[Dev] NeoScrypt GPU Miner - Public Beta Test
-
hi guys
i just hooked up a amd 6950 and i downloaded the cgminer, but when i run it, it just stops working and crashes. I’m using windows 8 64bit. and just updated my drivers to the latest. Could you know why this may be happening?
-
If you are using catalyst 14.something it won’t work.
Don’t go higher than 13.12
But there may also be win8 issues as I know of atleast one other person running win 8 that is having issues
-
my 6950 is really slow… its like just 4kH
@Yusuf, you may want to try a higher -w value…
-
If you are using catalyst 14.something it won’t work.
Don’t go higher than 13.12
But there may also be win8 issues as I know of atleast one other person running win 8 that is having issues
aah tmuir12, thanks. i think i was using the 14.xx catalyst from above. could you please link me to the pre compiled cgminer for 13.12?
thanks again!
-
I have two PCs with Radeon HD6950s, shaders unlocked. They’re not cutting edge, but they are decent miners. Slightly overclocked, they get about 450 KHS scrypt mining each, or about 7.5 MHS each mining Goestl.
The PCs are running Windows 7 X64. One has 13.12 drivers, one has 14.4. That really doesn’t seem to matter, at least on my two PCs. Swapping bin files had no impact.
I have tried the Neoscrypt GPU miner, versions 3.7.5 and 3.7.7b. I get accepted shares, and I can solo mine, but the mining is at very low hash rates.
After reading through this whole thread more than twice in case I missed something, I have tried a variety of different settings. The best I can get with HD6950s is about 3 KHS, and it renders the PC unusable, even at relatively low -I settings (-I 8. I get similar performance (~2 KHS) per core CPU mining, and the PC is pretty usable while minerd is running. So at the moment, CPU mining is about 4 times more productive and uses about 1/3 of the power as GPU mining on my hardware. But CPU mining is never going to be productive when most other miners are using GPUs at 25-50 KHS.
This is going to alienate me from Neoscrypt if this is not addressed. Since this beta has been release, difficulty has gone up, and that’s only going to get worse when there is widespread adoption of the GPU miner. I’m not going to buy new GPUs just to mine PXC, or FTC for that matter.
I’m not whining, it’s just a simple decision based on the relative performance of my hardware. Some of the comments here have been “when are we going live?” If this doesn’t get fixed, I’m not going to be able to reasonably compete any more, and I’ll have to find something else to mine where my hardware is competitive (like Groestl), and so will anyone else mining with similar hardware.
I don’t particularly want to abandon PXC, I’d much rather stay. If I can help the developers to resolve the performance issue by testing or providing log files or whatever, I’d be happy to, just let me know what I can do.
-
use only 3.7.7b miner for now
your settings are way off to get good hashrate. -I 8 is history now.
try -g 2 -w 24 -I 14 , if failing take some steps down, i have not a card similiar to yours, but your setting should be much like mine
when mining scrypt cards like yours been reported using intensivity 20+
-
my 6950 is really slow… its like just 4kH
@Yusuf, you may want to try a higher -w value…
I think AMD HD6XXX may be supported. my 6770 and 6870 only 2.5K/hash. If I use lasted 3.7.7b, It is 4K/hash ,but I think they should 20K/s hashrate.
-
You can’t set the same -w value for HD6XXX as for HD7XXX and R2XX anything lower than 64/128 and it will fail to start.
-
My settings are way off? I’ve tried alot of settings. How do you know what settings I’ve tried? I only mentioned that it renders the PC unusable even at relatively low intensity settings like -I 8, while CPU mining 4 times as fast is only minimally noticeable, I can leave minerd running all day long, and use the computer for everything except gaming or using flash while mining…
I’m a newbie here. I’m not a newbie to mining. I have tried more than reasonable troubleshooting.
I’ve been trying this for about 3 weeks, since 3.7.5 was just released. I’ve tried numerous settings. I’ve tried both 13.12 drivers and 14.4 drivers. I’ve tried installing the sdk to see if that helped. I’ve tried making the bin on 13.12 drivers and using it with identical parameters using 14.4. I’ve tried alot of things.
I’m open minded, and I want this to work. so I just tried your settings, -g 2 -w 24 -I 14. My GPU won’t start at those settings. -w 24 is the problem. it causes:
Error -54: Enqueueing kernel onto command queue. (clEnqueueNDRandKernel)
GPU 0 failure, disabling!
GPU 0 failure, disabling!
Thread 1 failure, exiting
Thread 0 failure, exiting
If I make the settings -g 2 -w 32 -I 14, it starts succesfully and mines at 3.8 KHS, a new record for me, but still less effective than CPU mining, and it renders the PC unusable. With these settings, driver does appear to make a difference. The 14.4 driver only gets 2.4 KHS and is noticeably more laggy, like borderline unstable even.
-g 2 -w 64 -I 14 yields 4.188 KHS. Better. Finally have performance equal to a core 2 duo CPU mining. Probably using twice the power to do it. Laggy as hell, even on the 13.12 drivers. cgminer became unresponsive after a while. I think it was still mining, but cgminer stopped updating the screen, and I couldn’t even quit, had to hit the red x to get it to stop. Still nowhere near a 6 core amd CPU for productivity or efficiency.
-g 2 -w 64 -I 12 yields 3.23 KHS, seems stable, but pointless.
-g 2 -w 128 -I 12 yields 3.0 KHS, not helpful.
-g 2 -w 128 -I 14 yields 4.0 KHS, then the graphics driver (13.12) blew up.
I’ll try anything you gentlemen suggest, but if I can’t get at least 15 KHS when newer cards are hitting close to 50, this is pointless. There is something wrong with the with how this miner/kernel interact with older radeons. Hopefully, it can be fixed. if not, I understand, and I’ll move on.
Edit - I 14 appears to produce less accepted shares. Hashrate measured at the pool is higher with -I 12.
-
aah tmuir12, thanks. i think i was using the 14.xx catalyst from above. could you please link me to the pre compiled cgminer for 13.12?
thanks again!
You download it from the AMD website, you need to hunt a bit on the page, but you can download previous versions from their website, they are big downloads though
http://support.amd.com/en-us/download
after selecting your graphics card details and operating system it will bring up the latest driver, look to the right of the page under ‘helpful links’ for ‘previous drivers and software’
-
Hi Henle, I think you your intensity may be set too high, anything higher than I 11 on my HD7800 gives poor results, I have found for Neoscrypt you need to set your intensity lower than Scrypt.
I get around 30Khash for my HD7800
-
Going as low as -I 8 reduces the performance down to 1.0 KHS. cgminer won’t even start lower than that. That’s about 1/10th the performance of CPU mining, and even at that low of an intensity setting, it still renders the PC virtually unusable.
Here’s the thing - I’ve got my core clock at 900, it really doesn’t like going much higher, even with a little extra vCore. The GPU never gets hotter than 74C, The fan doesn’t even spin up at all, even at -I 14, so maybe it’s not using much power. But at the same time, it renders Windows nearly unusable. Clearly, it’s using all available resources, but it’s not using them effectively.
I haven’t seen anyone post good results using HD5000/HD6000 hardware. I have seen I guess 3 or 4 people say they’re getting similar results to what I’m getting. If I had to guess, the code just will not perform on VLIW hardware.
I don’t think I’m going to buy new GPUs just to continue mining PXC. I’d have to get 50 blocks a day for 6 months to pay for them.
-
Yes, the cgminer 3.7.7b with 6950 is not very stable… able to run it for more than 24hrs at -I 14 -w 128 -g 2 but has difficulty shutting it down… have to kill it with -9 getting about 4.5KHash/s, there should be a lot of rooms for improvements.
-
ive been running stable for 24h now
-
just noticed that windows miner shared here cannot go below -w 48 when -I 14, only giving hw errors then
miner that i compiled myself goes to -w 24 at -I 14 without hw errors, just now cant remember settings i used to compile
values between those will not work but 24 or 48 will
link to miner without anydlls:(if u want to test and play)
http://www.filedropper.com/cgminer-w24
(daimyo, this is not same version u have, size 597kb?)
-
cisa, i downloaded your compiled version -w 48 -I with the libraries
-
cisa, i downloaded your compiled version -w 48 -I with the libraries
this is new one, compiled with different settings
earlier was not stable for me, stuck after 20hours
now trying this one if its keeps up hashing for days
here it stucks:
waiting for work to be available from pools
so it could be pool problem, not miner?
edit:
tested miner to have best results with 14.xx drivers/13.xx bin
-
same settings 3.7.7b, -g 2 -w 48 -I 14, 13.xx drivers
280x hynix= 100khs
have u tried bin files crossed? freshalgo did best when 14.xx created bin at 13.x drivers
Awesome thanks!
Working for me too. Got 100kh/s on each of my R9 280x. 14.xx driver.
-
cgminer 3.7.7b in http://www.ftc-c.com/pack/testnet.rar , It had fixed some problem.
-
Tried lizhi’s new 3.7.7b with Radeon HD 6950
-g 2 -w 64 -I 14 = 4.45 KHS. WU = 4.1 2.04 KHS at the pool. Extremely unstable. Screen flickering… Had a hard time killing it.
-g 2 -w 64 -I 12 = 3.2 KHS. WU = 3.7. 2.99 KHS at the pool… Laggy, but stable (normal for mining).
-g 1 -w 64 -I 12 = 3.2 KHS. WU = 3.7. 3.1 KHS at the pool. Mining
-g 1 -w 48 -I 12 = GPU failure.
-g 1 -w 32 -I 12 = 3.92 KHS. WU = 4.3. 3.3 KHS at the pool. Mining.
-g 2 -w 32 -I 12 = 4.0 KHS. WU = 1.8. 2.6 KHS at the pool. Mining.
-g 1 -w 32 -I 10 = 1.9 KHS. WU - 1.9. 1.75KHS at the pool. Almost usable as a PC while mining.
-g 1 -w 24 -I 12 = GPU failure.
-g 1 -w 16 -I 12 = 2.8 KHS. WU = 2.5. 3.0 KHS at the pool. Mining.
Slightly better numbers, but not significantly. Stability seems a little better.
