its crashing at -20? is what u mean? make sure u set that after it starts mining, it cant start with the offset that low it cant generate the dag as theres an initial spike in wattage at that point before stable hashing begins if -20 dont work start smaller u should get -18 at least with those stock core/volt settings cuz ur GPU is putting that 1074 at 881 that means ur asic is pretty decent better than most of mine almost all are over 900 stock
oh yea sorry im stoned and need to add when running offsets low after mining has begun u have to change them back to 0 before stopping mining too lol some how claymore miner spikes wattage when deleting the dag or something too
for reference heres my oldest 4GB MSI rig, bought this last summer sometime so gotta be 9-10 months old at least lil bit of hashing degradation from running for so long, fans used to only spin occasionally that one that dips from 26 down has been having that problem for a couple months now think its failing
for reference heres my oldest 4GB MSI rig, bought this last summer sometime so gotta be 9-10 months old at least lil bit of hashing degradation from running for so long, fans used to only spin occasionally that one that dips from 26 down has been having that problem for a couple months now think its failing
You are probably starving your GPU, to reduce power, but what is your hardware error volume. Run HWINFO and show us your memory hardware count after a reboot and one hour.
i dont bother with memory errors when im not overclocking the memory much because u dont get them that is only a problem for those trying to go above 27Mh on these GPU u dont see much for errors till u push past 1925 on mem that one GPU ran fine for over 6 months then started doing this one day after i plugged a monitor into it while it was mining these are the same settings ive used on all the 4GB card rigs since last summer each rig runs about 6amps either way that rig has only been running for about 35 hours so its still got another 130+ hours before ill shut down again unless win 10 decides it absolutely needs to update
i dont bother with memory errors when im not overclocking the memory much because u dont get them that is only a problem for those trying to go above 27Mh on these GPU u dont see much for errors till u push past 1925 on mem that one GPU ran fine for over 6 months then started doing this one day after i plugged a monitor into it while it was mining these are the same settings ive used on all the 4GB card rigs since last summer each rig runs about 6amps either way that rig has only been running for about 35 hours so its still got another 130+ hours before ill shut down again unless win 10 decides it absolutely needs to update
You can stop Windows 10 from restarting after an update (search Google) and set bios to restart if rig shuts down (depends on your MB. Mine is a Biostar TB85). Then, set miner to autostart on windows startup and... there you go. Rig works and if it shuts down, restart and start mining automatically. I have some problems with driver (16.9.2) which makes my rig restarts once a day, but no problem with that, since it keep running. My only problem is that, sometimes, after restart, don't load WattTool setting. I just have to keep checking if it is making 0,170 GH/s.
i dont bother with memory errors when im not overclocking the memory much because u dont get them that is only a problem for those trying to go above 27Mh on these GPU u dont see much for errors till u push past 1925 on mem that one GPU ran fine for over 6 months then started doing this one day after i plugged a monitor into it while it was mining these are the same settings ive used on all the 4GB card rigs since last summer each rig runs about 6amps either way that rig has only been running for about 35 hours so its still got another 130+ hours before ill shut down again unless win 10 decides it absolutely needs to update
Have you tried increasing the GPU speed to see if that helps ?
Hello drizzt_do_urden Recently, I'm getting too many memory errors.... I am using the same mod bios recommended here in the forum (memory timings from 1500 to 1625, 17xx, 19xx etc ... In Wattool I modified only 1150/930 in p3, p4, p5, p6 and p7, and 1950/930 in memory ... It flowed very well throughout this time (about 1 month) And now suddenly I get many memory errors and no frequency causes it to stop ... I'm using Claymore 9.4 Drivers 16.9.2
I've already tested Drivers 16.9.1 and claymore 9.3 but I get memory errors and sometimes claymore crashes ... I already tried to increase / decrease the memory clock and gpu clock and voltages but it does not work ...
I've been noticing that memory errors stop when I apply the GPU defaut clocks and keep the 1950/930 clock in memory only ...
I've tried all sorts of settings you can imagine ... and now I'm getting extremely strange results .... Now I'm getting 10mhs with any overclocking I do ...
Hello! Thanks! -Reset all settings...... OK -DDU driver....OK -Reinstall driver.....OK -reflash rom......OK
Now it seems to be more stable .... my ASIC Quality is 75.4%, I believe it is not very high and so I get memory errors by increasing a lot of memory clock even if I'm using 1000mv ... I'm afraid to increase the clocks because when it makes some mistake, the VGA becomes the same as before ... Last week, I took all day for it to work normally again ... any setting I put gave memory errors... The GPU Memory Controller sometimes gets 50%, sometimes 20% and sometimes 100% ... Is this normal or should it always remain 100%?
Hi all. Many thanks to drizzt_do_urden (and the rest of the contributors) for the specs. Running stable @ ~28,3 Mh/s (~71W/card - 840W total for rig) on 6 MSI RX 470 Gaming X 4Gb (Hynix), on a Biostar TB85, with 16.9.2 drivers (Windows 10 Pro). I have some "MSI RX 470 Gaming X 8GB" and "MSI RX470 ARMOUR OC 8GB" on the way. Any one has experience with these? On the beginning of the thread someone mentioned the 8Gb version, but no conclusion was achieved (I think) on stable specs...
Hi all. Just to let you know that I was able to optimize energy efficiency on my RX470 Gaming X 4GB. I basically copied (in WattTool) the P7 values to the P5 and P6 slots and my cards are still running stable @ ~28,3 MH/s but now they are using less than 70W of power. Varies from 65 to 68. Total for rig is now closer to 800W (almost 40W less).
It is better to set values in wattTool then to modifie in bios. If the card is freezing then automatically goes back to factory setting. but if you change values in bios then maybe the card won't start up anymore and you have to flash it again. WattTool profiles you can load just before the miner starts you just put it in the .bat file.
Thanks for the hint. Solved the problem by creating a .bat file with commands:
cd C:\"Path to WattTool"\ start watttool-0.92.exe "profile1.ini" "profile2.ini"... (without " - created these files by saving the profiles for each card)
Then, on Task Scheduler, put .bat file to load at startup, before the loading of the miner. All is well now: rig starts; loads watttool profiles; starts mining. When crashes, does all again.
I used the bios mod from the first post on 2 of the rx 470 gaming x 4g cards that I have but what I get is this: I have also tried myself to copy the 1500 timing value to all above that, but it's the same result.
I have restored all of my cards with the backup bios of it's original state so they all work but one of the cards just give me 4mh/s. Any thoughts?
hello all, i've buyed 3 new Msi 470 4gb and all of 3 are Samsung memory, i searched online for the best mod hashrate/watt but there are many versions, what is best mod performance for you? Thanks guys
Comments
make sure u set that after it starts mining, it cant start with the offset that low
it cant generate the dag as theres an initial spike in wattage at that point before stable hashing begins
if -20 dont work start smaller
u should get -18 at least with those stock core/volt settings
cuz ur GPU is putting that 1074 at 881 that means ur asic is pretty decent better than most of mine
almost all are over 900 stock
oh yea sorry im stoned and need to add
when running offsets low after mining has begun u have to change them back to 0 before stopping mining too lol
some how claymore miner spikes wattage when deleting the dag or something too
lil bit of hashing degradation from running for so long, fans used to only spin occasionally
that one that dips from 26 down has been having that problem for a couple months now
think its failing
because u dont get them
that is only a problem for those trying to go above 27Mh
on these GPU u dont see much for errors till u push past 1925 on mem
that one GPU ran fine for over 6 months then started doing this one day after i plugged a monitor into it while it was mining
these are the same settings ive used on all the 4GB card rigs since last summer
each rig runs about 6amps
either way that rig has only been running for about 35 hours so its still got another 130+ hours before ill shut down again
unless win 10 decides it absolutely needs to update
I have some problems with driver (16.9.2) which makes my rig restarts once a day, but no problem with that, since it keep running. My only problem is that, sometimes, after restart, don't load WattTool setting. I just have to keep checking if it is making 0,170 GH/s.
Recently, I'm getting too many memory errors....
I am using the same mod bios recommended here in the forum (memory timings from 1500 to 1625, 17xx, 19xx etc ...
In Wattool I modified only 1150/930 in p3, p4, p5, p6 and p7, and 1950/930 in memory ...
It flowed very well throughout this time (about 1 month)
And now suddenly I get many memory errors and no frequency causes it to stop ...
I'm using Claymore 9.4
Drivers 16.9.2
I've already tested Drivers 16.9.1 and claymore 9.3 but I get memory errors and sometimes claymore crashes ...
I already tried to increase / decrease the memory clock and gpu clock and voltages but it does not work ...
I've been noticing that memory errors stop when I apply the GPU defaut clocks and keep the 1950/930 clock in memory only ...
Could you advise me something?
Thanks man!
Now I'm getting 10mhs with any overclocking I do ...
-Reset all settings...... OK
-DDU driver....OK
-Reinstall driver.....OK
-reflash rom......OK
Now it seems to be more stable .... my ASIC Quality is 75.4%, I believe it is not very high and so I get memory errors by increasing a lot of memory clock even if I'm using 1000mv ...
I'm afraid to increase the clocks because when it makes some mistake, the VGA becomes the same as before ... Last week, I took all day for it to work normally again ... any setting I put gave memory errors...
The GPU Memory Controller sometimes gets 50%, sometimes 20% and sometimes 100% ... Is this normal or should it always remain 100%?
cd C:\"Path to WattTool"\
start watttool-0.92.exe "profile1.ini" "profile2.ini"... (without " - created these files by saving the profiles for each card)
Then, on Task Scheduler, put .bat file to load at startup, before the loading of the miner.
All is well now: rig starts; loads watttool profiles; starts mining. When crashes, does all again.
I have also tried myself to copy the 1500 timing value to all above that, but it's the same result.
I have restored all of my cards with the backup bios of it's original state so they all work but one of the cards just give me 4mh/s. Any thoughts?
then tell us the code
im surprised they mine
Thanks guys