Pages:
Author

Topic: Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 101. (Read 214410 times)

sr. member
Activity: 1484
Merit: 253
I tried to increase GPU voltage, it's helps on neoscrypt for sometime. Thread hangs not so quick, but on cryptonight it didn't helps at all. I use lowered clocks for my 580 cards. I use 1245MHz GPU clock. Lowering it decreases speed - and this is not good for me.
Increasing the voltage works well but it's not so good assuming ^2 power usage, so at home I can only lower MHz at GPU and mem. Ye, decreasing speed this way.
On cryptonight increasing voltage didn't work. I tried up to 1000mV with the same 1245Mhz vcore. On 17.11.4 card stable works with 1245MHz and 950mV. I stop it, becase I don't want to lose speed or inrease power consumption and temp.
I'll wait while zawawa take full control on 17.12.x drivers, or AMD release new version. If it didn't helps in future, I'll return to 17.11.4 driver.
newbie
Activity: 11
Merit: 0
I tried to increase GPU voltage, it's helps on neoscrypt for sometime. Thread hangs not so quick, but on cryptonight it didn't helps at all. I use lowered clocks for my 580 cards. I use 1245MHz GPU clock. Lowering it decreases speed - and this is not good for me.
Increasing the voltage works well but it's not so good assuming ^2 power usage, so at home I can only lower MHz at GPU and mem. Ye, decreasing speed this way.
newbie
Activity: 11
Merit: 0
Pls, Zawawa think 'bout the interface: you really need not to have Start/Stop button at every page.
Just for RX470/480 we may need this button at the first page, here:
https://s5.postimg.org/rijfmusd3/screenshot_3207.png
Bcause we never see the bottom of the miner, really, without any monitor at these rigs. Right now I have to click the first page, then press "Shift-Tab" and "Enter" to Start or Stop mining.
sr. member
Activity: 1484
Merit: 253
And strange bug with these 17.2.2 drivers. On all versions GGS from 1.1.8 to 1.1.12 after start mining and stopping after, 1st GPU (monitor connected) continue to show 100% activity. Core clock is maximum, memory clock at 300. Temperature is low, but a bit higher than at 300/300, fan drop to low or 0%. All functions work. Miners starting on it and works at 97-99% speed on this GPU. Computer hangs during rebooting. And not rebooting until power off and on.
Had the same bug, I found this was caused by not very heavy overclocking. Try pls lower your GPU Clock by 10+ and do so with your Mem clk as well. You can seek your sweet point, try and seek. Tryed at Neoscrypt only, the others algo nowadays aren't worth of talking with AMD cards.
As was said before by Zawawa, U will need to check and find, bcause THIS miner is completely different from others. This will be the answer to your coldboots and 100% loading of one of your cards when mining tread hungs.
I tried to increase GPU voltage, it's helps on neoscrypt for sometime. Thread hangs not so quick, but on cryptonight it didn't helps at all. I use lowered clocks for my 580 cards. I use 1245MHz GPU clock. Lowering it decreases speed - and this is not good for me.
I think it's drivers bugs, on 17.11.4 all was right.
newbie
Activity: 11
Merit: 0
And strange bug with these 17.2.2 drivers. On all versions GGS from 1.1.8 to 1.1.12 after start mining and stopping after, 1st GPU (monitor connected) continue to show 100% activity. Core clock is maximum, memory clock at 300. Temperature is low, but a bit higher than at 300/300, fan drop to low or 0%. All functions work. Miners starting on it and works at 97-99% speed on this GPU. Computer hangs during rebooting. And not rebooting until power off and on.
Had the same bug, I found this was caused by not very heavy overclocking. Try pls lower your GPU Clock by 10+ and do so with your Mem clk as well. You can seek your sweet point, try and seek. Tryed at Neoscrypt only, the others algo nowadays aren't worth of talking with AMD cards.
As was said before by Zawawa, U will need to check and find, bcause THIS miner is completely different from others. This will be the answer to your coldboots and 100% loading of one of your cards when mining tread hungs.
newbie
Activity: 11
Merit: 0
Thank U, zawawa, 1.1.12 and 1.1.13 R solid as rock!

https://s5.postimg.org/3p06hxyt3/screenshot_3203.png

I was wrong in my previous message, "dropping miner stats from pool" was mainly caused by DDoSing one of the pools by Nicehash. This was not your bug, but other miners that post their "Details" here https://s5.postimg.org/mhc1lnso7/screenshot_3204.png weren't however disconnected from stats for this time. Think of leaving the Details of YOUR miner somewhere via Stratum later, it's worth!
newbie
Activity: 6
Merit: 0
Dear zawawa, I have some test result for u..

You can see how different is the total shares in 10 minutes..(everything is same, intensity, algo, pool, clocks, settings etc..)..

https://preview.ibb.co/frJ1Pw/GGS_compare.jpg

Sorry for watermark , I will  say , "i was there" when your program reaches millions Smiley  

hello
i see you have rx580 and you can get very high hash in cryptonight
i have rx570 but i get only 650 hash
how do you get that?
tyvm
sorry for my english
valerio

I don't know how he got this but i did bios mods to get this.

https://imgur.com/a/yy3HL

i read that wrong i got my rx570 up to 725 but it was unstable but i modded both my rx 580 and rx 570's with bios from
https://anorak.tech/c/downloads

you do it at your own risk though.

stock my rx580's were 725ish rx570 is 640ish modded
850 to 900 rx580
750 rx570
sr. member
Activity: 728
Merit: 304
Miner Developer
Great work happy to have the devfee. I have tried 1.1.6,1.1.9,1.1.13

1.1.6 is probably the most stable.
1.1.9 is really close
1.1.13 is werid alot of bad submits

all using neoscrypt.

beginning with 1.1.9 and fan control if you don't manually adjust the temps to like 65/75 and fan 45/100 even if you uncheck the fan control box the gpus overheat and crash. its like gateless takes over the fans but fails to regulate on the default settings. I adjusted it so its fine on 1.1.9 but 1.1.13 bad shares start to come out.

Using the new Radeon drivers on all 3.

Just a FYI

Thanks a lot for the feedback. The new driver is so moody that it is beyond belief after all these years of AMD driver woes.
So sorry I am dragging you guys along with this mess. I *think* I fixed the known problems, but I am just making sure that is the case right now.
member
Activity: 242
Merit: 10
Dear zawawa, I have some test result for u..

You can see how different is the total shares in 10 minutes..(everything is same, intensity, algo, pool, clocks, settings etc..)..



Sorry for watermark , I will  say , "i was there" when your program reaches millions Smiley   

hello
i see you have rx580 and you can get very high hash in cryptonight
i have rx570 but i get only 650 hash
how do you get that?
tyvm
sorry for my english
valerio
member
Activity: 242
Merit: 10
when oc will be available?
newbie
Activity: 6
Merit: 0
Great work happy to have the devfee. I have tried 1.1.6,1.1.9,1.1.13

1.1.6 is probably the most stable.
1.1.9 is really close
1.1.13 is werid alot of bad submits

all using neoscrypt.

beginning with 1.1.9 and fan control if you don't manually adjust the temps to like 65/75 and fan 45/100 even if you uncheck the fan control box the gpus overheat and crash. its like gateless takes over the fans but fails to regulate on the default settings. I adjusted it so its fine on 1.1.9 but 1.1.13 bad shares start to come out.

Using the new Radeon drivers on all 3.

Just a FYI
newbie
Activity: 17
Merit: 0
Can you upgrade the source to VS2017 compiler?  I think your on 2013 and I only have 2015 and 17 :X
full member
Activity: 420
Merit: 184
kinda surprised at this moaning about big page files...

Normally I'd agree with you, but I just went to try GGS on my desktop with a single RX 570 (4GB) and 16GB of ram and GGS warned me I needed to set my pagefile to 24GB... Say what?



This requirement may not make much sense now, but it should when automatic algorithm switching is implemented.
Imagine each card working on its most profitable algorithm independently.

Ah, okay... that's rather ambitious of you! Thanks for the explanation.
sr. member
Activity: 728
Merit: 304
Miner Developer
kinda surprised at this moaning about big page files...

Normally I'd agree with you, but I just went to try GGS on my desktop with a single RX 570 (4GB) and 16GB of ram and GGS warned me I needed to set my pagefile to 24GB... Say what?



This requirement may not make much sense now, but it should when automatic algorithm switching is implemented.
Imagine each card working on its most profitable algorithm independently.
sr. member
Activity: 728
Merit: 304
Miner Developer
I think the error is not because of the custom pool. I tried what you told me but i get the same error like before when i try to press start.
I also tried to install from the setup but if i do not select the neoscrypt algo in the custom pool it starts and i get no error. If i select the neoscrypt then the program crashes with the error already submitted above Sad

Which video card are you using? Did you install the recommended driver?

What are the recomended drivers if i may ask. I had 15.12.

The current recommended driver is Adrenaline 17.12.2.

http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Adrenalin-Edition-17.12.2-Release-Notes.aspx
newbie
Activity: 15
Merit: 0
I think the error is not because of the custom pool. I tried what you told me but i get the same error like before when i try to press start.
I also tried to install from the setup but if i do not select the neoscrypt algo in the custom pool it starts and i get no error. If i select the neoscrypt then the program crashes with the error already submitted above Sad

Which video card are you using? Did you install the recommended driver?

What are the recomended drivers if i may ask. I had 15.12.
full member
Activity: 420
Merit: 184
kinda surprised at this moaning about big page files...

Normally I'd agree with you, but I just went to try GGS on my desktop with a single RX 570 (4GB) and 16GB of ram and GGS warned me I needed to set my pagefile to 24GB... Say what?

sr. member
Activity: 728
Merit: 304
Miner Developer
Dear zawawa, I have some test result for u..

You can see how different is the total shares in 10 minutes..(everything is same, intensity, algo, pool, clocks, settings etc..)..



Sorry for watermark , I will  say , "i was there" when your program reaches millions Smiley  

Hmm... Let me see...

Edit: I checked my rigs, but I didn't see any discrepancies between the locally reported hashrate and the hashrate reported by the pool. Strange, strange... Could you check the hashrate reported by the pool with the new version?
newbie
Activity: 126
Merit: 0
Dear zawawa, I have some test result for u..

You can see how different is the total shares in 10 minutes..(everything is same, intensity, algo, pool, clocks, settings etc..)..

https://preview.ibb.co/frJ1Pw/GGS_compare.jpg

Sorry for watermark , I will  say , "i was there" when your program reaches millions Smiley   
sr. member
Activity: 728
Merit: 304
Miner Developer
Dear Zawawa, i was busy few days, couldnt report back sorry. I can also confirm that , after 1.1.8 to -->1.1.9 some thing went wrong..Gpu reorder is okay but..Only efficient working verison is still 1.1.8  for me.. I couldnt test  after  1.1.9 ,  but  please  search the problem between 1.1.8 and 1.1.9..  

1.1.9 --> temp goes very high, opening lags 1 -2 minutes (with 16 gb system ram +58 gb virtual memory)
1.1.8--->  opening quite  fast,temp is  ok,  hash rate (cyrptonight) is better, than 1.1.9 ...just (gpu reorder problem for me)..

i will test and report with images..Thank you by the way..

edit: at the same time you released 1.1.3  , I will test this one.

edit2: in version 1.1.9 there is an extra file in work directory : GatelessGateSharp.vshost.exe  , at other  verisons I couldnt see  this  one , is this a bug, or you know that already ?

Beginning at Version 1.1.9, GGS runs a diagnostic test at each launch to work around the GPU order problem.
I just realized I could just save the results of the first test, though. We will see...
Pages:
Jump to: