1 2017-04-01 18:51:17 <arubi> devices: https://paste.ubuntu.com/24294492/
 2 2017-04-01 18:51:17 <arubi> shot in the dark, did anyone ever set up oclvanitygen or related in a VM with pcie passthrough to a GPU?  I can get gpu passthrough to the guest, I'm pretty certain (done this before a long time ago), my issue is getting oclvanitygen to use the gpu device.  it lists two opencl devices, both at index '0:', the cpu /and/ the gpu are both index '0:', so no matter which I choose the cpu does the work.  this is my output for getting the list of
 3 2017-04-01 18:52:02 <arubi> alternatively, if anyone can explain how I can easily rent gpu compute for oclvanitygen, that would also be good :)
 4 2017-04-01 18:53:56 <arubi> that space in the last line appears verbatim in the output
 5 2017-04-01 18:54:12 <arubi> er, the space right in the beginning
 6 2017-04-01 20:23:22 <Victorsueca> arubi: if you're going to rent GPU power for that you should use split key generation then
 7 2017-04-01 20:26:06 <arubi> Victorsueca, I wanna run it myself on some platform.  I did that a while ago, rented a bunch of servers running gpu's, but I think it might be cheaper now with amazon etc, it's not for financial use, so no money is involved with that key
 8 2017-04-01 20:26:46 <arubi> not for identification either, really just looking for one of the many keys that fits what I wanna do :)
 9 2017-04-01 20:29:38 <Victorsueca> arubi: then somebody might as well generate it on his own computer and share the private key with you or you're searching for still some grade of security?
10 2017-04-01 20:30:42 <arubi> it's going to take a while to find a key with one computer.  I calculated about a year's worth of searching with one gpu at 80 million keys a second
11 2017-04-01 20:30:58 <arubi> I don't mind letting someone else generate it for me, if they're up to it :)
12 2017-04-01 20:31:27 <Victorsueca> gee, what pattern are you searching for?
13 2017-04-01 20:33:05 <arubi> it's a pattern in the hash160, not the address, it's about 6-7 bytes of pattern
14 2017-04-01 20:33:43 <Victorsueca> in the ripemd thing you mean?
15 2017-04-01 20:33:59 <arubi> yep
16 2017-04-01 20:35:44 <arubi> hmm maybe more, now that I think about it it might take more than a year with just one..
17 2017-04-01 20:36:24 <arubi> need to go over this, hopefully I didn't miss a factor of x100 :), still it might be doable in a year, just more expensive
18 2017-04-01 20:38:21 <Victorsueca> maybe you need a rendering farm for this or something....
19 2017-04-01 20:39:07 <Victorsueca> or a ripemd160 ASIC + sah256 ASIC
20 2017-04-01 20:39:47 <arubi> I wanted to rent a rendering\compute resource, right, that's what I did a long time ago for something unrelated
21 2017-04-01 20:40:07 <arubi> I think amazon, google, offer gpu instances now, but I never tried
22 2017-04-01 20:41:04 <arubi> but first I wanna see that I can get oclvanitygen working, try to set up the hash160 pattern as it's just accepting addresses
23 2017-04-01 23:27:33 <omarshibli> I came up with this method:
24 2017-04-01 23:27:33 <omarshibli> I'm trying to figure out a way to use BIP32 child key derivation with an arbitrary string rather than a number or 'm/*' path.
25 2017-04-01 23:27:34 <omarshibli> 1. for a given input compute the sha512 hash
26 2017-04-01 23:27:35 <omarshibli> 2. covert output to hex
27 2017-04-01 23:27:37 <omarshibli> 3. breakup the hex number into 32 chunks, each chunk 4 characters
28 2017-04-01 23:27:39 <omarshibli> 4. map each chunk to decimal
29 2017-04-01 23:27:41 <omarshibli> 5. join all decimals with '/'
30 2017-04-01 23:27:43 <omarshibli> 6. prepend 'm/0'
31 2017-04-01 23:27:45 <omarshibli> the output would be in this form: m/0/321/323/323...
32 2017-04-01 23:27:47 <omarshibli> Is this secure key derivation?
33 2017-04-01 23:27:55 <omarshibli> a
34 2017-04-01 23:27:58 <omarshibli> 1. for a given input compute the sha512 hash
35 2017-04-01 23:27:58 <omarshibli> I came up with this method:
36 2017-04-01 23:27:58 <omarshibli> I'm trying to figure out a way to use BIP32 child key derivation with an arbitrary string rather than a number or 'm/*' path.
37 2017-04-01 23:28:00 <omarshibli> 2. covert output to hex
38 2017-04-01 23:28:02 <omarshibli> 3. breakup the hex number into 32 chunks, each chunk 4 characters
39 2017-04-01 23:28:04 <omarshibli> 4. map each chunk to decimal
40 2017-04-01 23:28:06 <omarshibli> 5. join all decimals with '/'
41 2017-04-01 23:28:08 <omarshibli> 6. prepend 'm/0'
42 2017-04-01 23:28:10 <omarshibli> the output would be in this form: m/0/321/323/323...
43 2017-04-01 23:28:12 <omarshibli> Is this secure key derivation?