to downgrade. C:\Windows\System32\amdocl64.dll C:\Windows\System32\amdoclcl64.dll C:\Windows\System32\OpenCL.dll C:\Windows\SysWOW64\amdocl.dll C:\Windows\SysWOW64\amdoclcl.dll C:\Windows\SysWOW64\OpenCL.dll
I run all 18 of my 5970's at 1.100v 870 clocks 500 memory
some cards require 1.112 volts or 850 clock.
Those memory clocks are exceeedingly high. 300-380 is the optimum range for 256 worksize. 128 worksize is about half that (150-190) but has slightly less hashes than 256 worksize.
|
|
|
wow, thanks for fast fix, thanks. eagerly awaiting next windows build.
|
|
|
|