Author

Topic: [ANN][GRS] Groestlcoin | 1st to activate Segwit | Building Lightning Network - page 219. (Read 558212 times)

newbie
Activity: 10
Merit: 0
Added GTX 780, latest ccminer 0.50 x64

CPU not possible, yamminer doesn't work

Also, looks like you're only getting 7.2MH/s. Please try the 18 April build of ccMiner. You should get 9MHs. See here: http://cryptomining-blog.com/2097-updated-ccminer-0-5-with-improved-performance-for-groestlcoin-grs/

Thanks

That's exactly the ccminer I am using. Same settings. Only 7.2MH/s
sr. member
Activity: 462
Merit: 252
Please use the new link: https://docs.google.com/spreadsheets/d/1hTOUdJZrKqrNDXTIbXgNvJtj_fnpECREcs-eHxaGrZc/edit?usp=sharing

We need NVidia GPU and CPU results.
Please share your settings.

If you have AMD GPU - please also share your settings, but this is lower priority. We've got lots of AMD results, but not enough CPU and NVidia.

Thank you!

Thanks, added some AMDs and CPUs, I don't have any Nvidia cards, but I will add more CPUs next week.
sr. member
Activity: 266
Merit: 250
Added GTX 780, latest ccminer 0.50 x64

CPU not possible, yamminer doesn't work

Also, looks like you're only getting 7.2MH/s. Please try the 18 April build of ccMiner. You should get 9MHs. See here: http://cryptomining-blog.com/2097-updated-ccminer-0-5-with-improved-performance-for-groestlcoin-grs/

Thanks
newbie
Activity: 10
Merit: 0
Hm I took the one for my CPU - 64bit win Haswell. Starts and crashes in a millisecond. Okay, give the other builds a try.

It's an Intel i7 3,4 Ghz, so i think that's not really an old one Smiley

sr. member
Activity: 266
Merit: 250
Added GTX 780, latest ccminer 0.50 x64

CPU not possible, yamminer doesn't work

Thank you for this!

yamminer does work, but looks like you need to try all builds to find the one that works best. There are like 8 builds of it.

Also, on older PCs the original CPU minerd works faster.
newbie
Activity: 10
Merit: 0
Added GTX 780, latest ccminer 0.50 x64

CPU not possible, yamminer doesn't work
sr. member
Activity: 266
Merit: 250
The CPU / GPU spreadsheet on groestlcoin.org

(https://docs.google.com/spreadsheet/ccc?key=0Atmcd1XbwbURdFVhT0w5RXdvcUdKdXVQVUt1blZZSHc#gid=0)

seems to be read-only. I wanted to put my values in but can't  Huh

Please use the new link: https://docs.google.com/spreadsheets/d/1hTOUdJZrKqrNDXTIbXgNvJtj_fnpECREcs-eHxaGrZc/edit?usp=sharing

We need NVidia GPU and CPU results.
Please share your settings.

If you have AMD GPU - please also share your settings, but this is lower priority. We've got lots of AMD results, but not enough CPU and NVidia.

Thank you!
full member
Activity: 238
Merit: 100
Markets GRS/BTC (SHBI) & GRS/LTC (SHLI) opened up:

https://www.swaphole.com/#!market/GRS/BTC
https://www.swaphole.com/#!market/GRS/LTC

Trailing Stop - 2FA - Fast Support
member
Activity: 86
Merit: 10
WOW, MINTPAL, what a stupid choice, you will get nothing more than dumpers! GRS will be dead like HIRO.

Remember this day, It is the day when died Tito Vilanova and GRS follows him.. Cry

the patient cyphermonks of Groestl will not be swayed by shits called Dumpers.



It reminds me of coffee divination self-appointed preachers  Wink
full member
Activity: 156
Merit: 100
Guido
WOW, MINTPAL, what a stupid choice, you will get nothing more than dumpers! GRS will be dead like HIRO.

Remember this day, It is the day when died Tito Vilanova and GRS follows him.. Cry

the patient cyphermonks of Groestl will not be swayed by shits called Dumpers.



have to sell out the first batch of miners
hero member
Activity: 812
Merit: 1000
WOW, MINTPAL, what a stupid choice, you will get nothing more than dumpers! GRS will be dead like HIRO.

Remember this day, It is the day when died Tito Vilanova and GRS follows him.. Cry

the patient cyphermonks of Groestl will not be swayed by shits called Dumpers.

hero member
Activity: 812
Merit: 1000
Good performance on many platforms Smiley
https://online.tugraz.at/tug_online/voe_main2.getvolltext?pCurrPk=58179


To quote from the conclusion:

In this work we have proposed two optimized algorithms for MixBytes, the MDS mixing layer ofGrøstl, which allow to speed up Grøstlon on various platforms. Furthermore, byte slicing provides the possibility to parallelize the Grøstl computation if the registers are large enough and parallel AES S-box table lookups are available. This is the case for Intel processors including the new AES instructions set or in general, using the vpaes implementation. Both implementations show that Grøstl can be implemented efficiently on very different platforms.

Democratizing Mining

Yes.  But just remember - this is only going to help narrow the gap somewhat compared to SHA256.

A really good GPU implementation is still likely to be, say, 8-16x better than a CPU implementation in hashes/second/$.  An ASIC implementation will have a very big gap.

(None of the GPU implementations are any good yet compared to what they should be, at least, until C&C release theirs.)

So don't make too many strong claims about how democratized it will be. Smiley  It's better than Bitcoin's algorithm.  Whether or not it's better than, say, Momentum, is an interesting question.


Hence democratizing… not democratized Smiley
member
Activity: 86
Merit: 10
WOW, MINTPAL, what a stupid choice, you will get nothing more than dumpers! GRS will be dead like HIRO.

Remember this day, It is the day when died Tito Vilanova and GRS follows him.. Cry
dga
hero member
Activity: 737
Merit: 511
Good performance on many platforms Smiley
https://online.tugraz.at/tug_online/voe_main2.getvolltext?pCurrPk=58179


To quote from the conclusion:

In this work we have proposed two optimized algorithms for MixBytes, the MDS mixing layer ofGrøstl, which allow to speed up Grøstlon on various platforms. Furthermore, byte slicing provides the possibility to parallelize the Grøstl computation if the registers are large enough and parallel AES S-box table lookups are available. This is the case for Intel processors including the new AES instructions set or in general, using the vpaes implementation. Both implementations show that Grøstl can be implemented efficiently on very different platforms.

Democratizing Mining

Yes.  But just remember - this is only going to help narrow the gap somewhat compared to SHA256.

A really good GPU implementation is still likely to be, say, 8-16x better than a CPU implementation in hashes/second/$.  An ASIC implementation will have a very big gap.

(None of the GPU implementations are any good yet compared to what they should be, at least, until C&C release theirs.)

So don't make too many strong claims about how democratized it will be. Smiley  It's better than Bitcoin's algorithm.  Whether or not it's better than, say, Momentum, is an interesting question.
hero member
Activity: 812
Merit: 1000
hero member
Activity: 812
Merit: 1000
I tried the apk on my galaxy nexus gsm:

Groestlhash: 10.4264 KH/s

So these initial benchmarks are getting ~10KH/thread
So would a 2 core ARM chip support 4 threads i.e. 40KH/s ?
Now I m sure we're gonna end up with much more than 10KH/s per thread.

Even at this early stage the power efficiencies are interesting. Im gonna make a big assumption but if 4 cores > 8 threads @10KH/thread = 80kH

R9290        150w      ~9000 KH/s
Cortex A9   1.9W          ~80 KH/s  --->> so theoretically 190w required to push 8000kH/s

So even at this early stage (permitting my assumptions about multi-thread performance are true) the Cortex is only ~30% less power efficient (hashrate/watts) then an R9290

Optimizations leading to 15kH/s or 20KH/s per thread or more should mean the Cortex A9 is more power efficient than the AMD GPU.

These new Cortex-A17 procs (2014) seems considerably more powerrful than the A9

hero member
Activity: 812
Merit: 1000
Good performance on many platforms Smiley
https://online.tugraz.at/tug_online/voe_main2.getvolltext?pCurrPk=58179


To quote from the conclusion:

In this work we have proposed two optimized algorithms for MixBytes, the MDS mixing layer ofGrøstl, which allow to speed up Grøstlon on various platforms. Furthermore, byte slicing provides the possibility to parallelize the Grøstl computation if the registers are large enough and parallel AES S-box table lookups are available. This is the case for Intel processors including the new AES instructions set or in general, using the vpaes implementation. Both implementations show that Grøstl can be implemented efficiently on very different platforms.

Democratizing Mining
member
Activity: 86
Merit: 10
Why so low price?  Roll Eyes

At this time, marketing is more important than the technical maturity

Miners not interested the possibility of mining on ARM / CPU, less power consumption, but what is the current profit.
hero member
Activity: 812
Merit: 1000
ARM's 2014 processors will blow today's smartphone chips away, with 3x the performance or 1/4 the battery drain

http://www.theverge.com/2012/10/30/3576560/arm-cortex-a57-cortex-a53-cpu-core

… just sayin'


NEW! A new Grøstl-256 NEON implementation using bitslicing (48.6 cycles/byte on Cortex-A8, Hercules eCAFE) is available in eBASH.


^^ sounds promising
Jump to: