With the H == 0 mod, I get about a 2.5 Mhash/s gain on my non overclocked HD 5770.
From about 166 Mhash/s to about 168.5 Mhash/s using -f 7.
If I lower the -f switch to less than 7, I can get a couple more Mhash/s but the cpu usage will rise and increase the power consumption.
Intel Core i7 860 with 8gb ram
Windows 7 64bit
HD 5770 (not overclocked)
Catalyst Version: 11.2
Stream Version: Not sure. Whatever came with the driver. Probably 2.4
Topic: python OpenCL bitcoin miner - page 17. (Read 1239060 times)
That's all you have to do. Open it in notepad/wordpad, change the lines in BitcoinMiner.cl, and see if your speed goes up or down. At this point nobody has been able to pinpoint exactly why it is slower for some and faster for others.
I think I do
I think its got to do with what SDk you are using I think this is faster using v2.4 and slower with v2.2. My theory anyway 
2 Mhash/s gain.
HD 6870 core 1038, mem 345 272Mhash/s
Now 274 Mhash/s
win7 32, cat 11.4 & 2.4 beta
HD 6870 core 1038, mem 345 272Mhash/s
Now 274 Mhash/s
win7 32, cat 11.4 & 2.4 beta
4 Mhash/s GAIN
Win7 64
5870s
cat 11.4
Win7 64
5870s
cat 11.4
same here but 5Mh, 2* 6870s and sdk 2.4 cat 11.4
4 Mhash/s GAIN
Win7 64
5870s
cat 11.4
Win7 64
5870s
cat 11.4
That's all you have to do. Open it in notepad/wordpad, change the lines in BitcoinMiner.cl, and see if your speed goes up or down. At this point nobody has been able to pinpoint exactly why it is slower for some and faster for others.
poclbm_py2exe_20110325.7z
If i just edit the bitcoinminer.cl in the above zip file, from
this
to this
& RUN pocblm.exe with options to mine in a pool, will i may get more hash?
Or i have to compile something?
Using windows & have zero programming knowledge.
If i just edit the bitcoinminer.cl in the above zip file, from
this
Code:
#ifdef VECTORS
if (belowOrEquals(H.x, targetH, G.x, targetG))
{
output[OUTPUT_SIZE] = output[nonce.x & OUTPUT_MASK] = nonce.x;
}
else if (belowOrEquals(H.y, targetH, G.y, targetG))
{
output[OUTPUT_SIZE] = output[nonce.y & OUTPUT_MASK] = nonce.y;
}
#else
if (belowOrEquals(H, targetH, G, targetG))
{
output[OUTPUT_SIZE] = output[nonce & OUTPUT_MASK] = nonce;
}
#endif
}
to this
Code:
#ifdef VECTORS
if (H.x == 0)
{
output[OUTPUT_SIZE] = output[nonce.x & OUTPUT_MASK] = nonce.x;
}
else if (H.y == 0)
{
output[OUTPUT_SIZE] = output[nonce.y & OUTPUT_MASK] = nonce.y;
}
#else
if (H == 0)
{
output[OUTPUT_SIZE] = output[nonce & OUTPUT_MASK] = nonce;
}
#endif
}
& RUN pocblm.exe with options to mine in a pool, will i may get more hash?
Or i have to compile something?
Using windows & have zero programming knowledge.
My results (+7 mHash/sec per GPU) were from:
Ubuntu 10.10 (x86_64)
ATIStream SDK 2.1
fglrx driver 11.3
HD 5870s
Ubuntu 10.10 (x86_64)
ATIStream SDK 2.1
fglrx driver 11.3
HD 5870s
What the heck ... it must be the SDK or fglrx driver version.
ubuntu 10.10 (x86_64)
ATIstreamSDK 2.1
fglrx driver 11.2
HD 5970
H == 0 mod
SLOWER 1-2% (4-5 Mhas/s)
ubuntu 10.10 (x86_64)
ATIstreamSDK 2.1
fglrx driver 11.2
HD 5970
H == 0 mod
SLOWER 1-2% (4-5 Mhas/s)
Okay, I tested the H == 0 mod on linux ubuntblow 10.10, HD5970 and it is 1-2% SLOWER. (Wonder why it is faster on Win?)
... poclbm fork for Win users ... ?
I have a Ubuntu 10.10 box and it gets about 1.1Mh/s faster with this mod.
Ubuntu 10.10, 5870, and this mod consistently knocks me down 5MH/s.
I made the modification and my 5870s went from 363 mHash/sec and 361 mHash/sec to 370 even on both. Ubuntu 10.10.
Okay, I tested the H == 0 mod on linux ubuntblow 10.10, HD5970 and it is 1-2% SLOWER. (Wonder why it is faster on Win?)
... poclbm fork for Win users ... ?
I have a Ubuntu 10.10 box and it gets about 1.1Mh/s faster with this mod.
Okay, I tested the H == 0 mod on linux ubuntblow 10.10, HD5970 and it is 1-2% SLOWER. (Wonder why it is faster on Win?)
... poclbm fork for Win users ... ?
It's not just windows. Its running around 5000kh/s faster on my debian server.
yup I am on win 7 64 bit
Okay, I tested the H == 0 mod on linux ubuntblow 10.10, HD5970 and it is 1-2% SLOWER. (Wonder why it is faster on Win?)
... poclbm fork for Win users ... ?
The if (H == 0) modification fixed my crossfire issue. Now all the GPUs are used at 97%-99%, before i had 1 GPU at 99% one at around 94% and one at around 85%.
Wow, really? How could that be I wonder?
Did you do anything else besides pasting in the above section of code?
I just made all the 3 changes posted, that's it. I'm using 3 hd 6950 on windows 7 64 bit.
Besides pasting in the
#ifdef VECTORS
if (H.x == 0)
-------
loop, what other 2 changes did you make exactly?
H.y == 0
H == 0
I believe he didn't paste but just manually change the 3 things (H.x == 0 and the 2 above)
The if (H == 0) modification fixed my crossfire issue. Now all the GPUs are used at 97%-99%, before i had 1 GPU at 99% one at around 94% and one at around 85%.
Wow, really? How could that be I wonder?
Did you do anything else besides pasting in the above section of code?
I just made all the 3 changes posted, that's it. I'm using 3 hd 6950 on windows 7 64 bit.
Besides pasting in the
#ifdef VECTORS
if (H.x == 0)
-------
loop, what other 2 changes did you make exactly?
Just curious did anyone manage to mine a successful block with the (H == 0) mod?
deepbit's pool is seeing the right hashrate so I am being paid in that way :p
Just curious did anyone manage to mine a successful block with the (H == 0) mod?
I just installed the 270.61 nvidia drivers, and it seems to make the drivers crash each time I exit poclbm by clicking the x, if I ctrl-c they seem to exit fine most of the time. Nothing serious yet, but I keep getting these in my system event log:
"Display driver nvlddmkm stopped responding and has successfully recovered."
"Display driver nvlddmkm stopped responding and has successfully recovered."
have you done any overclocking?
No
Completely remove driver using driver sweeper & install again with CUDA.
Jump to: