shawnp0wers
Newbie
 Offline
Activity: 19
Merit: 0
|
 |
January 08, 2012, 02:01:26 AM |
|
I got pooler's code to compile for PPC OSX as well (he made the code work for big endian systems now, thanks pooler!), but it fails on some systems I tested on. I'm going to recompile on Monday with a different machine and see how it goes. If I get a good binary, I'll post that as well.
ssvb's code would likely work better for PPC, but I'm having more serious problems compiling it. There are missing header files, etc. I'm sure it's because OSX 10.4 is so old. (I don't have any machines running 10.5, so I haven't tested ssvb's code there)
Anyway, I'll post a PPC binary if I can get it to work well. If I ever get ssvb's code to compile, I'll post that as well.
|
|
|
|
|
|
|
|
Unlike traditional banking where clients have only a few account numbers, with Bitcoin people can create an unlimited number of accounts (addresses). This can be used to easily track payments, and it improves anonymity.
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
|
|
|
|
ZedZedNova
Sr. Member
Offline
Activity: 475
Merit: 265
Ooh La La, C'est Zoom!
|
|
January 08, 2012, 07:11:04 AM |
|
I got pooler's code to compile for PPC OSX as well (he made the code work for big endian systems now, thanks pooler!), but it fails on some systems I tested on. I'm going to recompile on Monday with a different machine and see how it goes. If I get a good binary, I'll post that as well.
ssvb's code would likely work better for PPC, but I'm having more serious problems compiling it. There are missing header files, etc. I'm sure it's because OSX 10.4 is so old. (I don't have any machines running 10.5, so I haven't tested ssvb's code there)
Anyway, I'll post a PPC binary if I can get it to work well. If I ever get ssvb's code to compile, I'll post that as well.
I built pooler's code on my PowerMac G5 (PowerMac9,1 ppc970fx, 1.8 GHz) running Ubuntu 11.10 and achieved ~1.45 kH/s. ssvb's code on the same machine provides ~3.10-3.25 kH/s. I used CFLAGS="-O3 -mcpu=G5 -mtune=G5 -maltivec -fstrict-alias" when I built both. - Zed |
No mining at the moment.
|
|
|
shawnp0wers
Newbie
Offline
Activity: 19
Merit: 0
|
|
January 08, 2012, 06:09:27 PM |
|
I built pooler's code on my PowerMac G5 (PowerMac9,1 ppc970fx, 1.8 GHz) running Ubuntu 11.10 and achieved ~1.45 kH/s. ssvb's code on the same machine provides ~3.10-3.25 kH/s.
I used CFLAGS="-O3 -mcpu=G5 -mtune=G5 -maltivec -fstrict-alias" when I built both.
Cool, I didn't know about the -mtune flag, thanks. Still, I have to keep OSX on my machines, so compiling is... challenging.  |
|
|
|
|
ZedZedNova
Sr. Member
Offline
Activity: 475
Merit: 265
Ooh La La, C'est Zoom!
|
|
January 08, 2012, 06:52:41 PM |
|
I used CFLAGS="-O3 -mcpu=G5 -mtune=G5 -maltivec -fstrict-aliasing" when I built both.
I hate when I do stuff like that. I left off a critical three letters. Memory is not so good when it's late. - Zed |
No mining at the moment.
|
|
|
|
pooler (OP)
|
|
January 11, 2012, 10:04:25 PM
Last edit: January 12, 2012, 11:51:25 PM by pooler |
|
New version!The changes only affect 64-bit systems; they should bring an up to 8% speed increase to Intel processors, and a 20% increase to CPUs of the AMD K10 family. Binary for 64-bit Windows (courtesy of diki): https://github.com/downloads/pooler/cpuminer/pooler-cpuminer-win64-20120111.zipA binary for 64-bit Mac OS X will be available soon. As always, all benchmarks are welcome. In particular, I suspect that this version might actually be slower on Pentium D's... Can anybody confirm or deny this?
|
BTC: 15MRTcUweNVJbhTyH5rq9aeSdyigFrskqE · LTC: LTCPooLqTK1SANSNeTR63GbGwabTKEkuS7
|
|
|
|
|
ZedZedNova
Sr. Member
Offline
Activity: 475
Merit: 265
Ooh La La, C'est Zoom!
|
|
January 12, 2012, 03:36:12 AM |
|
Thanks for the new version.
I just loaded it onto my Mac Mini Core i5 2.3 GHz, seems to be about the same as the version from 3-January. I'm seeing 13.5 kH/s - 14.10 kH/s.
- Zed
|
No mining at the moment.
|
|
|
|
racerguy
|
|
January 12, 2012, 03:52:00 AM |
|
went from 6.3khash/s to 7.05khash/s on my q9400.
|
|
|
|
|
ThiagoCMC
Legendary
Offline
Activity: 1204
Merit: 1000
฿itcoin: Currency of Resistance!
|
|
January 12, 2012, 07:17:54 AM |
|
New version!The changes only affect 64-bit systems; they should bring an up to 8% speed increase to Intel processors, and a 20% increase to CPUs of the AMD K10 family. Binary for 64-bit Windows (courtesy of diki): https://github.com/downloads/pooler/cpuminer/pooler-cpuminer-win64-20120111.zipA binary for 64-bit Mac OS X will be available soon. As always, all benchmarks are welcome. In particular, I suspect that this version might actually be slower on Pentium D's... Can anybody confirm or deny this? WOW! From 3.04 to 5.23 @ Phenom X6! All my machines are Ubuntu 11.04 64 bits. I have only two intels... Thanks a lot!! Best! Thiago |
|
|
|
|
ThiagoCMC
Legendary
Offline
Activity: 1204
Merit: 1000
฿itcoin: Currency of Resistance!
|
|
January 12, 2012, 07:23:00 AM |
|
@pooler! Don't you wanna do the following job: Request for an Scrypt capable fork of cgminer - maybe within new CGMiner / RPC: https://bitcointalk.org/index.php?topic=56040.0Thanks!! Thiago |
|
|
|
|
rTech
Sr. Member
Offline
Activity: 305
Merit: 250
Trust but confirm!
|
|
January 12, 2012, 11:13:12 AM |
|
Nice work again... mostly 1-3 kh/s boost for my intel crunhers  keep it up...! |
|
|
|
|
|
pooler (OP)
|
|
January 12, 2012, 02:13:58 PM |
|
Thank you for the offer, but I don't have much experience with cgminer... and I am working on too many projects at the moment. I think that someone (who is too modest to admit he's a good coder ) has already started working on that, however. |
BTC: 15MRTcUweNVJbhTyH5rq9aeSdyigFrskqE · LTC: LTCPooLqTK1SANSNeTR63GbGwabTKEkuS7
|
|
|
|
localhost
|
|
January 12, 2012, 10:12:17 PM |
|
WOW!
From 3.04 to 5.23 @ Phenom X6!
That's weird, either you skipped a version, either the Linux version improved at a different pace than the Windows builds. I already have around 5.0/core on a Phenom X4 with the build of January ~5. |
-
|
|
|
|
exahash
|
|
January 12, 2012, 10:59:58 PM |
|
As always, all benchmarks are welcome.
In particular, I suspect that this version might actually be slower on Pentium D's... Can anybody confirm or deny this?
My Pentium D's went up. On one machine the last codebase pulled Jan 4 was doing 4.28 and this one pulled Jan 12 is doing 5.44. Additionally, both threads are now showing the same rate (2.72 kh/s). Previously one said 2.05 and the other 2.23. This is a Pentium D 920 (2 @ 2.80GHz). My Opteron 8218's went down slightly from about 15.7 kh/s over 6 threads to about 15.3 kh/s. I don't think the 8218 is K10 arch tho. No benches for my Sempron and Athlon boxes because they are mining BTC and when I simultaneously mined LTC for a few days I got too many SICK and DEAD notices from cgminer. I think X wanted more CPU than it could get, even with cpuminer at nice -n19 My Xeon boxes (E5335 8@ 2.00GHz) show the ~8% bump from 30.72 to 33.33. My other Xeon's (X3430 4@ 2.40GHz) went from 20.06 to 23.16 (~15%) only they were running code pulled Jan 9. I gotta run so I'll bench the Core2Duo's later. Thanks again Pooler! |
|
|
|
|
ssvb
Newbie
Offline
Activity: 39
Merit: 0
|
|
January 13, 2012, 01:53:51 AM |
|
Thank you for the offer, but I don't have much experience with cgminer... and I am working on too many projects at the moment. I think that someone (who is too modest to admit he's a good coder ) has already started working on that, however. This kind of brings a question about who maintains the "official" litecoin miner De facto your miner is the most used/popular at the moment. Do you have time/motivation for maintaining it in the foreseeable future? |
|
|
|
|
|
exahash
|
|
January 13, 2012, 03:09:25 AM |
|
I gotta run so I'll bench the Core2Duo's later.
I just checked the Core2's (6320 @ 1.86GHz) and went from 6.95 kh/s to 7.9 kh/s (~13.6%). The earlier number is from the miner pulled Dec 29. |
|
|
|
|
|
pooler (OP)
|
|
January 13, 2012, 05:06:26 AM |
|
Thank you for the offer, but I don't have much experience with cgminer... and I am working on too many projects at the moment. I think that someone (who is too modest to admit he's a good coder ) has already started working on that, however. This kind of brings a question about who maintains the "official" litecoin miner De facto your miner is the most used/popular at the moment. Do you have time/motivation for maintaining it in the foreseeable future? Don't worry, I do. I'm just saying that I don't want to take on more projects than I can handle. |
BTC: 15MRTcUweNVJbhTyH5rq9aeSdyigFrskqE · LTC: LTCPooLqTK1SANSNeTR63GbGwabTKEkuS7
|
|
|
|
tiberiandusk
|
|
January 13, 2012, 05:28:46 AM |
|
Got another 2khash/s per core for me.
|
|
|
|
|
localhost
|
|
January 13, 2012, 07:20:54 AM
Last edit: January 16, 2012, 10:40:45 AM by localhost |
|
Indeed a nice improvement again. From 5 khs to 6.5khs per core on my P2X4 955...
Edit: and I just noticed: my CPU, although still at 100% and with the above-mentioned hashrate gain, runs much cooler: 53°C instead of the previous 58°C!
|
-
|
|
|
ripper234
Legendary
Offline
Activity: 1358
Merit: 1003
Ron Gross
|
|
January 16, 2012, 12:13:41 PM |
|
Can anyone help me with my question posted above? I'm trying to run the optimized miner on Windows 7 64 bit, and it just crashes immediately without any error messages.
I see this in the system (application) event log:
Faulting application name: minerd.exe, version: 0.0.0.0, time stamp: 0x4eef5c94
Faulting module name: pthreadGC2.dll, version: 2.8.0.0, time stamp: 0x4be6d174
Exception code: 0xc0000005
Fault offset: 0x00000000000036a3
Faulting process id: 0x149c
Faulting application start time: 0x01ccc39bb160b781
Faulting application path: C:\Program Files (x86)\Litecoin\new_miner\minerd.exe
Faulting module path: C:\Program Files (x86)\Litecoin\new_miner\pthreadGC2.dll
Report Id: ef0fea24-2f8e-11e1-9e5e-14dae96c0870
So far I have been successfully using the miner-i7.exe release by ArtForz. Any idea why this crashes? I have Soluto installed, so that might be the reason why I'm not seeing any error message ... although I did try to exit it and that didn't help.
|
|
|
|
|
