Wolf's XMR/BCN/DSH CPUMiner - 2x speed compared to LucasJones' - NEW 06/20/2014

[Hotmetal]

Why does it automatically set thread count to ~cpu count?

Using your formula:

[2014-08-03 13:19:33] accepted: 12/12 (100.00%), 514.93 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:34] accepted: 13/13 (100.00%), 513.02 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:35] accepted: 14/14 (100.00%), 528.15 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:35] accepted: 15/15 (100.00%), 525.17 H/s at diff 444 (yay!!!)

Perhaps this should be added to the original post? Its a BIG improvement.

To answer my own question, it appears as though the improvement is noticed only when the server has a large number of cpus / cache etc.
Older machines seem to perform better with a larger amount of threads, even when cpu count is lower.

I'm breaking > 10 000 H/s without even breaking a sweat across a few machines.
Thank you

[ballot]

trying to compile for win 64 bit
with this instruction
https://bitcointalk.org/index.php?topic=594470.msg6542760#msg6542760
tried many different ./configure
everytime it success compile but the compiled binary crash while trying to run also i tried to strip the exe too 

[5w00p]

8 Core Xeon gets 370 hash/sec all day, all night with 8 threads in ubuntu 14.04 w/ hugepages=24.

[5w00p]

Why does it automatically set thread count to ~cpu count?

Using your formula:

[2014-08-03 13:19:33] accepted: 12/12 (100.00%), 514.93 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:34] accepted: 13/13 (100.00%), 513.02 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:35] accepted: 14/14 (100.00%), 528.15 H/s at diff 444 (yay!!!)
[2014-08-03 13:19:35] accepted: 15/15 (100.00%), 525.17 H/s at diff 444 (yay!!!)

Perhaps this should be added to the original post? Its a BIG improvement.

To answer my own question, it appears as though the improvement is noticed only when the server has a large number of cpus / cache etc.
Older machines seem to perform better with a larger amount of threads, even when cpu count is lower.

I'm breaking > 10 000 H/s without even breaking a sweat across a few machines.


Thank you

really, because 10,000/528 = ~19.  19 is a "few machines?"

edit: and that diff is pathetically low.  444 is for a pentium 2 or someshit.

[primer-]

Looking for instructions on how to enable MADV_HUGEPAGE on Centos 6.5 , anyone ??

[Hotmetal]

really, because 10,000/528 = ~19.  19 is a "few machines?"

edit: and that diff is pathetically low.  444 is for a pentium 2 or someshit.

Yes, 19 is a "few machines". There are plenty more to use across different pools / coins.
Difficulty is assigned by the network and will scale up and down. 444 is the starting difficulty for that pool and its gotten considerably higher now.

Lighten up, you sound like you're having a bad day

[primer-]

Looking for instructions on how to enable MADV_HUGEPAGE on Centos 6.5 , anyone ??

You can't, because CentOS is garbage. I had to add in a flag to disable some optimizations for that shit OS.

I know it is, is there any other way to get it to run on centos ? If i compile on ubuntu and run on centos it complains about glibc version (2.14 required, centos runs 2.12). Any other way ? Is it possible to include glibc in the static binary ?

EDIT : Removing MADV_HUGEPAGE from cpu-miner.c fixes the problem but with a significant loss in speed. I need the extra speed!!

[pallas]

Looking for instructions on how to enable MADV_HUGEPAGE on Centos 6.5 , anyone ??

You can't, because CentOS is garbage. I had to add in a flag to disable some optimizations for that shit OS.

I know it is, is there any other way to get it to run on centos ? If i compile on ubuntu and run on centos it complains about glibc version (2.14 required, centos runs 2.12). Any other way ? Is it possible to include glibc in the static binary ?

EDIT : Removing MADV_HUGEPAGE from cpu-miner.c fixes the problem but with a significant loss in speed. I need the extra speed!!

the static binary can include everything.
did you try:

./configure --enable-static

or

CFLAGS='-static'

?

[primer-]

Looking for instructions on how to enable MADV_HUGEPAGE on Centos 6.5 , anyone ??

You can't, because CentOS is garbage. I had to add in a flag to disable some optimizations for that shit OS.

I know it is, is there any other way to get it to run on centos ? If i compile on ubuntu and run on centos it complains about glibc version (2.14 required, centos runs 2.12). Any other way ? Is it possible to include glibc in the static binary ?

EDIT : Removing MADV_HUGEPAGE from cpu-miner.c fixes the problem but with a significant loss in speed. I need the extra speed!!

the static binary can include everything.
did you try:

./configure --enable-static

or

CFLAGS='-static'

?

Yes i have, i can't get past curl requirements even though i compiled curl from source with static flags.. I am not sure glibc is included in the static bin, can anyone confirm, dont want to waste time troubleshooting curl

[primer-]

Curl 7.31.1 compiled with :
./configure --disable-shared --enable-static --prefix=/usr/local --disable-ldap --disable-sspi
make -j 4;make install;

miner still errors out:
./autogen.sh
./configure CFLAGS="-static'
checking for the version of libcurl... 7.37.1
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2

Any ideas ?

[Hotmetal]

Curl 7.31.1 compiled with :
./configure --disable-shared --enable-static --prefix=/usr/local --disable-ldap --disable-sspi
make -j 4;make install;

miner still errors out:
./autogen.sh
./configure CFLAGS="-static'
checking for the version of libcurl... 7.37.1
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2

Any ideas ?

./configure is looking at curl prefix in /usr (depending on the distro). You might want configure curl like:
./configure --disable-shared --enable-static --prefix=/tmp/curl --disable-ldap --disable-sspi
make ; make install

then the miner:
./autogen.sh
./configure CFLAGS="-static' --with-libcurl=/tmp/curl
make

[primer-]

Curl 7.31.1 compiled with :
./configure --disable-shared --enable-static --prefix=/usr/local --disable-ldap --disable-sspi
make -j 4;make install;

miner still errors out:
./autogen.sh
./configure CFLAGS="-static'
checking for the version of libcurl... 7.37.1
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2

Any ideas ?

./configure is looking at curl prefix in /usr (depending on the distro). You might want configure curl like:
./configure --disable-shared --enable-static --prefix=/tmp/curl --disable-ldap --disable-sspi
make ; make install

then the miner:
./autogen.sh
./configure CFLAGS="-static' --with-libcurl=/tmp/curl
make

Tried it, no luck. Curl author should be hanged in public!!

[primer-]

Hm... try this:

Code:

make distclean
./autogen.sh
CFLAGS="-I/tmp/curl/include" LDFLAGS="-static -L/tmp/curl/lib" ./configure --with-libcurl=/tmp/curl

Same error,
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2


Dynamic compilation works just fine...

[primer-]

Hm... try this:

Code:

make distclean
./autogen.sh
CFLAGS="-I/tmp/curl/include" LDFLAGS="-static -L/tmp/curl/lib" ./configure --with-libcurl=/tmp/curl

Same error,
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2


Dynamic compilation works just fine...

Pastebin config.log.

http://pastebin.com/UMLA4K6B

[primer-]

Hm... try this:

Code:

make distclean
./autogen.sh
CFLAGS="-I/tmp/curl/include" LDFLAGS="-static -L/tmp/curl/lib" ./configure --with-libcurl=/tmp/curl

Same error,
checking for libcurl >= version 7.15.2... yes
checking whether libcurl is usable... no
configure: error: Missing required libcurl >= 7.15.2


Dynamic compilation works just fine...

Pastebin config.log.

http://pastebin.com/UMLA4K6B

You built libcurl with zlib, but you did not build a static copy; therefore it looks for one because you forced static compilation and cannot find it, thus the failure.

PFF, stupid curl configuration scrit, version 7.37 is buggy as hell!!

I got it compiled now with 7.34 :

Static CURL :

Code:

wget http://curl.haxx.se/download/curl-7.34.0.tar.gz
 ./configure --disable-shared --enable-static
make -j 4;make install;

cpu miner

Code:

./autogen.sh
./configure CFLAGS="-static"
make -j 4

[tgt]

1600 H/s:

4x Opteron 6344 (48 physical cores @ 2.6ghz static)
64G ram (12x4GB ddr3-
14.04, built from wolf's git source

inside KVM with cpu-passthrough (couldn't get it to run on 12.04 which is host's bare metal OS, could potentially be faster)
sysctl -w vm.nr_hugepages=144

pool reporting > 2kh, miner at 1600-1650kh/s.

https://i.imgur.com/qdQkiBi.png

Could probably get better performance by a) running it on baremetal, b) using numactl and a better task scheduler.

[liteon]

1600 H/s:

4x Opteron 6344 (48 physical cores @ 2.6ghz static)
64G ram (12x4GB ddr3-
14.04, built from wolf's git source

inside KVM with cpu-passthrough (couldn't get it to run on 12.04 which is host's bare metal OS, could potentially be faster)
sysctl -w vm.nr_hugepages=144

pool reporting > 2kh, miner at 1600-1650kh/s.



Could probably get better performance by a) running it on baremetal, b) using numactl and a better task scheduler.

Reduce it to sysctl -w vm.nr_hugepages=48
Will be at least 20% more power.

[5w00p]

really, because 10,000/528 = ~19.  19 is a "few machines?"

edit: and that diff is pathetically low.  444 is for a pentium 2 or someshit.

Yes, 19 is a "few machines". There are plenty more to use across different pools / coins.
Difficulty is assigned by the network and will scale up and down. 444 is the starting difficulty for that pool and its gotten considerably higher now.

Lighten up, you sound like you're having a bad day

Ok, well I guess every day I have is a "bad day" compared to the world you live in where six figures USD worth of hardware is at your disposal and you nonchalantly work them like rented mules mining XMR like it's going out of style.  Kick a brother down a Xeon server or two. will ya? 

[5w00p]

1600 H/s:

4x Opteron 6344 (48 physical cores @ 2.6ghz static)
64G ram (12x4GB ddr3-
14.04, built from wolf's git source

inside KVM with cpu-passthrough (couldn't get it to run on 12.04 which is host's bare metal OS, could potentially be faster)
sysctl -w vm.nr_hugepages=144

pool reporting > 2kh, miner at 1600-1650kh/s.



Could probably get better performance by a) running it on baremetal, b) using numactl and a better task scheduler.

I hit 1030 or so with 23 threads on AWS - you probably need to reduce your thread count.

Wolf, why would he want to REDUCE his hashrate?

He reports 1600 hash/sec and you report 1030. 

[Hotmetal]

Ok, well I guess every day I have is a "bad day" compared to the world you live in where six figures USD worth of hardware is at your disposal and you nonchalantly work them like rented mules mining XMR like it's going out of style.  Kick a brother down a Xeon server or two. will ya? 

It's more in the region of 7 figures USD worth of hardware. At the moment I'm just playing around on the portion of the development cluster.
Need to make sure things are running optimally before XMR mining is moved to production. I'm responsible that way