so, anyone had a chance to try out that new code posted by nvidia?
Meh I'll compile and try some Litecoin on my GTX780... 
Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 993. (Read 3426921 times)
January 22, 2014, 10:08:00 AM
January 22, 2014, 10:01:08 AM
so, anyone had a chance to try out that new code posted by nvidia?
January 22, 2014, 09:58:13 AM
Yep cracked it, DLL version crossovers..cleared them out and redownloaded the correct x86 2012 package then pulled that from the system 32 directory.
Thanks guys.
Now to find out how it performs on the following cards :-)
GTX 670
GTX 460
Quadro 4000
Quadro K4000
Thanks guys.
Now to find out how it performs on the following cards :-)
GTX 670
GTX 460
Quadro 4000
Quadro K4000
January 22, 2014, 09:52:20 AM
Yep tried chkdsk and reinstalling the C++ redistributables, And I'm using the DLL's from my working scrypt version of cudaminer.
I might try it on one of my other nvidia machines.
I might try it on one of my other nvidia machines.
Wrong dll's in the executables folder. Most likely libcurl architecture or version is wrong.
January 22, 2014, 09:51:25 AM
run chkdsk /f /r
then repair the Visual C++ Redistributable Packages.
January 22, 2014, 09:42:51 AM
run chkdsk /f /r
then repair the Visual C++ Redistributable Packages.
the error he got indicates a "bad image" (meant is an application or DLL binary image), which hints at disk corruption or a binary DLL that doesn't match his target hardware (maybe some 32 vs 64 bit confusion).
January 22, 2014, 09:41:19 AM
run chkdsk /f /r
then repair the Visual C++ Redistributable Packages.
January 22, 2014, 09:36:20 AM
A short overview of what kernels to use in the current github code. It gets kind of confusing now.
NOTE: for best speed on Linux with Fermi you want to comment out the two rule lines for fermi_kernel.cu (possibly also for test_kernel.cu) in Makefile.am such that the default rule for compute_10 takes over. Then run autogen.sh and configure again. It seems bizarre, but compiling the code for sm_20 (the actual target hardware) only makes it slower. This puzzling effect was NOT observed on Windows. Comment it out like this:
Christian
Code:
Fermi scrypt: F Kernel
Fermi scrypt-jane: X Kernel
Kepler compute 3.0 scrypt: Y Kernel <--- better than 2013-12-18 release IMHO
Kepler compute 3.0 scrypt-jane: K Kernel
Kepler Compute 3.5 scrypt: Z Kernel <--- most likely better than 2013-12-18 release
Kepler Compute 3.5 scrypt-jane: T Kernel
NOTE: for best speed on Linux with Fermi you want to comment out the two rule lines for fermi_kernel.cu (possibly also for test_kernel.cu) in Makefile.am such that the default rule for compute_10 takes over. Then run autogen.sh and configure again. It seems bizarre, but compiling the code for sm_20 (the actual target hardware) only makes it slower. This puzzling effect was NOT observed on Windows. Comment it out like this:
Code:
#fermi_kernel.o: fermi_kernel.cu
# $(NVCC) @CFLAGS@ -Xptxas "-abi=no -v" -arch=sm_20 --maxrregcount=63 $(JANSSON_INCLUDES) -o $@ -c $<
#test_kernel.o: test_kernel.cu
# $(NVCC) @CFLAGS@ -Xptxas "-abi=no -v" -arch=sm_20 --maxrregcount=63 $(JANSSON_INCLUDES) -o $@ -c $<
Christian
January 22, 2014, 09:35:12 AM
Yep tried chkdsk and reinstalling the C++ redistributables, And I'm using the DLL's from my working scrypt version of cudaminer.
I might try it on one of my other nvidia machines.
I might try it on one of my other nvidia machines.
January 22, 2014, 09:16:13 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
Check a page back, build 111 compiled by me.Thatnks patoberli I will take a look, If theres one thing I love about crypto its that its reinvigorated my urge to play with IT
Any clues on 0xc000007b errors with both versions once i've added in either msvcp100/110?
run chkdsk /f /r
then repair the Visual C++ Redistributable Packages.
January 22, 2014, 09:03:32 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
Check a page back, build 111 compiled by me.Thatnks patoberli I will take a look, If theres one thing I love about crypto its that its reinvigorated my urge to play with IT
Any clues on 0xc000007b errors with both versions once i've added in either msvcp100/110?
run chkdsk /f /r
then repair the Visual C++ Redistributable Packages.
January 22, 2014, 09:01:17 AM
You forgot an -L 2 there.
The -L is not yet rolled into the kernel launch configurations. This is intended, but not done yet.
Later on the launch config might look like this instead -K59x2/2. Then the only use for passing -L would be to tell autotune about the intended Lookup gap.
Gah, that was stupid
Thank you, working fine now, no more annoying autotune. January 22, 2014, 08:55:38 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
Check a page back, build 111 compiled by me.Thatnks patoberli I will take a look, If theres one thing I love about crypto its that its reinvigorated my urge to play with IT
Any clues on 0xc000007b errors with both versions once i've added in either msvcp100/110?
January 22, 2014, 08:35:24 AM
I just checked the Z config on the Vertcoin (script:2048) and it doesn't do well at all...
250khash against ~300khash (I think it is somewhat slower than the latest script code).
250khash against ~300khash (I think it is somewhat slower than the latest script code).
January 22, 2014, 08:31:26 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
Check a page back, build 111 compiled by me.Thatnks patoberli I will take a look, If theres one thing I love about crypto its that its reinvigorated my urge to play with IT
January 22, 2014, 08:23:50 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
Check a page back, build 111 compiled by me. January 22, 2014, 08:20:01 AM
Is there a compiled scrypt-jane version for Windows yet or do I need to work out how to compile code?
January 22, 2014, 08:16:22 AM
Lol, mine was also blocked. Maybe it's indeed because of that build, or maybe the admin changed something in his website code?!?
January 22, 2014, 08:13:02 AM
30 minutes after launching the new commit 111, my yacoin account was auto-locked because of too many attempts of connection.
It seems weird. Someone else have this problem ?
It seems weird. Someone else have this problem ?
January 22, 2014, 08:06:56 AM
With new v111 cudamine and the same setting i reached just the half
which could be the reason? Sorry, i'm a newbie
which could be the reason? Sorry, i'm a newbie
Don't do scrypt mining with the current code from github. You want to stick with the official 2013-12-18 release, unless you are really into adventure and experiments.
Jump to: