Ocтaвa caмo дa мaxнeш windows-a и вcичкo щe e нapeд
Topic: Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 157. (Read 214431 times)
@batko
Ocтaвa caмo дa мaxнeш windows-a и вcичкo щe e нapeд
Ocтaвa caмo дa мaxнeш windows-a и вcичкo щe e нapeд
Today I tested RX 480 4GB Nitro+ ELPIDA sgminer vs claymore 9.6
Same clock gpu default clock 1306, elpida memory overclocked to 2050mhz
RX 480 4GB NITRO+ 1306/2050 -36mV
With same setings Sgminer is faster
Same clock gpu default clock 1306, elpida memory overclocked to 2050mhz
RX 480 4GB NITRO+ 1306/2050 -36mV
With same setings Sgminer is faster
Guys,
What settings do you use for GG for ethereum with RX470s?
With Claymore, I get 28MH/s on 470s 4GB and 8GB, but with gatelessgate, the 4GB are stuck at 24.7MH/s (haven't tried it on the 8GB). I'm getting that hashrate with the default xintensity of 4620, worksize of 192 and 2 GPU threads contained in the batfile.
Thanks!
What settings do you use for GG for ethereum with RX470s?
With Claymore, I get 28MH/s on 470s 4GB and 8GB, but with gatelessgate, the 4GB are stuck at 24.7MH/s (haven't tried it on the 8GB). I'm getting that hashrate with the default xintensity of 4620, worksize of 192 and 2 GPU threads contained in the batfile.
Thanks!
me i don't use gateless to mine ETH or ZEC but use it for all my other coins like XMR PASC etc that I mine ...
This miner better for Cryptonight (XMR) on my R-380´s (560-635H/s, but ~2% rejected). Claymore 550-650H/s (less rejected, but 2.5% fee).
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
Are you using 16.04.1 or 16.04.2? 16.04.2 desktop includes kernel 4.8.0 instead of 4.4.0. With 4.8.0 I get /sys/class/drm entries for my Rx 470 even before installing the amdgpu-pro driver, while 4.4.0 did not.
I had problems with 4.4.0 and amdgpu-pro 16.40, and am only now getting around to testing 4.8.0.
Quote
clinfo-amdgpu-pro install
libdrm-amdgpu-pro-amdgpu1:amd64 install
libdrm-amdgpu-pro-dev:amd64 install
libdrm-amdgpu-pro-utils install
libdrm2-amdgpu-pro:amd64 install
libegl1-amdgpu-pro:amd64 install
libgbm1-amdgpu-pro:amd64 install
libgbm1-amdgpu-pro-base install
libgbm1-amdgpu-pro-dev:amd64 install
libgl1-amdgpu-pro-appprofiles install
libopencl1-amdgpu-pro:amd64 install
opencl-amdgpu-pro-icd:amd64 install
libdrm-amdgpu-pro-amdgpu1:amd64 install
libdrm-amdgpu-pro-dev:amd64 install
libdrm-amdgpu-pro-utils install
libdrm2-amdgpu-pro:amd64 install
libegl1-amdgpu-pro:amd64 install
libgbm1-amdgpu-pro:amd64 install
libgbm1-amdgpu-pro-base install
libgbm1-amdgpu-pro-dev:amd64 install
libgl1-amdgpu-pro-appprofiles install
libopencl1-amdgpu-pro:amd64 install
opencl-amdgpu-pro-icd:amd64 install
i'm using cpp-ethereum/build/ethminer/ethminer to mine ETH on this machines (several of them have same Xorg memory leakage)
bit offtopic but will ask if anyone have same memory leaks as i am:
Ubuntu 16.04.1 LTS
latest AMDPRO drivers
Code:
$ free
total used free shared buff/cache available
Mem: 8118412 6358508 640888 21812 1119016 1424880
Swap: 0 0 0
user@miner3:~$ top
top - 10:06:15 up 5 days, 18:06, 2 users, load average: 0.06, 0.05, 0.01
Tasks: 282 total, 2 running, 280 sleeping, 0 stopped, 0 zombie
%Cpu(s): 1.6 us, 0.5 sy, 0.0 ni, 97.7 id, 0.2 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 8118412 total, 639916 free, 6359212 used, 1119284 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 1424076 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1180 root 20 0 5516208 5.036g 41400 S 6.7 65.0 142:42.87 Xorg
1879 user 20 0 657276 37012 28080 S 6.7 0.5 67:47.74 gnome-terminal-
1 root 20 0 120824 6924 3936 S 0.0 0.1 0:02.17 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:00.02 kthreadd
3 root 20 0 0 0 0 S 0.0 0.0 0:04.21 ksoftirqd/0
5 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0H
7 root 20 0 0 0 0 S 0.0 0.0 12:53.01 rcu_sched
Ubuntu 16.04.1 LTS
latest AMDPRO drivers
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
Are you using 16.04.1 or 16.04.2? 16.04.2 desktop includes kernel 4.8.0 instead of 4.4.0. With 4.8.0 I get /sys/class/drm entries for my Rx 470 even before installing the amdgpu-pro driver, while 4.4.0 did not.
I had problems with 4.4.0 and amdgpu-pro 16.40, and am only now getting around to testing 4.8.0.
It turned out that it takes too long to make the new compiler compatible with GCN1, so I decided to keep the previous assembly version for it. Oh well.
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
Are you saying that mining with ADMGPU-PRO is finally working even when you mod your cards? I remember last year when I got my RX470s and AMDGPU-PRO was required to make them work, as soon as you modded them, millions of errors would be thrown in the syslog and the rig was highly unstable. This is the only reason I switched to window$...
My AMDGPU-PRO rigs are pretty stable so far. I'm overhauling my farm right now, so I will keep you updated.
Guys,
What settings do you use for GG for ethereum with RX470s?
With Claymore, I get 28MH/s on 470s 4GB and 8GB, but with gatelessgate, the 4GB are stuck at 24.7MH/s (haven't tried it on the 8GB). I'm getting that hashrate with the default xintensity of 4620, worksize of 192 and 2 GPU threads contained in the batfile.
Thanks!
What settings do you use for GG for ethereum with RX470s?
With Claymore, I get 28MH/s on 470s 4GB and 8GB, but with gatelessgate, the 4GB are stuck at 24.7MH/s (haven't tried it on the 8GB). I'm getting that hashrate with the default xintensity of 4620, worksize of 192 and 2 GPU threads contained in the batfile.
Thanks!
me i don't use gateless to mine ETH or ZEC but use it for all my other coins like XMR PASC etc that I mine ...I'm pretty much like most who feel thy are forced to use claymore and before anyone says this or that i don't care and also don't mean it how it sounds. but did have it hitting 26 MH with my 470's mining eth when i did try one less MH then i do with CM .. but CM is more Stable atm for long term mining .
yeah. claymore is good, also dual mine is useful.
I've tried ETH on SGminer and the fastest was 5.5.4 for some reason... it gave me just below 27 at the top, while in claymore I get 28.5 in dual mode and 27.5 eth only
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
Are you saying that mining with ADMGPU-PRO is finally working even when you mod your cards? I remember last year when I got my RX470s and AMDGPU-PRO was required to make them work, as soon as you modded them, millions of errors would be thrown in the syslog and the rig was highly unstable. This is the only reason I switched to window$...
Good news, guys! ZEC mining is working with fglrx and 7990, including the assembly version.
I will update the repository either tonight or tomorrow morning.
Edit: I just pushed the changes to the repo. Let me know how that goes.
I will update the repository either tonight or tomorrow morning.
Edit: I just pushed the changes to the repo. Let me know how that goes.
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
Thank you very much for working on the fglrx driver support! AMDGPU-PRO has virtually no GPU control on linux, other than fan and temp, and they require root privileges, which is nonsense (being a member of the "video" group aught to be enough, you would think).
I am trying to extend GG's Linux support by making it compatible with fglrx. Not exactly a fun thing to do, but still...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
By the way, I have been using AMDGPU-PRO 16.60 for a while on Ubuntu 16.04 LTS without problems. All I had to do was to upgrade the Linux kernel to 4.10.1. It's amazing that this requirement is documented nowhere...
That rig averages 4200Kh/s on XMR, but I feel it could do better. I do not have a lot of time to do tests though
. Also since GG, Wattool and GPUz all have their own way to "assign" cards it's kind of a nightmare figuring out which is which on that rig, lol.All cores run at 1100MHz and I'm bios undervolted at -137.5mv.
Trick I use is to set different mhz in bios for each card, so in my rig left to right, i have 1141, 1142, 1143, etc.
root@rig14:/home/dallas# clocks
1: 2012Mhz * # GPU 0
7: 1142Mhz *
1: 2013Mhz * # GPU 1
7: 1143Mhz *
1: 2006Mhz * # GPU 2
7: 1146Mhz *
1: 2015Mhz * # GPU 3
7: 1145Mhz *
1: 2014Mhz * # GPU 4
7: 1144Mhz *
1: 2011Mhz * # GPU 5
7: 1141Mhz *
Then I know GPU #2 reported above is the furthest card on the right, GPU #5 is furthest left
most common order is PCIe x16 is where GPU0 sits, looking at the DVI/HDMI connectors on the left of GPU0 is GPU1 and on the right are GPU2,GPU3, GPU4,GPU5, in that order.
That rig averages 4200Kh/s on XMR, but I feel it could do better. I do not have a lot of time to do tests though
. Also since GG, Wattool and GPUz all have their own way to "assign" cards it's kind of a nightmare figuring out which is which on that rig, lol.All cores run at 1100MHz and I'm bios undervolted at -137.5mv.
Trick I use is to set different mhz in bios for each card, so in my rig left to right, i have 1141, 1142, 1143, etc.
root@rig14:/home/dallas# clocks
1: 2012Mhz * # GPU 0
7: 1142Mhz *
1: 2013Mhz * # GPU 1
7: 1143Mhz *
1: 2006Mhz * # GPU 2
7: 1146Mhz *
1: 2015Mhz * # GPU 3
7: 1145Mhz *
1: 2014Mhz * # GPU 4
7: 1144Mhz *
1: 2011Mhz * # GPU 5
7: 1141Mhz *
Then I know GPU #2 reported above is the furthest card on the right, GPU #5 is furthest left
Sorry guys, my wife is not feeling well right now and I didn't have a chance to work on the miner today.
Just to let you know.
Just to let you know.
appreciated! Hope she'll get well soon.
Sorry guys, my wife is not feeling well right now and I didn't have a chance to work on the miner today.
Just to let you know.
Just to let you know.
No worries dude
So, what kind of timing/frequencies does everyone run around here?
I've got a rig with 6 RX470 4GB (2 Hynix, 2 Elpida and 2 Samsung).
On the hynix: 1375 strap and memory @ 2000
On the elpida: 1500 strap and memory @ 2000 (I think it can't go higher, Watttool just resets the memory to 2000 when I try something higher)
On the samsung: 1750 and memory @ 2000.
That rig averages 4200Kh/s on XMR, but I feel it could do better. I do not have a lot of time to do tests though
. Also since GG, Wattool and GPUz all have their own way to "assign" cards it's kind of a nightmare figuring out which is which on that rig, lol.All cores run at 1100MHz and I'm bios undervolted at -137.5mv.
Sorry guys, my wife is not feeling well right now and I didn't have a chance to work on the miner today.
Just to let you know.
Just to let you know.
well... its confirmed...
in current implementation of cryptonite samsung with 1625 straps give better hash at 1750 than 2000.
And now my question, once again to the algo devs: wtf?
in current implementation of cryptonite samsung with 1625 straps give better hash at 1750 than 2000.
And now my question, once again to the algo devs: wtf?
For my 470 Samsung cards, I loosen the 1625 strap just a bit and run mem @ 2060-2080mhz, which allows me to get good XMR and ETH rates. Its unfortunate there is no silver bullet, because ETH wants one thing, and XMR another, and ZEC yet another. So I've decided to just keep my core clocks lower and run XMR when its warm outside.
Also... 470 Hynix can do 750 XMR with 1375 strap @ 1780-1800mhz... but that makes it useless for ETH. But if you are setting up XMR rigs, and want lowest power option, that might be a good option.
470 hynix does 820 here with 1375 straps and higher mem freq.
Jump to: