More than once in the past week, I had a crash…
Here’s what I got:
Using Optiminer 1.5.0 on Ubuntu Linux 16.04.
I gave 1.5.0 another try today. Same issue, higher I/s but near 0 S/s using flypool. I cant be the only person seeing this problem?
So here is a rig Ubuntu 16.04 desktop, AMD 16.50 drivers, 7 Sapphire Nitro RX480 cards, and optiminer 1.4.0. Very stable and has been running for several weeks without issues.
I upload optiminer 1.5.0, unpack, launch miner and I get this:
Go back to 1.4.0 and all is fine again. I have tried the AMD 16.40 drivers and its the same. Gone through the release notes and nothing has changed in the configuration so that leaves my BIOS timing mod on the GPU’s. However, that does not make sense and I am not the only person using that mod so I should not be the only person having this problem.
SOLVED:
I saw in another post that -i 8 is not working in 1.5.0 even with 8GB cards (I was running with -i 10). I dropped to -i 7 and started working.
@Optiminer , can you update your miner so that it works out-of-the-box for CentOS 7.3 ?
On it, the libstdc++ provided is from GCC 4.8 instead of the needed 4.9.
Usually, that means you update libstdc++/gcc to the application dependency. Is optiminer supposed to work with every older library out there?
That would have been on bitcointalk? There was another miner there with a similar problem; I actually thought it might have been you, at the time I saw the post. It’s working! Super! 
Navigate to System and Security System
Advanced system settings
Click the Hardware tab
click Device Installation Settings
and select the “No, let me choose what to do option
Select “Never install install driver software from Windows Update
1 Like
Does Optiminer support SSL for SuprNova?
Hi guys is there any options in the miner to disable enable some cards?
Yes, run the optiminer binary with the -h option to see the flags you can use:
./optiminer-zcash -help
USAGE:
./optiminer-zcash [–benchmark ] [–list-devices]
[–force-generic-kernel] [–experimental-kernel]
[–nodevfee] [-i ] … [-m ]
[–watchdog-cmd ] [–watchdog-timeout
] [–max-log-files ]
[–max-log-file-size ] [-l ] [-v]
[-p ] [-u ] [-s host:port] [-d
] … [-c ] [–] [–version] [-h]
Where:
–benchmark
If specified runs a benchmark for the given number of seconds on all
specified devices and then exists the miner.
–list-devices
List all recognized devices with their platform and device id and
quit.
–force-generic-kernel
Force use of generic kernel even when a device-specific kernel is
available
–experimental-kernel
Unlock faster experimental kernel for certain graphic cards.
–nodevfee
If set, the dev fee will be disabled. Miner will run a bit slower.
Enable this if you do not want any more improvements of the miner and
want to earn less.
-i , --intensity (accepted multiple times)
Worker intensity. 0 means auto-detect based on available memory.
Higher values use more GPU memory. Can be specified once applying the
same intensity to all devices or once per device (same order as -d).
-m , --monitoring-port
The monitoring port to listen on for HTTP requests. Disabled by
default. If enabled, accepts requests from everywhere.
–watchdog-cmd
The watchdog command to execute. See --watchdog-timeout.
–watchdog-timeout
Timeout after which the watchdog triggers if a GPU does not produce
any solutions. It will execute the command specified by
–watchdog-cmd. You can use this command to do an appropriate action
(e.g. reset driver or reboot). 0 disables watchdog.
–max-log-files
Maximum number of rotated log files to keep.
–max-log-file-size
Maximum size of log file before it gets rotated.
-l , --log-file
Write logs to given file
-v, --verbose
Verbose logging.
-p , --password
Stratum password.
-u , --user
Stratum user.
-s host:port, --stratum host:port
Host and port of the stratum server to use.
-d , --device (accepted multiple times)
A OpenCL device id to use. If no devices are specified, all are used.
-c , --platform
The OpenCL platform id to use.
–, --ignore_rest
Ignores the rest of the labeled arguments following this flag.
–version
Displays version information and exits.
-h, --help
Displays usage information and exits.
(C) by Optiminer 2017
so if you want to run cards 0, 1, 3, 4, and 5 put this in the middle of your start shell command line like so:
while true
do
./optiminer-zcash -s $POOL -u $USER -p $PASSWORD -d 0 -d 1 -d 3 -d 4 -d 5 --watchdog-timeout 30 --watchdog-cmd “./watchdog-cmd.sh”
With optiminer 1.5 the miner is unstable using same parameter of OC than 1.4, I lower a little bit but every 6h or so the miner restart itself and every two day it hung. Should I lower OC mem of gpus or intensity first? I am using 6xRX480 with 1330/2100 (mod strap 1750) and i 5 and getting 300S/s each gpu. If I lower more I will be getting less Sol per gpu than version 1.4 (294S/s). I would like at least same Sol that 1.4. What do you recommend? Thanks in advance
I forgot to mention, OS is linux (ethos 1.9)
Lowering intensity doesnt solve the GPU hangs for me. I’m running -i 2 now, which is better, but far from stable. I run mostly rx470s @ 1125/1945 w/ 1500 straps.
I personally think its the straps are too tight. I think optiminer code is getting really efficient, and those timing mods are too tight now.
I can run ethereum and monero miners all day, but I think maybe they are a bit less efficient. When it comes to memory timings, a couple clock cycles can make all the difference.
I dont want to move to 1625 straps because of performance impact, so I’m currently working on tweaking my 1500 strap (ref: Custom RAM Timings for GPU's with GDDR5 - DOWNLOAD LINKS - UPDATED) so I can loosen them up just a bit and regain stability without impacting performance. But its a slow process
So if I made a recommendation to you, it would be lower your mem clock to 2050mzh, because that will loosen up your 1750 strap just enough I think. Then that may allow you to increase your --intensity again.
Wow, I can not get more than 2000 with R9 390 stock bios and opti 1.5 
would you mind to share more info?
Version 1.6.0 released! 
[1.6.0] Asm support for GCN 1 devices.
[1.6.0] Reduced CPU utilization.
[1.6.0] Fix segfault on reconnect.
[1.6.0] Version and Os exported in monitoring.
Speed increase about 3% for newer cards, the GCN 1 cards should see a much higher increase with the asm kernel now.
3 Likes
Iduro, I am using a 6xR9 390 but I can not get more than 340 S/s.
Optiminer, with the new release, for R9 390 it will be more stable?
i use ethOS
globalcore 1100
You tell me… I do not run heavily tuned GPUs… it is likely that the optimal power and clock setting change with each release.
what OS are you using?
Thanks @Optiminer! Just upgraded one of my 6-rig miners with Sapphire Nitro Fury X cards. Version 1.5 was doing 2633 and version 1.6 is knocking out 2720!