Topic: Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 103. (Read 214431 times)

Didn't Claymore's have a similar requirement?
All sorts of weird things happen with a small page file size, so it's better to stay on the safe side...

In my case it 8+8+4=20. But this cases is very rare. My 16Gb page file is enough. All of my VRAM never used in any cases, because of it 16Gb is quite enough for me.

No, I don't think that is enough. In my experience you sometimes need to have the page file be equal to the size of your combined VRAM, with a bit extra. So in your case I would set it at ~24 GB and see how it does.

I know it. But IMHO 12Gb RAM+16Gb page file size is quite enough for 2xRX 580 8Gb and 1 R9 270X 4Gb, isn't it? But program every launch gives error message. I think that it must be an option to disable this error message.

The miner won't be stable with a small page file size, though. I need to rule out this possible cause for instability.
You can manually increase the page file size in Advanced System Settings.

I have SSD system drive. And I want manage my virtual page size by myself. Please add an option to disable this warning with error sound.

You might want to just increase the page file size instead...
The latest driver is giving me too much headache.
Let me see if I can still work around driver issues...

So I tried it for a while on my RX580 x 6 GPU system.  

Kept hanging the system and had to force reboot multiple times.  On latest driver. Did well on Mh/s but just wasn't stable.  Also need to add decred, but I am sure on your list.

I think your going great so far, just not ready for real production yet.

Even did well on power usage, for Eth it was about -20W than Claymores. 

Too many bugs in 1.1.9 and 1.1.10 with 17.12.2 drivers.
After stopping miner he didn't release RX 580 GPU cores. It continued to be 100% busy. And any attempt to open RadeonSettings or other tool that reads GPU on low level leads to system hang.

And add please option to DISABLE with annoing warning of page size!

17.12.2 drivers fix equal sensors issue.
Please make an option to disable page file size warning!

I found a small bug to report.  If you open the program for the first time, then turn off your new AMD Fan settings, then hit start.  It thinks they are still on.  And the fan was frozen at 20%.  I noticed my GPU raise to about 84C before I killed the program.  Restarting the program with the option off, the fans are now being controlled by the OS (my preferred way).

This should be fixed in the new version.

Gateless Gate Sharp 1.1.10 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.10-alpha

* Improved device detection.
* Added a check for total page file size.

Bug fixes, bug fixes...

Any idea?

Yes, of course. I will try it now. But I use Claymore's miners for setting clocks and voltages for my cards... Let's see how I can change it in 17.12.2.

Could you try Adrenaline 17.12.2? It is THE recommended driver for GSS now.

1.1.10 alpha - the same bug:
All devices shows equal hardware sensors info fan/temp/clocks taken from 1st GPU.

Windows 10 x64 Pro, Relive 17.11.4

@zawawa are you going to integrate equihash? maybe exceed claymore speed 12.6?

Confirm, 1.1.9 alpha miner threads unresponsive. This may be recovered only with coldboot in my case (RX470*2). Restarting of thread never work in 1.1.9 and maybe the thread must be completely dropped with freeing its memory and only then we need to start new mining thread.
I have only tested it on neoscrypt.
Alpha 1.1.6 till 1.1.8 are Ok however. But all of the have some other problems as well.
For example I need  manually start/stop mining after reasonable amount of time ~ 60...90 minutes. Just because pool looses the miner. Maybe you need to add some more logic to miner, completely disconnecting from pool and reconnecting if and when we cannot mine any share for while (3x average time of finding shares) after 5 minutes of mining. Also we can switch to next custom pool in this case.
The next problem is not loading any of the previous custom pools and settings to the program fields when starting GGS. However they all are present in GatelessGateSharp.sqlite. This makes using of the GGS a headache. I use this custom pool settings:
pool.bsod.pw
1995
Ga59VHEiMTYPcnLm2mqmGfwSqUX294RB9q.470x2
c=SPK,d=128

Some piece of log for unresponsive thread case:
2018-01-01 13:37:55.0997 [4] Device #0 submitted a share.
2018-01-01 13:37:55.1477 [5] Share #38 accepted.
2018-01-01 13:37:56.3228 [6] Device #1: 686,77 Kh/s (NeoScrypt)
2018-01-01 13:37:56.7968 [4] Device #0: 708,77 Kh/s (NeoScrypt)
2018-01-01 13:38:06.3284 [6] Device #1: 682,01 Kh/s (NeoScrypt)
2018-01-01 13:38:06.3494 [5] Received new job: 3a
2018-01-01 13:38:13.8398 [5] Received new job: 3b
2018-01-01 13:38:14.8138 [6] Device #1 submitted a share.
2018-01-01 13:38:14.8648 [5] Share #39 accepted.
2018-01-01 13:38:16.3309 [6] Device #1: 690,27 Kh/s (NeoScrypt)
2018-01-01 13:38:26.0525 [6] Device #1 submitted a share.
2018-01-01 13:38:26.1015 [5] Share #40 accepted.
2018-01-01 13:38:26.3455 [6] Device #1: 323,31 Kh/s (NeoScrypt)
2018-01-01 13:38:26.9955 [5] Received new job: 3c
2018-01-01 13:38:36.3501 [6] Device #1: 682,54 Kh/s (NeoScrypt)
2018-01-01 13:38:46.3516 [6] Device #1: 689,51 Kh/s (NeoScrypt)
2018-01-01 13:38:56.3622 [6] Device #1: 681,50 Kh/s (NeoScrypt)
2018-01-01 13:38:59.5754 [1] Miner thread is unresponsive. Restarting...
2018-01-01 13:38:59.5754 [8] Miner thread for Device #0 started.
2018-01-01 13:38:59.5794 [8] Loaded Kernels\neoscrypt.cl for Device #0.
2018-01-01 13:39:06.3738 [6] Device #1: 685,24 Kh/s (NeoScrypt)
2018-01-01 13:39:08.8339 [8] Built NeoScrypt program for Device #0.
2018-01-01 13:39:08.8339 [8] Build options: -O5 -legacy -IKernels -DWORKSIZE=256
2018-01-01 13:39:16.3804 [6] Device #1: 684,43 Kh/s (NeoScrypt)
2018-01-01 13:39:19.9356 [5] Received new job: 3d
2018-01-01 13:39:26.3869 [6] Device #1: 660,76 Kh/s (NeoScrypt)
2018-01-01 13:39:27.5290 [5] Received new job: 3e
2018-01-01 13:39:34.9964 [5] Received new job: 3f
2018-01-01 13:39:36.3895 [6] Device #1: 683,76 Kh/s (NeoScrypt)
2018-01-01 13:39:45.6870 [6] Device #1 submitted a share.
2018-01-01 13:39:45.7370 [5] Share #41 accepted.
2018-01-01 13:39:46.3921 [6] Device #1: 685,95 Kh/s (NeoScrypt)
2018-01-01 13:39:56.3926 [6] Device #1: 687,36 Kh/s (NeoScrypt)
2018-01-01 13:39:58.7938 [6] Device #1 submitted a share.
2018-01-01 13:39:58.8428 [5] Share #42 accepted.
2018-01-01 13:40:03.4561 [6] Device #1 submitted a share.
2018-01-01 13:40:03.5051 [5] Share #43 accepted.
2018-01-01 13:40:06.4012 [6] Device #1: 685,59 Kh/s (NeoScrypt)
2018-01-01 13:40:07.9043 [6] Device #1 submitted a share.
2018-01-01 13:40:07.9563 [5] Share #44 accepted.
2018-01-01 13:40:07.9653 [6] Device #1 submitted a share.
2018-01-01 13:40:08.0153 [5] Share #45 accepted.
2018-01-01 13:40:10.0084 [1] Miner thread is unresponsive. Restarting...
2018-01-01 13:40:10.0084 [9] Miner thread for Device #0 started.
2018-01-01 13:40:10.0124 [9] Loaded Kernels\neoscrypt.cl for Device #0.
2018-01-01 13:40:13.8166 [6] Device #1 submitted a share.
2018-01-01 13:40:13.8656 [5] Share #46 accepted.

got it to mine on neoscript on zpool ty only getting around 650-700kh out of the cards though so waiting on tuning aspect reply to others post i was using AB for my overclock settings and fan speed and left it open all my cards are bios modded but only one is showing the 2150 memory on dashboard there all rx570 6,8gb 2, 4gb. Power use seems to be up about 300 watts also. TY for all your support i chose to work with your miner after seeing the amount of support you do on this thread i will dedicate one of my rigs to mine with SSG as long as its getting close to what it can get on other miners i like what you are doing with this Dev