Pages:
Author

Topic: [ANN]: cpuminer-opt v3.8.8.1, open source optimized multi-algo CPU miner - page 57. (Read 444067 times)

member
Activity: 473
Merit: 18
Should I be worried that my CPU, software, algo all supports AVX but it's not being used? Or does using AVX2 mean that AVX doesn't matter?

Code:
CPU: Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz
CPU features: SSE2 AES AVX AVX2
SW built on Nov  2 2017 with GCC 5.4.0
SW features: SSE2 AES AVX AVX2
Algo features: SSE2 AES AVX AVX2
Start mining with SSE2 AES AVX2
it uses the best available implementation, so yes, AVX doesn't matter when you have AVX2
newbie
Activity: 3
Merit: 0
Should I be worried that my CPU, software, algo all supports AVX but it's not being used? Or does using AVX2 mean that AVX doesn't matter?

Code:
CPU: Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz
CPU features: SSE2 AES AVX AVX2
SW built on Nov  2 2017 with GCC 5.4.0
SW features: SSE2 AES AVX AVX2
Algo features: SSE2 AES AVX AVX2
Start mining with SSE2 AES AVX2
full member
Activity: 239
Merit: 100
already try v3.7.2 but there is 2 red line inside my bat file
"stratum_recv_line failed"
"stratum connection interrupted"

pool problem
member
Activity: 312
Merit: 10
already try v3.7.2 but there is 2 red line inside my bat file
"stratum_recv_line failed"
"stratum connection interrupted"
full member
Activity: 1848
Merit: 148
tell me on which of the files will be mining above the algorithm m7m on the processor i7-3770
legendary
Activity: 1470
Merit: 1114
One thing I have found is that neoscrypt fails to initialize in the version i have compiled (both avx2 and sha)

Benchmark is broken for neoscrypt but pool mining works. Benchmark is broken for a few algos and I don't really
care to try to fix them.

all the algos in the table i posted benchmark fine with your version, but in this case, it actually breaks completely, even with real pools


p.s.
can you enable the api to report the hashrate before the 1st accepted share?

I can't do much if it's only your cross compile that's failing.

The hash rate is only meaningful when a share is submitted. Reporting the per-thread hash rate on every scan
would increase the traffic.
member
Activity: 473
Merit: 18
One thing I have found is that neoscrypt fails to initialize in the version i have compiled (both avx2 and sha)

Benchmark is broken for neoscrypt but pool mining works. Benchmark is broken for a few algos and I don't really
care to try to fix them.

all the algos in the table i posted benchmark fine with your version, but in this case, it actually breaks completely, even with real pools


p.s.
can you enable the api to report the hashrate before the 1st accepted share?
legendary
Activity: 1470
Merit: 1114
One thing I have found is that neoscrypt fails to initialize in the version i have compiled (both avx2 and sha)

Benchmark is broken for neoscrypt but pool mining works. Benchmark is broken for a few algos and I don't really
care to try to fix them.
legendary
Activity: 1470
Merit: 1114
Good to read you got it working. You should check that the compiled SW features match your CPU architecture for best
performance.

It should already, because it was compiled in the real machine, or it should not?

edit: in you opening post there are some minor typos here? (CLAGS is CFLAGS? obvious for an advanced user but deadly for a novice if it's the case, I'm the second type)

Users with openssl 1.0.1 (Ubuntu 14.04) may get better perforance by adding "-DUSE_SPH_SHA" to CLAGS.

Thanks for pointing out the typo, fixed. I don't expect too many newbies to be using old distros. You could consider the
error a competency test, if you can't find the error you shouldn't be trying to customize the compile.  You passed. Wink

The Virtualbox emulated CPU may not have the same architecture as the host and may be missing some newer features,
particularly with older versions of Virtualbox. On the other hand the old CPUs are not likely to have them either.
It's just something else to watch for.
sr. member
Activity: 840
Merit: 267
Chad Hodler since 2013
Good to read you got it working. You should check that the compiled SW features match your CPU architecture for best
performance.

It should already, because it was compiled in the real machine, or it should not?

edit: in you opening post there are some minor typos here? (CLAGS is CFLAGS? obvious for an advanced user but deadly for a novice if it's the case, I'm the second type)

Users with openssl 1.0.1 (Ubuntu 14.04) may get better perforance by adding "-DUSE_SPH_SHA" to CLAGS.
legendary
Activity: 1470
Merit: 1114
Can I use cpuminer-opt for LAN mining?
How should it be configured
thank you for your help!

I don't understand the question. If you mean solo mining connecting to the wallet instead of a pool,
then yes it should work. But be sure to include htttp:// in the URL because it will default to stratum otherwise.
full member
Activity: 168
Merit: 100
yescryptR16 algorithm support is not perfect
There will be an error

[2017-11-01 23:32:09] CPU #7: 448 H, 39.64 H/s
[2017-11-01 23:32:09] CPU #19: 430 H, 38.08 H/s
[2017-11-01 23:32:09] CPU #16: 491 H, 43.42 H/s
[2017-11-01 23:32:09] CPU #0: 490 H, 43.32 H/s
[2017-11-01 23:32:10] Stratum connection failed: Failed to connect to pool.bilbotel.fr port 6234: Connection refused
[2017-11-01 23:32:10] ...retry after 10 seconds
[2017-11-01 23:32:21] Stratum difficulty set to 1


That is a pool issue, likely overload as discussed in the yenten thread.

Can I use cpuminer-opt for LAN mining?
How should it be configured
thank you for your help!
legendary
Activity: 1470
Merit: 1114
yescryptR16 algorithm support is not perfect
There will be an error

[2017-11-01 23:32:09] CPU #7: 448 H, 39.64 H/s
[2017-11-01 23:32:09] CPU #19: 430 H, 38.08 H/s
[2017-11-01 23:32:09] CPU #16: 491 H, 43.42 H/s
[2017-11-01 23:32:09] CPU #0: 490 H, 43.32 H/s
[2017-11-01 23:32:10] Stratum connection failed: Failed to connect to pool.bilbotel.fr port 6234: Connection refused
[2017-11-01 23:32:10] ...retry after 10 seconds
[2017-11-01 23:32:21] Stratum difficulty set to 1


That is a pool issue, likely overload as discussed in the yenten thread.
legendary
Activity: 1470
Merit: 1114
Thanks for fix, you still have old link though so we cant actually download this fixed miner

Links have been updated. Pleas delete the bad links from your post.
full member
Activity: 154
Merit: 100
cpuminer-opt-3.7.2 is released.

Fixed yescryptr16.

Changed default sha256 and sha512 to openssl. This should be used when compiling with openssl 1.0.2
or higher (Ubuntu 16.04). This should increase the hashrate for m7m, xevan, skein, myr-gr & others
when openssl 1.0.2 is installed. This is due to AVX and AVX2 optimizations added to openssl 1.0.2.

Users with openssl 1.0.1 (Ubuntu 14.04) may get better performance by adding "-DUSE_SPH_SHA" to CLAGS.
When using sph_sha or openssl-1.0.1 the miner will still display AVX and AVX2 mining even though neither
sph_sha nor openssl-1.0.1 have those optimizations.

Windows binaries are compiled with -DUSE_SPH_SHA and won't get the speedup.

git: https://github.com/JayDDee/cpuminer-opt

tarball: https://drive.google.com/file/d/0B0lVSGQYLJIZZXoyZkRNLUc2cE0/view?usp=sharing

Windows binaries: https://drive.google.com/file/d/0B0lVSGQYLJIZTTV6YnhnV3F0eWc/view?usp=sharing

Thanks to all who helped testing SHA, particularly 4ward who discovered performance on Intel CPUs
had improved in openssl 1.0.2.


Thanks for fix, you still have old link though so we cant actually download this fixed miner
full member
Activity: 168
Merit: 100
yescryptR16 algorithm support is not perfect
There will be an error

[2017-11-01 23:32:09] CPU #7: 448 H, 39.64 H/s
[2017-11-01 23:32:09] CPU #19: 430 H, 38.08 H/s
[2017-11-01 23:32:09] CPU #16: 491 H, 43.42 H/s
[2017-11-01 23:32:09] CPU #0: 490 H, 43.32 H/s
[2017-11-01 23:32:10] Stratum connection failed: Failed to connect to pool.bilbotel.fr port 6234: Connection refused
[2017-11-01 23:32:10] ...retry after 10 seconds
[2017-11-01 23:32:21] Stratum difficulty set to 1


legendary
Activity: 1470
Merit: 1114
Will lyra2z330 gain benefit from thoese ssl/sha optimizations, too?

Another update, sorry for the spam.

Managed to install devtoolset-6 (GCC 6.3.1), it was quite straightforward:

... cut ...

You're trying to compile for a CPU archietcture simulated by Virtualbox. Even if the compile was successful it wouldn't work properly
on real hardware. I do not support attempts to cross compile on Linux. You're on your own.

FYI, with the help of a sysadmin, I got it to work (well, he did, most of it, and I ignore the details), so this miner is now running on Centos 6.9 and real hardware.

Good to read you got it working. You should check that the compiled SW features match your CPU architecture for best
performance.
legendary
Activity: 1470
Merit: 1114
cpuminer-opt-3.7.2 is released.

Fixed yescryptr16.

Changed default sha256 and sha512 to openssl. This should be used when compiling with openssl 1.0.2
or higher (Ubuntu 16.04). This should increase the hashrate for m7m, xevan, skein, myr-gr & others
when openssl 1.0.2 is installed. This is due to AVX and AVX2 optimizations added to openssl 1.0.2.

Users with openssl 1.0.1 (Ubuntu 14.04) may get better performance by adding "-DUSE_SPH_SHA" to CFLAGS.
When using sph_sha or openssl-1.0.1 the miner will still display AVX and AVX2 mining even though neither
sph_sha nor openssl-1.0.1 have those optimizations.

Windows binaries are compiled with -DUSE_SPH_SHA and won't get the speedup.

git: https://github.com/JayDDee/cpuminer-opt

tarball: https://drive.google.com/file/d/0B0lVSGQYLJIZamNFVnZGT1FFT1k/view?usp=sharing

Windows binaries: https://drive.google.com/file/d/0B0lVSGQYLJIZMXl6SWtmUE4yTm8/view?usp=sharing

Thanks to all who helped testing SHA, particularly 4ward who discovered performance on Intel CPUs
had improved in openssl 1.0.2.

Edit: Links updated for 3.7.2
sr. member
Activity: 840
Merit: 267
Chad Hodler since 2013
Will lyra2z330 gain benefit from thoese ssl/sha optimizations, too?

Another update, sorry for the spam.

Managed to install devtoolset-6 (GCC 6.3.1), it was quite straightforward:

... cut ...

You're trying to compile for a CPU archietcture simulated by Virtualbox. Even if the compile was successful it wouldn't work properly
on real hardware. I do not support attempts to cross compile on Linux. You're on your own.

FYI, with the help of a sysadmin, I got it to work (well, he did, most of it, and I ignore the details), so this miner is now running on Centos 6.9 and real hardware.
Pages:
Jump to: