Title: 5970 under Ubuntu Post by: xanadu on May 04, 2011, 07:36:20 PM I have a new system up and running using Ubuntu and a 5970 card.
Using the phoenix miner (vectors, -k poclbm, BFI_INT, aggression=8, worksize=128) it runs one core at 300/Mhash, but when I start the second core by specifying device= both miners drop down to 180-190/Mhash. Normally that would be a sign that both miners were using the same core, but I've double checked and indeed the device= is being properly set. Starting the miner without specifying just shows [ 0 ] Cypress [ 1 ] Cypress [ 2 ] AMD Sempron CPU usage also shoots up to 100% so it could be a resource constraint somewhere else? (No, I didn't start the miner running on the CPU!) Any ideas? I'm new to Ubuntu so this has been an interesting learning curve to get the machine running. Thanks, -X Title: Re: 5970 under Ubuntu Post by: netxshare on May 04, 2011, 07:39:28 PM Can I see your full command line?
Title: Re: 5970 under Ubuntu Post by: xanadu on May 04, 2011, 07:50:03 PM Can I see your full command line? python phoenix.py -u http://name:pass@ipaddress:8332 -k poclbm BFI_INT aggression=8 vectors worksize=128 device=0 (or 1) Title: Re: 5970 under Ubuntu Post by: xanadu on May 04, 2011, 08:17:25 PM I just used the command
Quote $ aticonfig --odgc --adapter=all Adapter 0 - ATI Radeon HD 5900 Series Core (MHz) Memory (MHz) Current Clocks : 725 1000 Current Peak : 725 1000 Configurable Peak Range : [550-1000] [1000-1500 GPU load : 46% Adapter 1 - ATI Radeon HD 5900 Series Core (MHz) Memory (MHz) Current Clocks : 725 1000 Current Peak : 725 1000 Configurable Peak Range : [550-1000] [1000-1500] GPU load : 51% It looks like both GPUs are running at 50% load Title: Re: 5970 under Ubuntu Post by: twitcoins on May 05, 2011, 01:59:39 AM Issue 'top' to keep an eye on your processes. Launch the first miner. I'm guessing you have a single-core sempron and it will spike the CPU to 100%.
If you have a multicore Sempron, I'm willing to bet you'll see %CPU spike over 100%, and with both miner's running I'd bet you peg both cores. Just a guess though. cat /proc/meminfo /proc/cpuinfo and paste the results here? Title: Re: 5970 under Ubuntu Post by: xanadu on May 05, 2011, 03:58:22 AM Issue 'top' to keep an eye on your processes. Launch the first miner. I'm guessing you have a single-core sempron and it will spike the CPU to 100%. cat /proc/meminfo /proc/cpuinfo and paste the results here? Absolutely dead on, it is a single core Semperon CPU, I was under the impression that mining put very little load on the CPU since the GPU was doing all of the calculating so I didn't spend anything more than the minimum for the CPU on the dedicated mining hardware. Here is the /proc/meminfo MemTotal: 2059804 kB MemFree: 1099720 kB Buffers: 117872 kB Cached: 320908kB SwapCached: 0kB Active: 351276 kB Inactive: 315408 kB Active(anon): 229692kB Inactive(anon): 2144 kB Active(file): 121584 kB Inactive(file): 313264 kB Unevictable: 16kB Mlocked: 16 kB HighTotal: 1187400kB HighFree: 547808 kB LowTotal: 872404 kB LowFree: 551932 kB SwapTotal: 2094076 kB SwapFree: 2094076 kB Dirty: 68 kB Writeback: 0 kB AnonPages: 227936 kB Mapped: 101988 kB Shmem: 3936 kB Slab: 161460 kB SReclaimable: 147012 kB SUnreclaim: 14448kB KernelStack: 2112 kB PageTables: 4620 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 3123976 kB Committed_AS: 1220480 kB VmallocTotal: 122880 kB VmallocUsed: 60360 kB VmallocChunk: 56476 kB HardwareCorrupted: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 4096kB DirectMap4k: 12280kB DirectMap4M: 897024 kB Here is the CPU info: processor : 0 vendor_id : AuthenticAMD cpu family : 16 model : 6 model name : AMD Sempron(tm) 140 Processor stepping : 2 cpu MHz : 2700.000 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc up nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt npt lbrv svm_lock nrip_save bogomips : 5400.75 clflush size : 64 cache_alignment : 64 address sizes : 48 bits physical, 48 bits virtual power management: ts ttp tm stc 100mhzsteps hwpstate Title: Re: 5970 under Ubuntu Post by: RedSky on May 08, 2012, 08:55:03 AM Any update on a solution? I just updated Ubuntu to 12.04 and am running into the same problem...
Title: Re: 5970 under Ubuntu Post by: zvs on May 08, 2012, 07:49:00 PM Any update on a solution? I just updated Ubuntu to 12.04 and am running into the same problem... i have 3 5970's on some random dual core intel CPU with one core disabled, device 0 at agg 9, rest at agg 12 use SDK 2.1? and run them at 150mem with 128 worksize or 300mem @ 256 for maybe 1% more Title: Re: 5970 under Ubuntu Post by: RedSky on May 09, 2012, 03:51:13 AM I don't think it's the configuration settings - pretty much tried every reasonable combination there. I also tried downgrading the SDK to 2.4 from 2.6 (albeit not all the way to 2.1), with no luck.
The problem is that the GPU cores only see 50% load. It's like the miner (or maybe the driver) distributes the load for one core across both instead of running each under full load. This is my first time running the 5970 so it could be I'm missing something basic, but my problem seems to match exactly that of the OP. Title: Re: 5970 under Ubuntu Post by: ThiagoCMC on May 09, 2012, 08:56:12 PM I have 4 5970 at water using:
Ubuntu 12.04 64 bits server Catalyst 11.12 SDK v2.4 CGMiner 2.4.1 Clocks 790 / 150 = 355 MHash per GPU Cheers! Thiago Title: Re: 5970 under Ubuntu Post by: RedSky on May 10, 2012, 09:40:20 PM Haha well thanks for rubbing it in...
The only difference with my system though seems to be the Catalyst driver version. I'll try rolling that back and see what happens. |