nxcess
2[H]4U
- Joined
- Dec 22, 2004
- Messages
- 2,706
I use phoenix from guiminer and haven't had an issue with btcguild. If you need help, I can do a quick teamview session to troubleshoot it right now. PM me if you need it.
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
All of my rigs, first GPU starts with DEVICE=1
[Edit]
I take that back, on my Windows machines, DEVICE=0 is the first GPU. In Linux, it is 1.
210
I was running the beta before but it expired and wouldn't run even if I uninstalled and redownloaded/installed it. this one at least runs.
in other news the Phoenix miner on the other card inexplicably started picking up shares so I'm hashing 860 here. Bailing wire and duct tape but it works...lol.
Using the default OpenCL miner for GUIminer, I'm pulling down 378 Mhash/s on a Gigabyte 6950 2Gb card. Current speeds in CCC:
GPU: 930
Mem: 1250
Power: 20%
Fan: 55%
Temp: 73C
Seem peachy?
Following flags are active -
-v 2 -w128 -f2 BFI_INT AGGRESSION=6
I want to get this card nice and stable before I add a 5830.
Using the default OpenCL miner for GUIminer, I'm pulling down 378 Mhash/s on a Gigabyte 6950 2Gb card. Current speeds in CCC:
GPU: 930
Mem: 1250
Power: 20%
Fan: 55%
Temp: 73C
Seem peachy?
Following flags are active -
-v 2 -w128 -f2 BFI_INT AGGRESSION=6
I want to get this card nice and stable before I add a 5830.
I actually have no idea, I haven't had the urge to look into till just now.can you unlock the shaders in that one?
I don't know the exact voltage as CCC has the slider allowing max draw, but I was stable at 8% over. 20% is only a hard limit that the card is allowed to draw.What voltage did you need to get 930?
I tried the tweaks posted earlier - http://hardforum.com/showpost.php?p=1037460056&postcount=305, but I saw no appreciable difference. I haven't figured out how to configure one of the other kernels to surpass the stock OpenCL.As Lorien has pointed out earlier in the thread the modded Phoenix kernal will hash faster...but is very sensitive to other uses of the rig, even just a browser can knock 50Mh/s
I have not done the shader mod, but how do I lower my Mem clock that far? CCC restricts me to 1250 as a bottom limit.
I actually have no idea, I haven't had the urge to look into till just now.
I don't know the exact voltage as CCC has the slider allowing max draw, but I was stable at 8% over. 20% is only a hard limit that the card is allowed to draw.
I am envious of the clocks you guys are getting, 915 seems to be my limit and I need 1.18 v to do it
pretty happy with my hashrate though
Hey, what are folks out there paying per kWh?
Hey, what are folks out there paying per kWh?
1) Guiminer 20110701 released: https://github.com/downloads/Kiv/poc...r-20110701.exe
http://forum.bitcoin.org/index.php?t...1551#msg311551
This has the latest versions of all miners. It also includes fallback support when #n pool connection attempts fail.
2) POCLBM and Phatk kernel optimization by bitless - Status: Redundant.
This is the original 3% kernel tweak. It has now been rolled into the versions of phoenix and poclbm that come with the latest guiminer. There is no need to edit the kernel.cl files anymore.
3) Phatk kernel syntax error correction. Status: Redundant
This has already been rolled into newer versions of the phatk kernel.
4) POCLBM kernel optimization by goxed. Status: Obsolete
The __local tweak has been confirmed to cause a slow down in hash rate. Please, do not use it anymore. https://forum.bitcoin.org/index.php?topic=23309.msg303916#msg303916
DiabloD3 said:I am locking this topic.
As per OpenCL specification:
6.5.2 __local (or local)
The __local or local address space name is used to describe variables that need to be
allocated in local memory and are shared by all work-items of a work-group. This qualifier can
be used with arguments to functions (including __kernel functions) declared as pointers, or
with variables declared inside a __kernel function.
6.5.4 __private (or private)
All variables inside a function (including __kernel functions), or passed into the function as
arguments are in the __private or private address space. Variables declared as pointers
are considered to point to the __private address space if an address space qualifier is not
specified except for arguments declared to be of type image2d_t and image3d_t which
implicitly point to the __global address space.
In other words, the hardware threads are stomping each others work and producing nonsense.
5) Further improved phatk OpenCL Kernel (> 3% increase) for Phoenix by Diapolo. Status: Updated regularly
Download links and instructions here: http://forum.bitcoin.org/index.php?topic=25860.0;all
There are too any changes to list here, the gist of it is that it reduces the number of ALU OPs required to process a hash.
Kernel performance (BFI_INT active / APP KernelAnalyzer CAL 11.3 profile):
HD5870
original phatk: 1391 ALU OPs
2011-07-01: 1387 ALU OPs
2011-07-03: 1383 ALU OPs
2011-07-06: 1378 ALU OPs
2011-07-07: 1378 ALU OPs
2011-07-11: 1376 ALU OPs
2011-07-17: 1374 ALU OPs (17 less compared to original phatk)
HD6970
original phatk: 1706 ALU OPs
2011-07-01: 1709 ALU OPs
2011-07-03: 1705 ALU OPs
2011-07-06: 1701 ALU OPs
2011-07-07: 1701 ALU OPs
2011-07-11: 1700 ALU OPs
2011-07-17: 1698 ALU OPs (8 less compared to original phatk)
The fewer the Arithtmetic Logic Unit OPerations per second, the more hashes you can process per cycle. Newer versions incorporate the changes from all previous versions.
Getting 430-435 Mhash/sec on my 5870. 965 core, 300 memory. Looks good? Using the 07-01-2011 Phoenix with the kernel linked by Lorien above. BTCGuild
WAT!Damn Lorien
and here I am too drunk to implement...lol...try busting the latest not on Friday night
just kidding
BITCOIN SCAMMER
scammer purchased bitcoin and paid with paypal, after receiving the bitcoin, creates a dispute with paypal *Unauthorized Claim* and paypal blocks the transfer.