There is some really weird stuff going on with CPU capabilities check on AMD... According to cpuminer-opt binary: AMD Phenom(tm) II X4 965 Processor CPU features: SSE2 AVX AVX2SW built on Jun 5 2016 with GCC 4.8.3 Build features: SSE2 Algo features: SSE2 AES AES not available, starting mining with SSE2 optimizations... According to CPUID CPU-Z:  I'm pretty sure that this is the problem with -march=native on AMD CPU's without AES-NI. Yes very weird. I haven't found anything to suggest they've implemented CPUID differently than Intel, yet these flags are taken directly from the CPUID register. I would be nice if CPUz had open source so I could just copy their code. I found this, closest I found to your CPU. Something I can work with to figure out what's wrong. http://users.atw.hu/instlatx64/AuthenticAMD0100F42_K10_Deneb_CPUID.txt. Edit: Did you compile on a VM and run CPUz on the host? |
|
|
|
Does it come with power, if not what connector does it have? This looks like a nice little hasher if x11 wasn't so depressed. |
|
|
|
I have also seen this behavior but not for the last few days.
The individual miners do not show up but, in my case, the summary shows 0 miners but a hashrate existed for the algo.
Edit: and I was getting valid shares and being paid for the hashes
good, that's what really matters! Sometimes these issues are from crackfoo making modifications/upgrades to the boxes, etc. Which makes sense since its saturday =) I looked closely at the graph and there was no apparent deviation and it should have ben noticeable. When I first posted I simply added up the hash of the other workers and compared it to the sum and it was pretty close so I assumed I was losing hash. It seems now I'm was being credited for the hash so I guess it's not a big deal. |
|
|
|
|
I've found another problem, this time a little more serious.
Miners are disappearing.
Twice now I've noticed a miner missing from the list, The miner is ok without any interrruptions and has
been mining for several hours but it doesn't show up in the list and it's hash isn't counted in the sum.
If I stop the miner and restart it, it immediately pops up on the list and starts registering hash.
The second time it occurred mining lyra2v2, I don't recall the first.
|
|
|
|
Problems with DNET? Seems to be stuck loading blocks.
Probably needs to add some nodes: Nope not sure what the issue is.... I have lots of nodes... The one's I listed are on the DNET ANN page, and the wallet synced ok. Unfortunately, no getwork support to have fun solo mining the low diff. Good now, Thanks. |
|
|
|
cpuminer-opt v3.3.5 is available. Fixed hodl on Windows with non-AES CPUs. New build procedure for Windows, see README.md. This fix is only significant for Windows users who compile their own on non-AES CPUs. Cryptomining Blog does not build binaries for non-AES CPUs. Rewrote CPU capabilities check to provide more info and be less verbose. It is implemented in permissive mode meaning that mining will always be attempted. The miner may crash or perform at less than optimum. This should close any remaining issues for now. It is recommended that all users upgrade even those who don't like to update every release. Download source: https://drive.google.com/file/d/0B0lVSGQYLJIZVVp5aGxFVm1qVE0Check CMB or OP for Windows binaries availability. |
|
|
|
I hope everything is great and you are having a good time with Revolver.
One week has passed since launch. Lets briefly recap it.
From our perspective launch was fair - AMD and internal\external CPU miners were available upon launch and we saw some solid and steady net hash increase over time. CCMiner for NVidia users arrived just couple days later and showed significantly better performance than AMD counterpart at the moment. So right now AMD, NVidia and CPU miners are happily hashing together.
You don't have a link to cpuminer-opt in the OP. It's twice as fast. |
|
|
|
Agreed on the reasoning but I prefer having a root login for the simple reason it requires a second password. If the user password gets cracked with sudo enabled the whole machine is compromised. Without sudo root would have to be cracked. Sudo can then be used to provide limited root priviledges like installling packages and mounting drives (is that still priviledged) but nothing too dangerous. |
|
|
|
My plan is to do a "net" install of Debian, basically the bare minimum, then figure out what I need to type next. Essentially a "step-by-step" from complete scratch/zero. It's in a VM, so no biggie if I have to do over and over just to get it right.
My only experience with minimal installs is Centos. It provide the install time options for both a minimal install and development tools. After that install some of the missing dependencies for cpuminer-opt then pick off any stragglers when the compiler complains about them. That would make for a pretty lean machine. But I'm sure Debian can do it just as well. |
|
|
|
Does ubuntu have a "console" install? Non-graphical type? Of course I will go to their website. I know how to run a Pi2 a bit, and I think Rasbian is based on Debian, so ... well, that's just what I've been using for 20 years (as a newbie, 20 years a Linux newbie, go figure). Ubuntu is more recent (about 10+ years old), and is also Debian based.
Anyway, queuing download for 16.04 LTS ubuntu 64-bit, desktop and server versions.
Last I checked (many moons ago) the server edition of ubuntu was text only so I'msure they do. I don't think you need to switch. If you 've been using Debian for so long and you know how to compile it should be a brease. I only recommended Ubuntu because it's trivial to install and has a lot more bells and whistles to make it more desktop friendly. Strip that away and it's basically all Debian. |
|
|
|
I am proposing a new direction for the CPU capabilities check... Checking CPU capatibility...
Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz
CPU features: SSE2 AES AVX AVX2
SW built on Jun 4 2016 with GCC 4.8.4
Build features: SSE2 AES AVX AVX2
Algo features: SSE2 AES
Start mining with AES-AVX optimizations...
More technical detail but last line summarizes for users. Better? It would also be nice to include the -march flag either as specified or translated from native. Haven't found a way except indirectly via build features. Fortunately the ambuguity does not affect the miner's decision to use AES. mine with sse2: core2 -> SSE2 corei7 -> SSE2 corei7-i -> SSE2 AES mine with aes: corei7-avx -> SSE2 AES AVX corei7-avx-i -> SSE2 AES AVX corei7-avx2 -> SSE2 AES AVX AVX2 |
|
|
|
My 2 E5 Xeon 2670 dual platform, a total of 16 nuclear 32 threads, HMQ1725 algorithm is only 361K, this is not the normal bar @3.3.4 version
Can you provide more information? Are you saying HMQ1725 on v3.3.4 is slower than previous versions? Did you compile yourself or use a precompiled binary? Also please post the startup messages showing the CPU capabilities. |
|
|
|
 I don't disagree. But I'll have to spend some time playing with git before I comit cpuminer-opt to it. And that will have to wait until the development slows. It will happen eventually. Still looking for an endian.h package that will work with Windows so I can enable hodl on non-AES CPUs.
|
|
|
|
|
Problems with DNET? Seems to be stuck loading blocks.
|
|
|
|
|
A little more info for users with early corei CPUs. I found this which lists the mapping between -march
and the processor family.
This help clarify how CMB has built their binaries for Intel CPUs.
They have chosen to only build for CPUs with AES and AVX. This includes:
- westmere (-march=corei7-avx)
- sandybridge ivybridge (-march=corei7-avx-i)
- haswell broadwell (-march=corei7-avx2)
This excludes the following architectures that are supported by cpuminer-opt with SSE2 optimizations:
- nehalem (-march=corei7)
- core2 (-march=core2)
In order to use cpuminer-opt on these CPUs it is necessary to compile from source on that machine using
-march=native which will automatically select the correct architecture for your CPU.
For those who desire to specify the architecture manually please note that in GCC 4.9 and above you may
specify the family name, ie -march=haswell rather than -march=corei7-avx2.
I'm still waiting for feedback from more Westmere and Nehalem users. I presume the lack of feedback means
things are working (or not) as expected.
I still don't have any issues that would justify a new release so I'll wait a little longer.
|
|
|
|
The Google Drive "direct" link somehow works on Chrome and Windows, but don't work with wget on Ubuntu urban@storm:~$ wget -4 https://drive.google.com/uc?export=download&id=0B0lVSGQYLJIZVTFNSkZ0elRQZ2M
[1] 10374
urban@storm:~$ --2016-06-04 11:03:26-- https://drive.google.com/uc?export=download
Resolving drive.google.com (drive.google.com)... 216.58.209.206
Connecting to drive.google.com (drive.google.com)|216.58.209.206|:443... connected.
HTTP request sent, awaiting response... 400 Bad Request
2016-06-04 11:03:26 ERROR 400: Bad Request.
You may need quotes around your URL pi@raspberrypi:~ $ wget -O cpuminer-opt.tar.gz -4 "https://drive.google.com/uc?export=download&id=0B0lVSGQYLJIZVTFNSkZ0elRQZ2M"
--2016-06-04 12:39:10-- https://drive.google.com/uc?export=download&id=0B0lVSGQYLJIZVTFNSkZ0elRQZ2M
Resolving drive.google.com (drive.google.com)... 64.15.112.84, 64.15.112.103, 64.15.112.118, ...
Connecting to drive.google.com (drive.google.com)|64.15.112.84|:443... connected.
HTTP request sent, awaiting response... 302 Moved Temporarily
Location: https://doc-00-7o-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/dm3nstjm774g792jpfu2h5q95p8c1g2j/1465041600000/02937440540891330699/*/0B0lVSGQYLJIZVTFNSkZ0elRQZ2M?e=download [following]
Warning: wildcards not supported in HTTP.
--2016-06-04 12:39:14-- https://doc-00-7o-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/dm3nstjm774g792jpfu2h5q95p8c1g2j/1465041600000/02937440540891330699/*/0B0lVSGQYLJIZVTFNSkZ0elRQZ2M?e=download
Resolving doc-00-7o-docs.googleusercontent.com (doc-00-7o-docs.googleusercontent.com)... 216.58.214.65
Connecting to doc-00-7o-docs.googleusercontent.com (doc-00-7o-docs.googleusercontent.com)|216.58.214.65|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [application/gzip]
Saving to: ‘cpuminer-opt.tar.gz’
cpuminer-opt.tar.gz [ <=> ] 17.50M 464KB/s in 40s
2016-06-04 12:39:55 (446 KB/s) - ‘cpuminer-opt.tar.gz’ saved [18348257]
pi@raspberrypi:~ $
or try curl pi@raspberrypi:~ $ curl -L "https://drive.google.com/uc?export=download&id=0B0lVSGQYLJIZVTFNSkZ0elRQZ2M" > cpuminer-opt.tar.gz
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 383 0 383 0 0 94 0 --:--:-- 0:00:04 --:--:-- 94
100 17.4M 0 17.4M 0 0 352k 0 --:--:-- 0:00:50 --:--:-- 384k
pi@raspberrypi:~ $
Thanks for this. For me on Linux it works with curl but not wget or Firefox. |
|
|
|
i read the INSTALL file, it certainly wasn't there
README.md Edit: there are 2 files of interest. The other one is RELEASE_ANNOUNCEMENT which describes changes features and performance and is a copy of the link in my sig. |
|
|
|
I continue to encourage Windows users to compile their own. That is always the best way
How would I go about doing that? LOL. Is there a "free" Windows compiler out there? I have Xeon E5640. I do have access to Visual Studio Community 2015; supposed to be the same as Professional without CodeLens. On the other hand, is there some "newbie" guide to doing this with a fresh install of say, Debian 8.4? (Or would you recommend Ubuntu?) I just made a Debian 8.4 VM. Maybe I'll do it over again. This is compiled on windows with a free compiler, but it is a pain to set up if you don't know what you are doing. It can compile using mingw-w64 and msys, but you need to first compile all of the dependencies and link everything correctly. That was not the case for me using msys and mingw_w64. Nothing to compile except cpuminer-opt itself. |
|
|
|
I continue to encourage Windows users to compile their own. That is always the best way
How would I go about doing that? LOL. Is there a "free" Windows compiler out there? I have Xeon E5640. I do have access to Visual Studio Community 2015; supposed to be the same as Professional without CodeLens. On the other hand, is there some "newbie" guide to doing this with a fresh install of say, Debian 8.4? (Or would you recommend Ubuntu?) I just made a Debian 8.4 VM. Maybe I'll do it over again. cpuminer-opt doesn't compile on any form of VS. There are two viable options. 1. Compile natively on Windows with mingw and msys. Instructions are in the README.md file. Installing msys and mingw_w64 can be a bit confusing. They are two seperate installs with different installers but they are built to work together. 2. Install a Linux VM and compile and run cpuminer from the vm, full speed. I would recommend Ubuntu or Mint. Fedora and centos are also good choices. Debian is not for newbies but if you already have it installed go with it. Once the OS is installed you have to install the development tools. I don't have the minimalist command to get everything (note to self to do that), I usually start with gcc and any of its dependencies and iterate from there until it works. In addition to the develpopment tools there are some libraries that need to be installed. They are listed in README.md. Usng a VM may be preferable if you intend to mine hodl with an older CPU because that doesn't work on Windows. |
|
|
|
Thank you very much. This is important information. As mentioned in a subsequent post
your CPU actually uses the nehalem compile arch, the westmere build is incompatible.
This solves part of the mystery. There are two compile targets for 1st generation corei CPUs,
Nehalem which will not run the miners' AES code, and Westmere which will.
For now I don't have the means and time to try my own compile, but whenever I can help with a run/test, just PM me and I'll give is a try. And you... keep up the good work ! Thank you. And thank you for the kind words. If you ever get the urge to try to compile your owm I'm here to help. |
|
|
|
|
Show Posts
