Bitcoin Forum
January 18, 2018, 08:53:56 AM *
News: Latest stable version of Bitcoin Core: 0.15.1  [Torrent].
 
   Home   Help Search Donate Login Register  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 [29] 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 ... 165 »
  Print  
Author Topic: [ANN]: cpuminer-opt v3.7.10, open source optimized multi-algo CPU miner  (Read 390416 times)
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 27, 2016, 08:43:17 PM
 #561

Hey man, I have a suggestion for you. You don't need to find out if the hardware have AES-NI in the beginning. If you use EVP alone, it finds out by itself if the hardware is AES-NI capable, and if that's the case, it uses it automatically. You don't have to do anything additional.

I already tried it and it works perfectly. Just my two cents. 

I don't know what EVP is? Where does it come into play at compile time, run time?

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
1516265637
Hero Member
*
Offline Offline

Posts: 1516265637

View Profile Personal Message (Offline)

Ignore
1516265637
Reply with quote  #2

1516265637
Report to moderator
pallas
Legendary
*
Offline Offline

Activity: 1526


Black Belt Developer


View Profile
April 27, 2016, 09:05:55 PM
 #562

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 27, 2016, 10:55:22 PM
 #563

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 27, 2016, 11:09:00 PM
 #564

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 27, 2016, 11:18:09 PM
 #565

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 27, 2016, 11:21:38 PM
 #566

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 01:28:11 AM
 #567

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.


Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 28, 2016, 01:51:13 AM
 #568

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 02:30:08 AM
 #569

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

Did you mistype your first sentence? If evp was faster than Wolf I would have expected a more emphatic response.
If it is faster how do I implement it?

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 28, 2016, 02:43:42 AM
 #570

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

Did you mistype your first sentence? If evp was faster than Wolf I would have expected a more emphatic response.
If it is faster how do I implement it?

It is faster and you already did. EVP is already implemented in the file hodl.cpp It will select AES-NI automatically if the machine has it. All you need to do is get rid of the Wolf code and let your own code by itself; it will select AES-NI automatically, you don't have to change anything.

Just try it and you will see.

Note: get rid of all the code that you are using for the AES-NI section in the hodlcoin source, you don't need it.

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 03:31:41 AM
 #571

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

Did you mistype your first sentence? If evp was faster than Wolf I would have expected a more emphatic response.
If it is faster how do I implement it?

It is faster and you already did. EVP is already implemented in the file hodl.cpp It will select AES-NI automatically if the machine has it. All you need to do is get rid of the Wolf code and let your own code by itself; it will select AES-NI automatically, you don't have to change anything.

Just try it and you will see.

Note: get rid of all the code that you are using for the AES-NI section in the hodlcoin source, you don't need it.

You're full of shit, Wolf is 30% faster.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 28, 2016, 03:39:27 AM
 #572

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

Did you mistype your first sentence? If evp was faster than Wolf I would have expected a more emphatic response.
If it is faster how do I implement it?

It is faster and you already did. EVP is already implemented in the file hodl.cpp It will select AES-NI automatically if the machine has it. All you need to do is get rid of the Wolf code and let your own code by itself; it will select AES-NI automatically, you don't have to change anything.

Just try it and you will see.

Note: get rid of all the code that you are using for the AES-NI section in the hodlcoin source, you don't need it.

You're full of shit, Wolf is 30% faster.

That's what you get when you are trying to help.

Never mind, forget all that I told you. Have a nice one.

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 03:45:27 AM
 #573

Evp, AFAIK, is an abstraction layer of openssl which automatically selects the best instruction set based on the current cpu. Not sure it applies to this project, maybe to hodlcoin or others with "standard" algos.

OK, it's associated with openssl, The only thing close I found for the acronym was enhanced virus protection.
Few coins use openssl for their hashing algos so I don't see the point, and hodl already uses it for both the
AES_NI and non-AES_NI implementations.

@joblo:

    You are already using it in hodl.cpp:

                        EVP_EncryptInit(&ctx, EVP_aes_256_cbc(), key, iv);
                        EVP_EncryptUpdate(&ctx, cacheMemoryOperatingData, &outlen1, cacheMemoryOperatingData2, cacheMemorySize);
                        EVP_EncryptFinal(&ctx, cacheMemoryOperatingData + outlen1, &outlen2);
                        EVP_CIPHER_CTX_cleanup(&ctx);

All that I'm saying is that you shouldn't do anything else but let it run by itself. in other words, don't try to select the best arch; just use native and don't fork the source code to use 'special case' for AES-NI, since EVP will select it automatically. Meaning: don't use '#ifdef AES_NI', it's not necessary.

Hope that I'm clear this time Smiley
 

I beg to differ. The NO_AES_NI check is needed to prevent the compiler from trying to compile AES_NI instructions
on an incompatible CPU. Without it the compile fails.

Then just get rid of that code, you don't need that part because like I already said: EVP algo will detect AES-NI capability and will use it automatically. You don't need extra code for that, just get ride of it.

I already did and it works perfect.

If you've been digging into the code you realize I have implemented both existing hodl miners, the original hodlminer and the
AES optimized hodlminer-wolf. The original hodlminer uses EVP so why is it slower than Wolf? Openssl/evp is likely a general
purpose tool while the Wolf miner is optimized for the specific task.

If you've followed previous discuusions about design changes you'll know I'm a tough sell. Without going into
detail there are only three things that would motivate me to make design changes to a stable product:
support for more algos, higher hashrates or Windows support.



Ok, the optimized Wolf miner runs slower than EVP when the machine is AES-NI capable. Again, I'm just giving you a suggestion, you are free to do whatever you want of course Smiley

Whenever you got the time, just give it a try on a machine with AES-NI running just EVP and you will see what I'm talking about. And I'm just talking about the part for hodlcoin, wich is the only part that I've checked out.

Did you mistype your first sentence? If evp was faster than Wolf I would have expected a more emphatic response.
If it is faster how do I implement it?

It is faster and you already did. EVP is already implemented in the file hodl.cpp It will select AES-NI automatically if the machine has it. All you need to do is get rid of the Wolf code and let your own code by itself; it will select AES-NI automatically, you don't have to change anything.

Just try it and you will see.

Note: get rid of all the code that you are using for the AES-NI section in the hodlcoin source, you don't need it.

You're full of shit, Wolf is 30% faster.

That's what you get when you are trying to help.

Never mind, forget all that I told you. Have a nice one.

Are you serious? That's a pretty childish response.

If evp is so much better and is already implemented in the orignal hodlminer Why is it so much slower than Wolf?

You've made a substantial claim without any data. My data contradicts your claims, Your move.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 28, 2016, 04:39:11 AM
 #574

Look, I was giving you a suggestion, but you said I'm full of shit. So that's it.

If you are saying that EVP on machines that are AES-NI is slower than Wolf's, it's all fine, I really don't care. On my AMD FX machine EVP is faster than Wolfs. Period.

I already wasted too much time in something that it's not even my business, I thought I was giving you a good advise, my bad.

I apologize if my claim was wrong. (on my machine it's true, but maybe on yours isn't, I don't care anymore)

Like I said, have a nice night, it's time for to go to bed.

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
pallas
Legendary
*
Offline Offline

Activity: 1526


Black Belt Developer


View Profile
April 28, 2016, 07:36:37 AM
 #575

Wolf0 doesn't have an AMD cpu and it is indeed slower on that kind of processor.
Still, EVP can only be applied to algos which use plain AES, which are very few.
The others need specific code, like joblo is making.

joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 01:48:18 PM
 #576

Look, I was giving you a suggestion, but you said I'm full of shit. So that's it.

If you are saying that EVP on machines that are AES-NI is slower than Wolf's, it's all fine, I really don't care. On my AMD FX machine EVP is faster than Wolfs. Period.

I already wasted too much time in something that it's not even my business, I thought I was giving you a good advise, my bad.

I apologize if my claim was wrong. (on my machine it's true, but maybe on yours isn't, I don't care anymore)

Like I said, have a nice night, it's time for to go to bed.

There's no need to pout. You're making a claim that the AES_NI optimized hodlminer-wolf is slower than the
original on AMD CPUs with AES_NI. That is a serious claim and no one else has reported this so you need proof.
This goes beyond being childish, you have made a serious charge against a respected dev whio was paid to
develop the AES_NI enhanced miner.

You've already gone too far, you need to follow through with proof or make a full retraction with an apology to Wolf.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
GoldTiger69
Hero Member
*****
Offline Offline

Activity: 501


View Profile WWW
April 28, 2016, 06:21:50 PM
 #577

Look, I was giving you a suggestion, but you said I'm full of shit. So that's it.

If you are saying that EVP on machines that are AES-NI is slower than Wolf's, it's all fine, I really don't care. On my AMD FX machine EVP is faster than Wolfs. Period.

I already wasted too much time in something that it's not even my business, I thought I was giving you a good advise, my bad.

I apologize if my claim was wrong. (on my machine it's true, but maybe on yours isn't, I don't care anymore)

Like I said, have a nice night, it's time for to go to bed.

There's no need to pout. You're making a claim that the AES_NI optimized hodlminer-wolf is slower than the
original on AMD CPUs with AES_NI. That is a serious claim and no one else has reported this so you need proof.
This goes beyond being childish, you have made a serious charge against a respected dev whio was paid to
develop the AES_NI enhanced miner.

You've already gone too far, you need to follow through with proof or make a full retraction with an apology to Wolf.

You have to calm down, I haven't make any charge against nobody, what are you talking about? listen to your self!

All I did was telling you that on my machine, EVP was using AES-NI and it is faster than wolfs code. Period. There is no 'charges' against nobody, please!!

And like I said, that only applied to the source code on the hodlcoin, I even told you the file where it was used: hodl.cpp

My suggestion was that you could use EVP alone (whitout wolfs code) because it uses AES-NI automatically. Where is the accusation??

You are making a stupid and simple suggestion a matter way, way out of proportion. You doesn't even understand that I was trying to suggest something to improve your code.

So, end of the story. I'm not gonna continue with this stupid conversation, it went from a simple suggestion to "charges against a respected dev'. Go figure!

I can help you to restore/recover your wallet or password.
https://bitcointalk.org/index.php?topic=1234619.0
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 06:41:59 PM
 #578

Look, I was giving you a suggestion, but you said I'm full of shit. So that's it.

If you are saying that EVP on machines that are AES-NI is slower than Wolf's, it's all fine, I really don't care. On my AMD FX machine EVP is faster than Wolfs. Period.

I already wasted too much time in something that it's not even my business, I thought I was giving you a good advise, my bad.

I apologize if my claim was wrong. (on my machine it's true, but maybe on yours isn't, I don't care anymore)

Like I said, have a nice night, it's time for to go to bed.

There's no need to pout. You're making a claim that the AES_NI optimized hodlminer-wolf is slower than the
original on AMD CPUs with AES_NI. That is a serious claim and no one else has reported this so you need proof.
This goes beyond being childish, you have made a serious charge against a respected dev whio was paid to
develop the AES_NI enhanced miner.

You've already gone too far, you need to follow through with proof or make a full retraction with an apology to Wolf.

You have to calm down, I haven't make any charge against nobody, what are you talking about? listen to your self!

All I did was telling you that on my machine, EVP was using AES-NI and it is faster than wolfs code. Period. There is no 'charges' against nobody, please!!

And like I said, that only applied to the source code on the hodlcoin, I even told you the file where it was used: hodl.cpp

My suggestion was that you could use EVP alone (whitout wolfs code) because it uses AES-NI automatically. Where is the accusation??

You are making a stupid and simple suggestion a matter way, way out of proportion. You doesn't even understand that I was trying to suggest something to improve your code.

So, end of the story. I'm not gonna continue with this stupid conversation, it when from a simple suggestion to "charges against a respected dev'. Go figure!

You don't realize the sum of your claims. You claim the original hodlminer is faster than Wolf's AES_NI optimized version.
I'm not saying it isn't true in your case but you need to show proof to convince me. I can't replicate your results and
they in fact prove the opposite.  You're on AMD and I'm on Intel, there're may be something there that should be pursued but you
haven't given me anything I can work with. You're coming accross as a troll. Prove me wrong.

The original hodlminer was not promoted to use AES_NI and Wolf was contracted to implement one. His version
produced higher hashrates and no one has complained. Either no one is mining hodl with AES enabled AMD CPUs,
they aren't paying attention, or you're wrong. There is some foundation to your claim because Wolf is optimized
for Intel, its performance on AMD may not be optimum. You can provide the answer if you stop throwing a hissy
fit and stick to the issue. It's up to you.

Edit: In your last 5 posts you haven't provided one piece of new information. Instead you're acting like a spoiled child
whining because I didn't give you praise. Grow up.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
Jullovich1
Newbie
*
Offline Offline

Activity: 3


View Profile
April 28, 2016, 06:56:14 PM
 #579

Is a windows version available?




1BER28Sc19azZjKVs4CUWcqBvo4o4cFb5W
joblo
Legendary
*
Offline Offline

Activity: 1008


View Profile
April 28, 2016, 11:05:25 PM
 #580

I've retested blakecoin in cpuminer-opt v3.1.18 and it is now listed as supported.

Principal developer of cpuminer-opt, the optimized multi-algo CPU miner.
BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
https://bitcointalk.org/index.php?topic=1326803.0
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 [29] 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 ... 165 »
  Print  
 
Jump to:  

Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!