Posts: 1
Threads: 0
Joined: Jul 2017
07-06-2017, 04:36 AM
(This post was last modified: 07-06-2017, 04:38 AM by systemBuilder.
Edit Reason: add some facts.
)
(06-18-2017, 09:16 PM)epixoip Wrote: I've not investigated why Keccak is so much faster. Might have something to do with the compiler.
As far as "fraction of the price" goes:
7700K $300, motherboard $40 == $340 total
1800X $460, motherboard $65 == $525 total
7700K setup is 35% cheaper. Plus the operating costs are lower since it draws less power. Also don't forget that the 1800X lacks onboard graphics, so you'll have to buy a discrete GPU as well. We used a spare 290X that we had lying around. With the 7700K we used the onboard graphics, so no discrete GPU is needed. Plus you can use the onboard graphics for an additional boost in hashrate as well.
Wow what a fanboy post! Let me respond. Pick an 8-core(16 thread) unlocked AMD processor against a 4-core(8 thread) unlocked Intel processor ... Ryzen 1700 vs. i7-7700k. Ryzen 1700 is cheaper ($299 this week). NOBODY buys an i7-7700k without a GPU. NOBODY uses i7-7700k integrated graphics. The AMD Motherboard is cheaper. AMD WINS!
Posts: 1
Threads: 0
Joined: Jul 2017
(07-06-2017, 04:36 AM)systemBuilder Wrote: (06-18-2017, 09:16 PM)epixoip Wrote: I've not investigated why Keccak is so much faster. Might have something to do with the compiler.
As far as "fraction of the price" goes:
7700K $300, motherboard $40 == $340 total
1800X $460, motherboard $65 == $525 total
7700K setup is 35% cheaper. Plus the operating costs are lower since it draws less power. Also don't forget that the 1800X lacks onboard graphics, so you'll have to buy a discrete GPU as well. We used a spare 290X that we had lying around. With the 7700K we used the onboard graphics, so no discrete GPU is needed. Plus you can use the onboard graphics for an additional boost in hashrate as well.
Wow what a fanboy post! Let me respond. Pick an 8-core(16 thread) unlocked AMD processor against a 4-core(8 thread) unlocked Intel processor ... Ryzen 1700 vs. i7-7700k. Ryzen 1700 is cheaper ($299 this week). NOBODY buys an i7-7700k without a GPU. NOBODY uses i7-7700k integrated graphics. The AMD Motherboard is cheaper. AMD WINS!
I had to just register so to reply this answer dumb answer.
AMD Motherboard is cheaper so you still need a GPU. Again the original thread message is that AMD is better CPU. CPU I say it one more time CPU! So you now quote an even worse performing CPU to just win over your mates! I'm not sure.
I think your mommy is calling you!
From reading all messages in this thread, there is more room for Fanboys or who cares, it's irrelevant as is your answer.
End of Line.
Posts: 40
Threads: 5
Joined: Dec 2013
From the results I have seen, there are at least 3 facts to point out:
1) Hashcat for CPUs must be an Intel-paid software or should be.
It''s the single most optimized app in the whole planet capable of showing Kabylake 7700K to be 620% faster than 1800X.
It's simple, plain ridiculous.
The OpenCL coding of hashcat or/and the OpenCL compiler could have a first position in Guiness Book of records for Intel friendly code and AMD hating code.
2) Thank God nobody uses hashcat for CPU password cracking
3) VEGA is faster than 1080 in hashcat GPU cracking.
P.S
Epixoip''s comments and benchmarks of RyZen 1800X revealed one of the biggest AMD haters of AMD writing publicly.
You should immediately stop reading his posts.
Posts: 5,185
Threads: 230
Joined: Apr 2010
Sorry I didn't want to interrupt this thread, but as you commented about the hashcat code there's no way to ignore.
You're absolutely right when you say that the hashcat code is optimized for Intel and NVidia, but you're also absolutely wrong about AMD. Hashcat is very much optimized for AMD, the same way it is as for Intel and NVidia. There's no reason to not optimize for any of them. The project is free since the first minute and I've rejected *every* donation offer since ever. I myself own both NV and AMD cards and before I release anything new, I always test it on all vendors.
You can be sure that hashcat gets the best out of the AMD hardware possible.
You talk about VEGA is faster than 1080. That's nice if it's true. Can you please provide a hashcat -b output? Thanks!
Posts: 347
Threads: 3
Joined: May 2010
(07-10-2017, 10:03 PM)NikosD Wrote: From the results I have seen, there are at least 3 facts to point out:
1) Hashcat for CPUs must be an Intel-paid software or should be.
It''s the single most optimized app in the whole planet capable of showing Kabylake 7700K to be 620% faster than 1800X.
It's simple, plain ridiculous.
The OpenCL coding of hashcat or/and the OpenCL compiler could have a first position in Guiness Book of records for Intel friendly code and AMD hating code.
2) Thank God nobody uses hashcat for CPU password cracking
3) VEGA is faster than 1080 in hashcat GPU cracking.
P.S
Epixoip''s comments and benchmarks of RyZen 1800X revealed one of the biggest AMD haters of AMD writing publicly.
You should immediately stop reading his posts.
So, this thread is still active.
1) Putting the stupid intel shit aside, 620% is for a specific algo, but despite that, the 7700k is indeed faster than the 1800X, that is a fact, and it is not ridiculous.
Your sarcasm has no power here, hashcat code is open, feel free to optimize stuff, and complain about the compilers at the appropriate places if you desire (not here).
2) English is not my first language, I did not know "nobody" meant "a lot", I'm glad a lot of people use hashcat for cpu hashcracking.
3) I do not see a fact here, I see nothing.
Also you made typos in your last part, let me fix it for you :
Quote:NikosD's comments and lack of facts revealed one of the biggest AMD hardcore fanboy writing publicly.
You should immediately stop reading his posts.
Based on facts, and getting back to the topic (even if it was already answered), Intel and Nvidia are better compared to AMD (for hashcat at least, but this is hashcat forum so it is the only thing that matters).
P.S.
Feel free to answer, but be advised that you exceeded your quota of free ranting, any more ranting must be backed up by actual facts.
Posts: 54
Threads: 2
Joined: Mar 2017
I would have thought that it would be quite difficult for hashcat to be open source and be programmed in a way that is blatently pro Intel and anti AMD without anyone noticing....
@NikosD, I don't see why you are still talking about this. You make a claim, people countered it, you reject those arguments and say there isn't any proof, someone then spends money and their own time to prove that you are wrong and now you blame the code, again with no evidence yourself. Does someone now need to explain the code and technical details of why hashcat is fast on Intel compared to AMD???
Posts: 8
Threads: 3
Joined: Apr 2016
07-19-2017, 02:25 PM
(This post was last modified: 07-19-2017, 02:30 PM by kost.)
Partial bench of my 7700k, it failed out half way through due to some self-test salt length error, not sure why:
Code: hashcat (v3.6.0-198-g2dd18339) starting in benchmark mode...
OpenCL Platform #1: NVIDIA Corporation
======================================
* Device #1: GeForce GTX 1080, skipped.
OpenCL Platform #2: Intel(R) Corporation
========================================
* Device #2: Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz, 4080/16320 MB allocatable, 8MCU
Hashtype: MD4
Speed.Dev.#2.....: 1387.6 MH/s (6.00ms)
Hashtype: MD5
Speed.Dev.#2.....: 805.5 MH/s (10.35ms)
Hashtype: Half MD5
Speed.Dev.#2.....: 475.5 MH/s (17.55ms)
Hashtype: SHA1
Speed.Dev.#2.....: 405.3 MH/s (20.60ms)
Hashtype: SHA-256
Speed.Dev.#2.....: 150.1 MH/s (55.65ms)
Hashtype: SHA-384
Speed.Dev.#2.....: 45382.6 kH/s (92.02ms)
Hashtype: SHA-512
Speed.Dev.#2.....: 49863.4 kH/s (83.77ms)
Hashtype: SHA-3 (Keccak)
Speed.Dev.#2.....: 16234.2 kH/s (64.32ms)
Hashtype: SipHash
Speed.Dev.#2.....: 1150.1 MH/s (7.24ms)
Hashtype: Skip32 (PT = $salt, key = $pass)
Speed.Dev.#2.....: 36539.5 kH/s (57.14ms)
Hashtype: RIPEMD-160
Speed.Dev.#2.....: 203.2 MH/s (41.08ms)
Hashtype: Whirlpool
Speed.Dev.#2.....: 6953.6 kH/s (75.09ms)
Hashtype: GOST R 34.11-94
Speed.Dev.#2.....: 4968.4 kH/s (52.31ms)
Hashtype: GOST R 34.11-2012 (Streebog) 256-bit
Speed.Dev.#2.....: 3076.5 kH/s (84.53ms)
Hashtype: GOST R 34.11-2012 (Streebog) 512-bit
Speed.Dev.#2.....: 2739.0 kH/s (94.95ms)
Hashtype: DES (PT = $salt, key = $pass)
Speed.Dev.#2.....: 42340.7 kH/s (49.24ms)
Hashtype: 3DES (PT = $salt, key = $pass)
Speed.Dev.#2.....: 11364.7 kH/s (91.89ms)
Hashtype: phpass, WordPress (MD5), phpBB3 (MD5), Joomla (MD5)
Speed.Dev.#2.....: 227.0 kH/s (17.73ms)
Hashtype: scrypt
Speed.Dev.#2.....: 0 H/s (0.33ms)
Hashtype: PBKDF2-HMAC-MD5
Speed.Dev.#2.....: 233.0 kH/s (33.97ms)
Hashtype: PBKDF2-HMAC-SHA1
Speed.Dev.#2.....: 142.3 kH/s (56.08ms)
Hashtype: PBKDF2-HMAC-SHA256
Speed.Dev.#2.....: 63830 H/s (62.81ms)
Hashtype: PBKDF2-HMAC-SHA512
Speed.Dev.#2.....: 22617 H/s (89.42ms)
Hashtype: Skype
Speed.Dev.#2.....: 514.4 MH/s (16.23ms)
Hashtype: WPA/WPA2
Speed.Dev.#2.....: 17487 H/s (57.81ms)
Hashtype: WPA/WPA2 PMK
Speed.Dev.#2.....: 2988.4 kH/s (0.02ms)
Hashtype: IKE-PSK MD5
Speed.Dev.#2.....: 59111.2 kH/s (70.66ms)
Hashtype: IKE-PSK SHA1
Speed.Dev.#2.....: 33255.0 kH/s (62.79ms)
Hashtype: NetNTLMv1 / NetNTLMv1+ESS
Speed.Dev.#2.....: 790.9 MH/s (10.54ms)
Hashtype: NetNTLMv2
Speed.Dev.#2.....: 55096.3 kH/s (75.82ms)
Hashtype: IPMI2 RAKP HMAC-SHA1
Speed.Dev.#2.....: 69160.8 kH/s (60.38ms)
Hashtype: Kerberos 5 AS-REQ Pre-Auth etype 23
Speed.Dev.#2.....: 5097.2 kH/s (50.60ms)
Hashtype: Kerberos 5 TGS-REP etype 23
Speed.Dev.#2.....: 5281.0 kH/s (48.84ms)
Hashtype: DNSSEC (NSEC3)
Speed.Dev.#2.....: 116.1 MH/s (71.96ms)
Hashtype: PostgreSQL CRAM (MD5)
Speed.Dev.#2.....: 198.3 MH/s (42.13ms)
Hashtype: MySQL CRAM (SHA1)
Speed.Dev.#2.....: 99309.4 kH/s (84.13ms)
Hashtype: SIP digest authentication (MD5)
Speed.Dev.#2.....: 21607.9 kH/s (48.29ms)
Hashtype: SMF (Simple Machines Forum) > v1.1
Speed.Dev.#2.....: 305.9 MH/s (27.28ms)
Hashtype: vBulletin < v3.8.5
Speed.Dev.#2.....: 195.3 MH/s (42.74ms)
Hashtype: vBulletin >= v3.8.5
Speed.Dev.#2.....: 144.4 MH/s (57.86ms)
Hashtype: IPB2+ (Invision Power Board), MyBB 1.2+
Speed.Dev.#2.....: 117.0 MH/s (71.33ms)
Hashtype: WBB3 (Woltlab Burning Board)
Speed.Dev.#2.....: 47388.6 kH/s (88.13ms)
Hashtype: OpenCart
Speed.Dev.#2.....: 56572.4 kH/s (73.83ms)
Hashtype: Joomla < 2.5.18
Speed.Dev.#2.....: 791.3 MH/s (10.53ms)
Hashtype: PHPS
Speed.Dev.#2.....: 194.7 MH/s (42.89ms)
Hashtype: Drupal7
Speed.Dev.#2.....: 1114 H/s (57.17ms)
Hashtype: osCommerce, xt:Commerce
Speed.Dev.#2.....: 511.8 MH/s (16.30ms)
Started: Wed Jul 19 13:19:52 2017
Stopped: Wed Jul 19 13:21:12 2017
Posts: 5,185
Threads: 230
Joined: Apr 2010
Don't mix in latest github version benchmarks. I've optimized the code to better perform on CPUs, so comparison would be invalid.
Posts: 8
Threads: 3
Joined: Apr 2016
(07-19-2017, 05:04 PM)atom Wrote: Don't mix in latest github version benchmarks. I've optimized the code to better perform on CPUs, so comparison would be invalid.
Why wouldn't I? I don't see how it is harmful. If anything it can help OP decide based on latest results.
Posts: 5,185
Threads: 230
Joined: Apr 2010
Because older version performed slower. That means you'd need to update the old benchmarks, too.
|