Topic: [ANN]: cpuminer-opt v3.8.8.1, open source optimized multi-algo CPU miner - page 66. (Read 444067 times)

I used the ./build.sh script to compile with HW SHA enabled and libssl-dev 1.0.2g-1ubuntu4.6 installed.

It is slower without the HW SHA.

./cpuminer -a m7m --benchmark

         **********  cpuminer-opt 3.6.6  ***********
     A CPU miner with multi algo support and optimized for CPUs
     with AES_NI and AVX extensions.
     BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
     Forked from TPruvot's cpuminer-multi with credits
     to Lucas Jones, elmad, palmd, djm34, pooler, ig0tik3d,
     Wolf0, Jeff Garzik and Optiminer.

CPU: Intel(R) Celeron(R) CPU J3455 @ 1.50GHz
CPU features: SSE2 AES SHA
SW built on Jul 23 2017 with GCC 5.4.0
SW features: SSE2 AES
Algo features: SSE2 AES AVX SHA
Start mining with SSE2 AES

[2017-07-23 21:24:36] 4 miner threads started, using 'm7m' algorithm.
[2017-07-23 21:24:55] CPU #2: 131.07 kH, 6991.75 H/s
[2017-07-23 21:24:55] CPU #3: 131.07 kH, 6983.48 H/s
[2017-07-23 21:24:55] Total: 262.14 kH, 13.98 kH/s, 40C
[2017-07-23 21:24:55] CPU #0: 131.07 kH, 6799.19 H/s
[2017-07-23 21:24:55] CPU #1: 131.07 kH, 6794.54 H/s
[2017-07-23 21:25:00] CPU #3: 34.92 kH, 7024.97 H/s
[2017-07-23 21:25:00] Total: 428.13 kH, 27.61 kH/s, 40C
[2017-07-23 21:25:00] CPU #2: 34.96 kH, 6746.14 H/s
[2017-07-23 21:25:00] CPU #0: 34.00 kH, 6795.39 H/s
[2017-07-23 21:25:00] CPU #1: 33.97 kH, 6724.83 H/s
[2017-07-23 21:25:05] CPU #3: 35.12 kH, 6922.89 H/s
[2017-07-23 21:25:05] Total: 138.05 kH, 27.19 kH/s, 39C
[2017-07-23 21:25:05] CPU #2: 33.73 kH, 6864.23 H/s
[2017-07-23 21:25:05] CPU #0: 33.98 kH, 6790.99 H/s
[2017-07-23 21:25:05] CPU #1: 33.62 kH, 6689.45 H/s
[2017-07-23 21:25:10] CPU #3: 34.62 kH, 6955.82 H/s
[2017-07-23 21:25:10] Total: 135.95 kH, 27.30 kH/s, 40C

I used debian 9 with openssl 1.1 and got the same speed as ubuntu with the libssl-dev 1.0.2g-1ubuntu4.6 installed.

./cpuminer -a m7m --benchmark

         **********  cpuminer-opt 3.6.6  ***********
     A CPU miner with multi algo support and optimized for CPUs
     with AES_NI and AVX extensions.
     BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
     Forked from TPruvot's cpuminer-multi with credits
     to Lucas Jones, elmad, palmd, djm34, pooler, ig0tik3d,
     Wolf0, Jeff Garzik and Optiminer.

CPU: Intel(R) Celeron(R) CPU J3455 @ 1.50GHz
CPU features: SSE2 AES SHA
SW built on Jul 24 2017 with GCC 6.3.0
SW features: SSE2 AES SHA
Algo features: SSE2 AES AVX SHA
Start mining with SSE2 AES SHA

[2017-07-24 02:23:37] 4 miner threads started, using 'm7m' algorithm.
[2017-07-24 02:23:50] CPU #3: 131.07 kH, 10.00 kH/s
[2017-07-24 02:23:50] Total: 131.07 kH, 10.00 kH/s, 37C
[2017-07-24 02:23:50] CPU #2: 131.07 kH, 9817.99 H/s
[2017-07-24 02:23:51] CPU #1: 131.07 kH, 9474.24 H/s
[2017-07-24 02:23:51] CPU #0: 131.07 kH, 9293.29 H/s
[2017-07-24 02:23:55] CPU #0: 37.17 kH, 9407.19 H/s
[2017-07-24 02:23:55] CPU #1: 37.90 kH, 8790.50 H/s
[2017-07-24 02:23:55] CPU #3: 50.01 kH, 9838.35 H/s
[2017-07-24 02:23:55] Total: 256.15 kH, 37.85 kH/s, 38C
[2017-07-24 02:23:55] CPU #2: 49.09 kH, 9702.19 H/s
[2017-07-24 02:24:00] CPU #1: 43.95 kH, 9022.16 H/s
[2017-07-24 02:24:00] CPU #3: 49.19 kH, 9912.67 H/s
[2017-07-24 02:24:00] Total: 179.41 kH, 38.04 kH/s, 38C

Debian 9 has libssl-dev 1.1.

libssl-dev 1.1 is needed but Ubuntu 16.04 has 1.0.2. None of the major distributions seem to have it yet.

How did you compile, did you change any options or just use the build script?
The option to add sha is "-msha" but the default compile probably includes it.
You might have to do something like "-maes -mavx" "-maes -msse4.2" instead
of the default "-march=native" to compile without sha.

Edit: correction, Goldmont doesn't have AVX.

Which package needs to version 1.1?
libssl-dev or libcurl4-openssl-dev?

I have Ubuntu 16.04

How do I compile without SHA?

I thought the hash rate was pretty good for a 10w CPU.

I just enjoy playing, not any profit really.

That's very interesting. SHA is not mentioned here...

https://ark.intel.com/products/95594/Intel-Celeron-Processor-J3455-2M-Cache-up-to-2_3-GHz

But here it is...

https://en.wikipedia.org/wiki/Goldmont

OK so you have a Goldmont and it is the only Intel CPU to support SHA at this time.

It looks like all the boxes are checked, CPU, compiler, build and algo. The only question seems to be
the version of openssl. Cpuminer doesn't check the version but it needs to be 1.1 to have HW SHA support.
The only way to be sure is to compare the hash rate with a non-SHA compile.

Edit: strictly speaking, when cpuminer says it's mining with SHA it means it is using the openssl version
instead of its own SW version, but it's no guarantee that openssl is using the HW SHA instructions.

What about shipping a Docker image or atleast a Dockerfile with all the latest dependencies self contained?What about shipping a Docker image or atleast a Dockerfile with all the latest dependencies self contained?

It looks like HW SHA to me? This is only a 10w CPU.

         **********  cpuminer-opt 3.6.6  ***********
     A CPU miner with multi algo support and optimized for CPUs
     with AES_NI and AVX extensions.
     BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
     Forked from TPruvot's cpuminer-multi with credits
     to Lucas Jones, elmad, palmd, djm34, pooler, ig0tik3d,
     Wolf0, Jeff Garzik and Optiminer.

CPU: Intel(R) Celeron(R) CPU J3455 @ 1.50GHz
CPU features: SSE2 AES SHA
SW built on Jul  4 2017 with GCC 5.4.0
SW features: SSE2 AES SHA
Algo features: SSE2 AES AVX SHA
Start mining with SSE2 AES SHA

[2017-07-23 15:21:13] Starting Stratum on stratum+tcp://xmg.suprnova.cc:7128
[2017-07-23 15:21:13] 4 miner threads started, using 'm7m' algorithm.

[2017-07-23 15:21:13] Stratum difficulty set to 8
[2017-07-23 15:21:17] m7m block 1418024, diff 7.180
[2017-07-23 15:21:19] m7m block 1418025, diff 7.180
[2017-07-23 15:21:19] CPU #2: 19.29 kH, 8515.43 H/s
[2017-07-23 15:21:19] CPU #0: 22.44 kH, 9851.52 H/s
[2017-07-23 15:21:19] CPU #1: 21.60 kH, 9509.43 H/s
[2017-07-23 15:21:19] CPU #3: 20.52 kH, 9059.83 H/s
[2017-07-23 15:21:27] CPU #2: 67.76 kH, 8803.45 H/s
[2017-07-23 15:21:27] Accepted 1/1 (100%), 132.33 kH, 37.22 kH/s, 44C

Thanks but that CPU doesn't have HW SHA. This information is displayed
when the program starts.

Compiling is still the biggest challenge to using SHA. Few Linux distros include openssl 1.1 which added SHA support.

Intel J3455 ASRock J3455M with 1866 memory dual channel
--benchmark results
 
lbry         Total: 5287.13 kH, 1341.57 kH/s, 43C
sha256t      Total: 4643.44 kH, 6343.56 kH/s, 42C
skein         Total: 15.05 MH, 3223.62 kH/s, 45C
myr-groestl   Total: 7336.60 kH, 1480.04 kH/s, 44C
m7m         Total: 254.99 kH, 38.28 kH/s, 45C

An advertised feature of Ryzen PRO is "Built-in AES 128-bit Encryption engine" which I wonder if it will benefit mining.

It should be the equivalent of 2 Ryzen 1700X or 1800X CPUs, so a pretty darned good miner if you can afford the price.

Any thoughts on AMD Threadripper for mining?

Looking at it.

Any plans to add the skunk algorithm from SIGT?  https://bitcointalksearch.org/topic/ann-signatum-new-algorithm-fair-launch-no-premine-2030529

Disappointing results. AVX2 provides 256 bit vectors vs 128 bit for AVX. Ignoring any memory accesses the AVX2
code uses half the instructions to do the same processing so should be twice as fast.

Most of the AVX2 code is in the x11 chain so it should see a bigger difference than x17.

I had heard that Ryzen had a "faked" AVX2 and your test seems to prove it and that it performs worse than AVX.
I was concerned it wouldn't be any faster than AVX but never considered that it would be slower.
The only reason I can think of to include it is for compatibility with any hard coded AVX2. IMO they should have
left it out.

It appears that AVX2 should continue to be avoided on AMD CPUs.

Thanks for testing.

Preliminary results:

Ryzen 1700X @ stock settings with 8 threads set in the batch file.

Lyra2v2 = 1001 Kh/s on avx2
             = 1016 Kh/s on avx
Timetravel = 799 Kh/s on avx2
                = 843 Kh/s on avx
X11          = 686 Kh/s on avx2
                = 711 Kh/s on avx
X17          = 294 Kh/s on avx2
                = 294 Kh/s on avx

X17 was the same for both options, but I was surprised to see avx beating avx2 for the other algorithms tested. My understanding was that it would likely be the other way around....

Hello, with a Intel CPU i7 3820 what is the best .exe files to use on windows ?
cpuminer-aes-avx.exe
cpuminer-aes-avx2.exe
cpuminer-aes-sse42.exe
cpuminer-sse2.exe
cpuminer-sse42.exe

?

Lyra2v2, Lyra2z and Deep have the most AVX2 and should see the biggest improvement. All the X algos, qubit, timetravel
and a few others also have varying amounts of AVX2 code.

If you can list the algos you would like tested I can try it out on both Ryzen R7 1700X and R5 1600 CPUs. Possibly on Threadripper too towards the end of summer if I talk myself into getting one.

I have not found a way to compile on Windows with SHA. I have tried 3 different mingw environments without success.
One user was successful but didn't share his procedure.