[ANNOUNCE] Tenebrix, a CPU-friendly, GPU-hostile cryptocurrency

[Lolcust]

There seems to be room for optimization arising from how APUs can use memory, but I guess only time (and engineers and programmers) can tell.

 

[1714218972]

1714218972

[1714218972]

1714218972

[1714218972]

1714218972

[Bobnova]

I'll be pretty excited if APUs turn out to do that.
I've been saying that AMD should work on microcode to offload floating point stuff to the GPU core for a while now.

Any thoughts on my linux tenebrix-miner compiling issue?

EDIT:
I commented out the libcurl check line that was erroring out, as I know I have libcurl.
Now it configures happily, and gives me this on "make":

Code:

bobnova@bobnova-P67A-UD4-B3:~/Tenebrix-miner$ make
make  all-recursive
make[1]: Entering directory `/home/bobnova/Tenebrix-miner'
Making all in compat
make[2]: Entering directory `/home/bobnova/Tenebrix-miner/compat'
make[3]: Entering directory `/home/bobnova/Tenebrix-miner/compat'
make[3]: Nothing to be done for `all-am'.
make[3]: Leaving directory `/home/bobnova/Tenebrix-miner/compat'
make[2]: Leaving directory `/home/bobnova/Tenebrix-miner/compat'
Making all in x86_64
make[2]: Entering directory `/home/bobnova/Tenebrix-miner/x86_64'
/usr/bin/yasm -f elf64 sha256_xmm_amd64.asm
rm -f libx8664.a
ar cru libx8664.a sha256_xmm_amd64.o 
ranlib libx8664.a
make[2]: Leaving directory `/home/bobnova/Tenebrix-miner/x86_64'
make[2]: Entering directory `/home/bobnova/Tenebrix-miner'
gcc -DHAVE_CONFIG_H -I. -pthread -fno-strict-aliasing  @LIBCURL_CPPFLAGS@  -DHAS_YASM -O3 -Wall -msse2 -MT minerd-cpu-miner.o -MD -MP -MF .deps/minerd-cpu-miner.Tpo -c -o minerd-cpu-miner.o `test -f 'cpu-miner.c' || echo './'`cpu-miner.c
gcc: @LIBCURL_CPPFLAGS@: No such file or directory
cpu-miner.c: In function ‘parse_arg’:
cpu-miner.c:786:3: warning: passing argument 2 of ‘json_load_file’ makes integer from pointer without a cast
/usr/local/include/jansson.h:225:9: note: expected ‘size_t’ but argument is of type ‘struct json_error_t *’
cpu-miner.c:786:3: error: too few arguments to function ‘json_load_file’
/usr/local/include/jansson.h:225:9: note: declared here
make[2]: *** [minerd-cpu-miner.o] Error 1
make[2]: Leaving directory `/home/bobnova/Tenebrix-miner'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/bobnova/Tenebrix-miner'
make: *** [all] Error 2

[racerguy]

is there anyway to limit the amount of connection the client makes?  Think it's crashing my router when it goes too high.

[Bobnova]

Hey wanna know what's funny?  The windows miner works just fine in Linux connected to the linux-built client.  Go figure.
It's a bit on the slow side, 1.0kh/s rather than 1.06-1.1, but I'll take it for being able to mine in linux.

[EskimoBob]

Hey wanna know what's funny?  The windows miner works just fine in Linux connected to the linux-built client.  Go figure.
It's a bit on the slow side, 1.0kh/s rather than 1.06-1.1, but I'll take it for being able to mine in linux.

Wine+miner did not help me at all with miner. It crashed on start. How did you run the Windows version in Linux?

[Bobnova]

What I did:

1) Install all dependancies for main linux client and linux build of the miner.
2) Curse the miner's configuration file.
3) Run the main linux client.
4) Download the windows package.
5) Fire up a terminal and cruised on over to whatever I put the windows package.
6) "chmod +x minerd.exe"
7) "./minerd.exe --algo scrypt --s 5 --threads 6 --url http://127.0.0.1:8697 --userpass 1:1"  Swap out the 6 for however many threads you want.  Using 100% of the CPU's available threads cripples it for daily use for me.
Party.

Note that Wine is never actually called, I don't see any evidence of it either.  It may be running in stealth mode, I don't know.
In any case, it works though it's ~10% slower than the same minerd in windows.
Also note that the windows client is not being run.

Let us know how that works for you.

Lastly:  I'm on Ubuntu 11.4 with all the current updates.

[phelix]

[...]

I have a question so the difficulty changes every week, is that time based or block based?

Same as BTC, but faster.

how many coin will there be in existence?

no limit, except for natural balance of generation / loss that will be achieved at some point.

[...]

could I sum this fork up like this?:
- coins I hold will always suffer from inflation
- sooner or later network hashrate will drop and tenebrix will freeze orphaned as namecoin
- will be dominated by FPGAs   (Why would ArtForz design such a thing...  hmmm....)

btw: anyone made the big start money this time?

[kano]

OK, repeat question ... but generalised this time.
Anyone know where to find the autogen.sh, configure or Makefile missing from in https://github.com/Lolcust/Tenebrix
It has none - that link is the link in the first post for source and yet I guess some people have compiled this on linux.
(yes I can go and grab the files from bitcoin - but why have a git source without the compile files?)
I guess you all got the source from somewhere else? (not the place listed in the first post)

[dishwara]

What does it mean when it says boo or yay sometimes?  When I get a yay, it shows up as a transaction, but not necessarily as 25 TBX?  I really appreciate the quietness of mining.  

I think it counts near-solutions and orphans as transactions, looking into it.

While i mined only 15 blocks(375 coins), overview shows Number of transactions: 20
There is no other Transactions besides "Mined" in Transactions tab.

[Lolcust]

- coins I hold will always suffer from inflation

Now this inflation argument is incredibly weird.

Not only do terms inflation/deflation mean things that do not directly apply to infinitely divisible (and also infinitely re-nominable) abstract mathematical constructs (thus rendering the entire branch of discussion pretty moot), but you seem to completely ignore the fact that stable output means that relative increase in "amount" of coins will be on a constant decline.

By the way, ironically, that is exactly the way everyone's favorite "deflationary" commodity, gold, behaves in real life (you see, presence of abstract theoretical limit to gold extraction from a planet does not really affect market behavior of gold in real life within the planning scope equal to our expected lifespan, which I hope we can agree is the largest reasonable timeframe to plan for   )

Oh, wait a moment, gold production  per annum has actually grown since 1990s. Holy shit, Tenebrix is more "deflationary" than gold (that is, if you believe abstract constructs spawned by eldritch mathematications can really be "deflationary" in the usual sense  )

Also, you seem to ignore natural losses.

You know what ? I will implement some namecoinish coin-purging trick (without demurrage and/or mining rate decrease, of course, principles are principles) just so that when the next deflation guy pops up, I'll just link him to that line of code on github and be done with it  

- sooner or later network hashrate will drop and tenebrix will freeze orphaned as namecoin

Namecoin orphaned ? You must not have been paying attention, mate.

Also, what do ya suggest, crazy skewed retargets, lol ?  

will be dominated by FPGAs   (Why would ArtForz design such a thing...  hmmm....)

Do you have any shred of an idea just how much a good FPGA (just the raw hardware, not R&D costs and shenanigans) costs ?

If it becomes lucrative to mine Tenebrix with dedicated hardware like that, that means Tenebrix is a success.

Also, it is questionable whether FPGA implementation will actually outperform CPU of comparable price (or perhaps, a private pool of cpu miners, at that kind of price ), and by what margin, in terms of raw performance (not in terms of per watt performance, lol)

btw: anyone made the big start money this time?

Depends on how you define money  :-P

But on the other hand, some people implement things that are most favorably described as "sovereign bonds", for the love of Cthulhu  

[EskimoBob]

OK, repeat question ... but generalised this time.
Anyone know where to find the autogen.sh, configure or Makefile missing from in https://github.com/Lolcust/Tenebrix
It has none - that link is the link in the first post for source and yet I guess some people have compiled this on linux.
(yes I can go and grab the files from bitcoin - but why have a git source without the compile files?)
I guess you all got the source from somewhere else? (not the place listed in the first post)

those files are not missing, those files will be created when you run the required commands to build the client or the miner.
I think you are confusing Tenebrix and Tenebrix-miner

cd  to a directory you like to use for building your stuff from source

Code:

$ git clone https://github.com/Lolcust/Tenebrix 

$cd Tenebix 

I use another terminal window to read help files
(you are still in Thenebix directory)

Code:

$ cat README.rst 

make sure you have all the required stuff installed

Lets build the client.

Code:

$ qmake 
$ make

as a result now you have a 'bitcoin-qt' and you can run it:

Code:

$ ./bitcoin-qt 

If you are asking about Tenebrix-miner:

Code:

$ git clone https://github.com/Lolcust/Tenebrix-miner
$ cd Tenebrix-miner
$ cat README
...
$ ./autogen.sh   

if this has no errors, run:

Code:

$ CFLAGS="-O3 -Wall -msse2" ./configure

and finally:

Code:

$ make 

Start you CLIENT and then run the miner.

[caston]

I've been trying to run it on EC2 with limited success. I am a EC2 newbie. I just used the Windows version (even though I am mining on Linux at home) as it would be easier to install the binaries and get up and running faster. It does mine but the hash rate is pretty slow and about half what I am getting at home. I picked a fairly low end configuration though and have only tried 2 threads so far.

The application does start and get connections and downloads blocks but the window seems to lock up and go all transparent.

If I actually knew what I was doing and my synapses made awesome connections much faster I may have tried to get this going last night (instead
of setting it up on my Linux box) and then used EC2 to cluster mine the crap out of this baby.

My assumption though is that if this works other people are already doing it and that may explain why my block rate decreased quite a lot since last night when I first started mining.

[caish5]

I'm noticing 2 k/hash per sec per core on a Q9400 C2Q, but only 1.67 on an i7-9600 both at stock clocks. Can anyone else confirm this. The i7 is on windows 7 and the core2 is on ubuntu.

[racerguy]

I'm noticing 2 k/hash per sec per core on a Q9400 C2Q, but only 1.67 on an i7-9600 both at stock clocks. Can anyone else confirm this. The i7 is on windows 7 and the core2 is on ubuntu.

I'm mining with a q9400 overclocked to 3.2ghz on windows and getting about 1.45khsh/s, looks like the linux miner is faster somehow.

[caston]

I'll be interested to see how many blocks my machine has done when I get home from work. My expectations though is that it probably lost the RPC connection as connections drop in and out.

[Bobnova]

I'm assuming you mean an i7 960, if so remember that the 960 has twice the threads of the core2.
HyperThreading  works nicely for TBX, so the end result is that you get 8 threads of 1.67kh (13.36kh/s total) on the 960 and four threads of 2kh/s (8kh/s total) on the core2.

[racerguy]

anyone have trouble keeping up with the blockchain?  The client I have running on my sisters comp keeps falling behind the blockchain, i've fixed it for now by connecting through the tor network (can't forward the port cause i'm already doing that for my comp).

[EskimoBob]

What do you guys think will be a expected hash rate form the "Intel Core2 Duo CPU E8400  @ 3.00GHz" and what is the recommended number of threads?
 

[racerguy]

What do you guys think will be a expected hash rate form the "Intel Core2 Duo CPU E8400  @ 3.00GHz" and what is the recommended number of threads?
 

I'm getting 1.08khash/s from my sisters q8200 @2.33ghz so my guess would be 1.3-1.4khash for you unless you're mining on linux.  I'm running 1 thread per core atm, haven't tried anymore and don't want to because i'm mining as a background process anyway.

[Bobnova]

Probably around 1.5kh, and two threads as it's a dual core non-HT CPU.

Those things also overclock like mad though, very few fail to make it to 4ghz.