Pages:
Author

Topic: [ANN]: cpuminer-opt v3.8.8.1, open source optimized multi-algo CPU miner - page 101. (Read 444043 times)

hero member
Activity: 630
Merit: 502
I'm planning on upgrading to the AMD Ryzen R7-1800X next month so which coins should I look to mine?

Your choice.

Let me rephase by which coins currently offer the best ROI? Or are none of them worth mining as it will cost more for electricity? How comparable are current AMD CPUs compared to Intel and will there be optimizations for Ryzen?
legendary
Activity: 1470
Merit: 1114
I'm planning on upgrading to the AMD Ryzen R7-1800X next month so which coins should I look to mine?

Your choice.
hero member
Activity: 630
Merit: 502
I'm planning on upgrading to the AMD Ryzen R7-1800X next month so which coins should I look to mine?
legendary
Activity: 1470
Merit: 1114
I just pushed an update to git for the aclocal.m4 issue causing compile problems
when cloned from git. Those who have run into this problem may find it easier to
compile. It was supposed to be in 3.5.6 but got lost in one of my many resets this release.
legendary
Activity: 1470
Merit: 1114
It nearly want you to have a loader that will analyze your current CPU capability and use the proper exe  Grin so many of them  Shocked

I do the best I can. The startup messages are really important to make sure you're getting the best performance,
and can explain crashes. You want the CPU and SW features to match and the algo will choose the best it can.
full member
Activity: 224
Merit: 100
CryptoLearner
It nearly want you to have a loader that will analyze your current CPU capability and use the proper exe  Grin so many of them  Shocked
legendary
Activity: 1470
Merit: 1114

cpuminer-sse2.exe
cpuminer-sse42.exe
cpuminer-aes-sse42.exe
cpuminer-aes-avx.exe
cpuminer-aes-avx2.exe

It's actually sse4.2 not 4.1, I just checked.


And so it shall be. These will be the names of the Windows binaries in the next release.
The included README.txt file will have descriptions to help.

Prepare your scripts.
legendary
Activity: 1260
Merit: 1046
Thank you Joblo for the 3.5.6 release.
I compiled it with the build.sh and Ubuntu 16.04 : all works fine.
I use it for ZCoin and all is ok :-).

Is there a possibility to ask in the command line for a particular difficulty and staying allways at this difficulty ?

I don't know if the command line diff really does anything useful. Some pools allow diff as a password paramter or
offer a different port for different diff. In most pools vardiff usually works well adusting automatically.
Ok, thank you again :-).
legendary
Activity: 1470
Merit: 1114
Thank you Joblo for the 3.5.6 release.
I compiled it with the build.sh and Ubuntu 16.04 : all works fine.
I use it for ZCoin and all is ok :-).

Is there a possibility to ask in the command line for a particular difficulty and staying allways at this difficulty ?

I don't know if the command line diff really does anything useful. Some pools allow diff as a password paramter or
offer a different port for different diff. In most pools vardiff usually works well adusting automatically.
legendary
Activity: 1260
Merit: 1046
Thank you Joblo for the 3.5.6 release.
I compiled it with the build.sh and Ubuntu 16.04 : all works fine.
I use it for ZCoin and all is ok :-).

Is there a possibility to ask in the command line for a particular difficulty and staying allways at this difficulty ?
legendary
Activity: 1470
Merit: 1114
im just used builtin script    ./build.sh   after its finish with error  change is neede from " march-native" to msse2   am im right , means i have to edit it by hand ?


You don't have to edit the script, just enter the commands manually.
full member
Activity: 144
Merit: 100
Eager to learn
im just used builtin script    ./build.sh   after its finish with error  change is neede from " march-native" to msse2   am im right , means i have to edit it by hand ?
legendary
Activity: 1470
Merit: 1114
ok , thanks i will do

Makefile:2451: die Regel für Ziel „algo/echo/aes_ni/cpuminer-hash.o“ scheiterte
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Fehler 1
make[2]: *** Auf noch nicht beendete Prozesse wird gewartet …
mv -f algo/hodl/.deps/cpuminer-hodl_arith_uint256.Tpo algo/hodl/.deps/cpuminer-hodl_arith_uint256.Po
mv -f algo/hodl/.deps/cpuminer-hodl.Tpo algo/hodl/.deps/cpuminer-hodl.Po
mv -f algo/haval/.deps/cpuminer-haval.Tpo algo/haval/.deps/cpuminer-haval.Po
make[2]: Verzeichnis „/home/boss/joblo6“ wird verlassen
Makefile:3496: die Regel für Ziel „all-recursive“ scheiterte
make[1]: *** [all-recursive] Fehler 1
make[1]: Verzeichnis „/home/boss/joblo6“ wird verlassen
Makefile:671: die Regel für Ziel „all“ scheiterte
make: *** [all] Fehler 2
strip: ‚cpuminer‛: Keine solche Datei
boss@boss-MS-7253:~/joblo6$

cpu is Athlon 64 x2 5600+ with nstruction set extensions   
MMX
SSE
SSE2
3DNow!
SSE3

same error on all other amd i own athlon II x250 , 64x2 4200+

What options did you compile with? You might have to help with the translation.
I noticed this "algo/echo/aes" in the text. If this is where the error occurred your
cpu does not have aes. Try compiling with -msse3, if that doesn't work -msse2.
It shouldn't make a difference.
full member
Activity: 144
Merit: 100
Eager to learn
ok , thanks i will do

Makefile:2451: die Regel für Ziel „algo/echo/aes_ni/cpuminer-hash.o“ scheiterte
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Fehler 1
make[2]: *** Auf noch nicht beendete Prozesse wird gewartet …
mv -f algo/hodl/.deps/cpuminer-hodl_arith_uint256.Tpo algo/hodl/.deps/cpuminer-hodl_arith_uint256.Po
mv -f algo/hodl/.deps/cpuminer-hodl.Tpo algo/hodl/.deps/cpuminer-hodl.Po
mv -f algo/haval/.deps/cpuminer-haval.Tpo algo/haval/.deps/cpuminer-haval.Po
make[2]: Verzeichnis „/home/boss/joblo6“ wird verlassen
Makefile:3496: die Regel für Ziel „all-recursive“ scheiterte
make[1]: *** [all-recursive] Fehler 1
make[1]: Verzeichnis „/home/boss/joblo6“ wird verlassen
Makefile:671: die Regel für Ziel „all“ scheiterte
make: *** [all] Fehler 2
strip: ‚cpuminer‛: Keine solche Datei
boss@boss-MS-7253:~/joblo6$

cpu is Athlon 64 x2 5600+ with nstruction set extensions   
MMX
SSE
SSE2
3DNow!
SSE3

same error on all other amd i own athlon II x250 , 64x2 4200+
legendary
Activity: 1470
Merit: 1114
looks gooad an simple , but all the instructions works togeher as mentonied ?

this one im not sure  "cpuminer-aes-sse42"    but you are the experts

until now i was not able to compile  the amd´s with ss2,sse3 support on cpuminer-opt   thats why im using epsylon´s miner

Post your compile errors and i'll look into it, -marh=native works for Intel but AMD is trickier. The key features
in decreasing performance: AVX2, AVX, AES, SSE4.2, SSE2. Choose the best that will compile and run, AES needs
to be specified seperately, is "-maes -msse4.2", "-maes -mavx".

AES-SSE42 is the difference between the first Nehalem that didn't have AES and Westmere that did. How that maps
to AMD is unknown to me. It will be trial and error. Felix may be able to help as he is an AMD user, though he doesn't
compile.
full member
Activity: 144
Merit: 100
Eager to learn
looks gooad an simple , but all the instructions works togeher as mentonied ?

this one im not sure  "cpuminer-aes-sse42"    but you are the experts

until now i was not able to compile  the amd´s with sse2,sse3 support on cpuminer-opt   thats why im using epsylon´s miner
legendary
Activity: 1470
Merit: 1114
and if it could work on Amd´s sse2 only cpu´s  e.g. sse2,sse3 instruction-set

with epsylons miner i  squezze out from an athlon II x250  98 kh/s at "timetravel"  im sure joblo can top it Wink  to move over 100

btw.  v3.5.6 no gain in my case  looks like on the top of my stuff without burn them

Not for SSE2, only AVX and AVX2. sse3 means nothing to cpuminer, the next jump in performance after sse2 is AES and AVX.
Let me know if multi is faster, I can always import Epsylon's improvements.
hero member
Activity: 700
Merit: 500
core2 seems more appropriate than corei7, but i also like plain cpuminer-sse2.exe Cheesy

Less typing vs a reference to the actual arch name. I guess the arch name isn't much use for users
of the binaries, only to those that compile. I may drop core from all the names. How this? It shows
a logical progression of features, more user frendly maybe. SSE4.2 (it's actually 4.2 i just checked)
is insignificant to cpuminer but it helps distinguish core2 from early core-i.

cpuminer-sse2
cpuminer-sse42
cpuminer-aes-sse42
cpuminer-aes-avx
cpuminer-aes-avx2

It's actually sse4.2 not 4.1, I just checked.


i support this, what do others think?
legendary
Activity: 1470
Merit: 1114
core2 seems more appropriate than corei7, but i also like plain cpuminer-sse2.exe Cheesy

Less typing vs a reference to the actual arch name. I guess the arch name isn't much use for users
of the binaries, only to those that compile. I may drop core from all the names. How this? It shows
a logical progression of features, more user frendly maybe. SSE4.2 (it's actually 4.2 i just checked)
is insignificant to cpuminer but it helps distinguish core2 from early core-i.

cpuminer-sse2
cpuminer-sse42
cpuminer-aes-sse42
cpuminer-aes-avx
cpuminer-aes-avx2

It's actually sse4.2 not 4.1, I just checked.
full member
Activity: 144
Merit: 100
Eager to learn
and if it could work on Amd´s sse2 only cpu´s  e.g. sse2,sse3 instruction-set

with epsylons miner i  squezze out from an athlon II x250  98 kh/s at "timetravel"  im sure joblo can top it Wink  to move over 100

btw.  v3.5.6 no gain in my case  looks like on the top of my stuff without burn them
Pages:
Jump to: