this program causes severe memory leak errors on a couple machines with 7950s and 7970s... not sure what the issue is
There are not memory leaks, just use lower "-hc" value, read "readme" for details.
|
|
|
And for those that are interested this works with ATI 7990 graphics cards ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) I am running with i7-990x cpu and plenty of RAM . 2 x ATI 7990s (MSI brand) at default clock settings on 13.12 drivers. No Crossfire. I tweaked a little and I've found this worked best for me Bat files..: NsXpmMiner.exe -o ypool.net -u user.worker -p ** -di 0 -t 8 -hc 256 NsXpmMiner.exe -o ypool.net -u user.worker -p ** -di 1 -t 8 -hc 256 NsXpmMiner.exe -o ypool.net -u user.worker -p ** -di 2 -t 8 -hc 256 NsXpmMiner.exe -o ypool.net -u user.worker -p ** -di 3 -t 8 -hc 256 Have you tried "-hc 320"? Must be a bit faster. Also may be "-hc 512" for 2-3 cards and "-hc 256" or "-hc 320" for the rest.
|
|
|
Probably my miner is good enough if so many people use it and don't mind paying dev fee. And for that dev fee I will do my best to make it the fastest. If you don't like it, it's a good reason to release miner that is faster than mine. Money => competition => more miners and faster miners, it's good for everyone. If you don't agree with dev fee, use "-nofee" mode or use different miner (4 different miners are already available for now). Announcement #1: in 3-5 days I'll release new version, +20...25% speed on all cards. So it will cover dev fee again and with "-nofee" option it will work even faster than current version without this option ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Prototype shows about 200K 4ch/h and 140 7ch/h on 290X, CPD is about 2.3, but still a lot of work required to make it stable and make it work on slower cards properly. Announcement #2: I will release solo miner. But firstly I will prepare and release another version that can work properly when we reach Diff 11, it seems we will hit it very soon and current version cannot handle it properly with good mining speed. So be patient please ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
I see discussion in this topic is about dev fee now, not about the miner itself ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) I'm going to change dev fee address in new version to collect some statistics about popularity of new version, so please don't miss it and check for my new address in time ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
Does this miner support 7850 1G or 2G??
It should work if you specify "-hc 256" or may be some less value.
|
|
|
When waiting for a new version of the miner ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) In a few days, I'm working on further speedup.
|
|
|
i have two rigs with same hardware and 5x280x
one has 2gb ram and the other 4gb ram
in the 2gb ram i can set hc512 and it works.
in the 4gb i can't even when its in the bat file, it works automaticly with 248 hashcnt. when i set gpu alloc100 the miner doesn't even start trt400+-
the performance diffrence is so big between the two? will there be a fix?
Try to enable alloc100 and set "-hc 320".
|
|
|
Hi Claymore, I have many problem with many my grids: x 5 r9-270, x 6 r9-270, x 2 r9 - 270 When I run, I received 1 or 2 cards with stop driver VGA. I use version 2.1, try with many config such as: ss - 2, HC: 512, 256 and default core, mem but still error. I can't fix it. My command: NsXpmMiner.exe -o ypool.net -u abcd.1 -p 1 -di 0 -ss 2 -sd 0 -hc 512 di 0-> 5 Ex: Card 0 stop driver, card 1 stop driver and no have TRT, CHD.
But no this error with other grid:5 x r9 280x or 5 x HD 7950. Please see & resolve it for me. Thanks
Try NsXpmMiner.exe -o ypool.net -u abcd.1 -p 1 -di 0 -ss 2 -hc 256 Also try to remove "GPU_MAX_ALLOC_PERCENT" environment variable if you use it. Also disable overclocking, install(reinstall) catalyst v13.12 and reboot computer. Helps in 99% cases. Hi Claymore, I remove GPU_MAX_ALLOC_PERCENT 100, restart pc and it's ok. I run 6 card 270 no problem. But low speed. GPU_MAX_ALLOC_PERCENT: TRT 680 Remove GPU_MAX_ALLOC_PERCENT: TRT 730 Can you fix this error and upgrade version 2.2 so that I still can use GPU_MAX_ALLOC_PERCENT 100 ? Thanks TRT 680 is ok for 270 cards. You can use GPU_MAX_ALLOC_PERCENT but use "-hc 256" or even less value (e.g. 200) that will work for you.
|
|
|
Hi Claymore, I have many problem with many my grids: x 5 r9-270, x 6 r9-270, x 2 r9 - 270 When I run, I received 1 or 2 cards with stop driver VGA. I use version 2.1, try with many config such as: ss - 2, HC: 512, 256 and default core, mem but still error. I can't fix it. My command: NsXpmMiner.exe -o ypool.net -u abcd.1 -p 1 -di 0 -ss 2 -sd 0 -hc 512 di 0-> 5 Ex: Card 0 stop driver, card 1 stop driver and no have TRT, CHD.
But no this error with other grid:5 x r9 280x or 5 x HD 7950. Please see & resolve it for me. Thanks
Try NsXpmMiner.exe -o ypool.net -u abcd.1 -p 1 -di 0 -ss 2 -hc 256 Also try to remove "GPU_MAX_ALLOC_PERCENT" environment variable if you use it. Also disable overclocking, install(reinstall) catalyst v13.12 and reboot computer. Helps in 99% cases.
|
|
|
How can i see program start time or running time ?
I missed it in v2.1, added in v2.2, it will be available in a few days. Any efficiency enhancement? Yes, it will be faster.
|
|
|
How can i see program start time or running time ?
I missed it in v2.1, added in v2.2, it will be available in a few days.
|
|
|
@Claymore
Total Round Time (TRT): 11795
If you get TRT = 0 or such big value, set "-hc 256" (see readme). Also make sure that Catalyst 13.12 is used, disabled overclocking and reboot computer.
|
|
|
When you mention we need to mine 24hours before speed is picking up, is that the flow of coins in pool, or the acual mining speed your talking about?
It is pool-related and also to reduce "luck factor". Mining speed is the same all the time.
|
|
|
With a 8 core processor, should -t 8 or higher number be used? Again 2 gpu rig Thnx for your other tips with set max, it seems to work on -hc 256
The best approach is to test how it works on your hardware ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) If CPU is slow and GPUs are fast, "-t 8" may help. Otherwise it is useless.
|
|
|
Was curious if anyone had much experience with R9 290's on subpar processors?
I've got a 2.8 GHz G1830 Haswell Celeron dual core.
I have 4x R9 290s.
After 24 hours, I'm getting about 1.4 cpd per card. 4ch is 116k, and 7 ch is 83. They are overclocked, 1400/985, as that seemed to improve the 7ch from 75 to 83.
Based on previous reports, this seems a little low? I'm running it -ss 2 for each card. CPU usage fluctuates between 50-67%.
Is this the norm, or should I be getting better numbers? What can I change to improve the cpd?
You should get about 100 7ch/h. What if start only one card, what 7ch/h does it show? What if specify "-t 8"?
|
|
|
Everyone that is giving stats like the 280x at 1.? cpd. Are these numbers from after about 24 hours of mining? Or should I see that within the first 30 mins of mining?
You will see estimated CPD immediately, however, you should wait about 30 minutes to get correct value. To get correct average CPD you need several days of mining.
|
|
|
@Claymore:
I'm able to run my 4x280X rig without problems when i set 3 cards to hc=512 and last one with hc=256. All cards get 110-115 4ch.
What about CPU load?
|
|
|
I've noticed that setting gpu_max_allocae blabla variable has close implication on HashCount. So make: set gpu_ blabla=100 in .bat file and next deal with -hc. I also noticed that -hc 640 in command line made it hc=512 in miner info. So HC greater than 512 is cut?
If GPU does not allow to allocate enough memory, HashCnt will be decreased automatically, so even if 512 is specified you can see less value, set "GPU_MAX_ALLOC_PERCENT 100" to avoid it. HC cannot be higher than 512, too much memory is required for higher values and OpenCL fails. When i do set set "GPU_MAX_ALLOC_PERCENT 100", it seems not mining at all. Trt 0 and showing nothing in stats. Could you plz post an exact bat config so i can see where i am putting it wrong? Do i use the " before and after comand? Not having bad results, but i see that it could be better compared to others. 8core cpu on most rigs, 2 gpus each. Try "NsXpmMiner.exe -o ypool.net -u YourLogin.WorkerName -p YourPassword -di 0 -hc 256 -ss 2"
|
|
|
What settings for 7870 people? -hc 256? -hc 384?
Always use default values. But if miner is unstable - try "-hc 256". It depends not only on card, but also how many GPUs you have installed. So just use default settings and change them only if you really need it.
|
|
|
Claymore
whether it is possible to make the gpu-engine parameters and gpu-memclock in programm? because by default the program exposes 925 and 1300 and my card holds by default 1300 and 1550
sanks
Use "Claymore GPUs Control Utility v1.0" for overclocking and temperature control for all cards.
|
|
|
|