hashcat Forum
Catalyst v13.1 - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Deprecated; Previous versions (https://hashcat.net/forum/forum-29.html)
+--- Forum: General Help (https://hashcat.net/forum/forum-8.html)
+--- Thread: Catalyst v13.1 (/thread-1963.html)

Pages: 1 2 3


Catalyst v13.1 - atom - 01-18-2013

I've checked out the new Catalyst v13.1 driver a bit.

All test have been done on Ubuntu 12.04.1 lts.

Performance comparison are against Catalyst v12.8.

Here are my results:

On hd7970:
  • Good: Oracle-DES speed increased from 1200M to 1310M
  • Bad: ASIC hangs still occour randomly (to often to ignore)
  • Bad: SHA1 speed dropped from 2850M to 2160M
  • Bad: NTLM speed dropped from 16000M to 15000M
  • Bad: descrypt speed dropped from 83M to 60M (was at 103M with Catalyst v12.1)

On hd6990:
  • Good: No ASIC hangs
  • Bad: SHA512 speed dropped from 220M to 174M
  • Bad: Oracle-DES speed dropped from 282M to 230M

Both cards:
  • Good: No more kernel compiler segfaults
  • Good: All unit-tests passed
  • Bad: Still segfaults when accessing ADL and X11 is not running
  • Bad: All other algorithms speed dropped slightly (5% or less)
  • Bad: Installed memory still reported incorrectly
  • Bad: 32-bit OpenCL libraries are still not copied on 64-bit OS's



RE: Catalyst v13.1 - undeath - 01-18-2013

looks like another one to skip. AMD fails so much. Thank you for testing, atom!


RE: Catalyst v13.1 - gummeln - 01-18-2013

I still get the nasty ERROR: clBuildProgram() -11 with AMD 7870, W7 X64 and new 13.1 Cats.


RE: Catalyst v13.1 - undeath - 01-18-2013

nobody ever said this driver is officially supported.


RE: Catalyst v13.1 - ati6990 - 01-18-2013

Amd Teams sucked at driversupport ....


RE: Catalyst v13.1 - gummeln - 01-18-2013

(01-18-2013, 12:16 PM)undeath Wrote: nobody ever said this driver is officially supported.

I know, I just wished it worked Wink


RE: Catalyst v13.1 - Hash-IT - 01-18-2013

I have said this before but this driver problem is up to us to sort out, not atom.

We are paying customers of AMD's products so we "should" have some weight to our complaints. If anyone is interested in seeing a better driver release please take the time to e-mail AMD support and tell them.

Some time ago AMD made a derogatory comment about hashcat-plus, they said they were not interested in our "little hash cracking program".

Please help atom, myself and eventually yourselves by e-mailing the customer support at AMD to complain. If you were very keen there is nothing wrong with using a few e-mail address's more than once Smile

If I am honest hashcat-plus is the ONLY reason I bought my AMD GPU. It will be the ONLY reason I will be buying more in the future. If you are the same then please tell AMD !

I understand there is a chance AMD will simply ignore our comments, however by doing nothing we get the drivers we deserve.

Thanks.


RE: Catalyst v13.1 - eljolot - 01-18-2013

(01-18-2013, 01:35 PM)Hash-IT Wrote: I have said this before but this driver problem is up to us to sort out, not atom.

We are paying customers of AMD's products so we "should" have some weight to our complaints. If anyone is interested in seeing a better driver release please take the time to e-mail AMD support and tell them.

Some time ago AMD made a derogatory comment about hashcat-plus, they said they were not interested in our "little hash cracking program".

Please help atom, myself and eventually yourselves by e-mailing the customer support at AMD to complain. If you were very keen there is nothing wrong with using a few e-mail address's more than once Smile

If I am honest hashcat-plus is the ONLY reason I bought my AMD GPU. It will be the ONLY reason I will be buying more in the future. If you are the same then please tell AMD !

I understand there is a chance AMD will simply ignore our comments, however by doing nothing we get the drivers we deserve.

Thanks.
little? >.< that's offending, let's send thousands of mails Tongue


RE: Catalyst v13.1 - Hash-IT - 01-18-2013

(01-18-2013, 03:32 PM)eljolot Wrote: little? >.< that's offending, let's send thousands of mails Tongue

Yeah ! Smile

It would be best to write them slightly differently each time and over a period of days / weeks otherwise they will assume they are being spammed.

Also please remember not to be rude back to them. Just make your case that you don't find their driver support very good and be polite. Remember you are the paying customer but being rude to them will just make them difficult.


RE: Catalyst v13.1 - eljolot - 01-18-2013

you right Big Grin I guess we should to make point why we want that, maybe they are no taking this serious or are lazy to see what's wrong in his code.

We can make point that the hash cracking is a way to test how secure the password encryptation is.

and that is not only a little program, is a program that also can decrypt payloads of a virus like http://hashcat.net/oclGaussCrack/


(01-18-2013, 05:00 PM)Hash-IT Wrote:
(01-18-2013, 03:32 PM)eljolot Wrote: little? >.< that's offending, let's send thousands of mails Tongue

Yeah ! Smile

It would be best to write them slightly differently each time and over a period of days / weeks otherwise they will assume they are being spammed.

Also please remember not to be rude back to them. Just make your case that you don't find their driver support very good and be polite. Remember you are the paying customer but being rude to them will just make them difficult.