Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 95.

UnclWish

sr. member

Activity: 1484

Merit: 253

Quote from: zawawa on January 07, 2018, 06:39:18 PM

Gateless Gate Sharp 1.1.18 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.18-alpha

* Adjusted the watchdog process.

@UnclWish There shouldn't be any difference... Let's try this version and see how it goes.

Yes. thanks. I'm just was unlucky ))))

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: UnclWish on January 07, 2018, 12:32:35 PM

It's strange, but on 1.1.17 amount of finded shares became less than it was in 1.1.16... Pool where I mine all day shows accepted speed about 10-20% less than it must to be...

I see a strange drop in profitability on cryptonight.usa.nicehash.com. You are using NiceHash, too, right?
Let me check other pools...

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Gateless Gate Sharp 1.1.18 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.18-alpha

* Adjusted the watchdog process.

@UnclWish There shouldn't be any difference... Let's try this version and see how it goes.

OhMyHolyToken

jr. member

Activity: 98

Merit: 2

Any plans for Equihash (AMD)?

UnclWish

sr. member

Activity: 1484

Merit: 253

It's strange, but on 1.1.17 amount of finded shares became less than it was in 1.1.16... Pool where I mine all day shows accepted speed about 10-20% less than it must to be...

nerdralph

sr. member

Activity: 588

Merit: 251

Quote from: zawawa on January 07, 2018, 03:25:55 AM

Quote from: nerdralph on January 06, 2018, 11:14:07 PM

[...]
I think you'll have to write most of the code in straight asm, as it will be a long time before the optimizer gets good. Some instruction intrinsics like alignbit have only been added in the last few months, and as far as I can tell, there are no optimization hints to indicate different instruction timing.
https://reviews.llvm.org/D34046

I know for a fact that AMDGPU supports the inline GCN assembly for quite some time now. It is just so much easier for me to rewrite kernels in the inline GCN assembly. I might develop a Visual Studio extension for the GCN assembly and work with it. We will see...

I agree with you that llvm-AMDGPU supports inline asm, and that it is easier to write kernels using inline asm vs 100% asm. API setup (Cl1.2/2.0 vs ROCm vs gallium...) and kernel parameter passing is a PITA with pure asm, so I think the best option is to write the majority of the kernel in inline asm, using OpenCL primarily for the API. But since the AMD drivers don't include a version of llvm that supports inline asm, you'll need to distribute the kernels in binary form after building them with a recent version of llvm.

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: maverik77 on January 07, 2018, 10:05:33 AM

How come once I set everything does not save me? I tried but always resets and I have to set it every time

You could delete GatelessGateSharp.sqlite.

maverik77

newbie

Activity: 31

Merit: 0

I've already set up privileges .... run as administrator ...

UnclWish

sr. member

Activity: 1484

Merit: 253

Quote from: maverik77 on January 07, 2018, 10:05:33 AM

How come once I set everything does not save me? I tried but always resets and I have to set it every time

Try to start with administrator privelegies.

maverik77

newbie

Activity: 31

Merit: 0

How come once I set everything does not save me? I tried but always resets and I have to set it every time

icedevil

member

Activity: 242

Merit: 10

When oc will be enable?

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: treanski on January 07, 2018, 06:50:09 AM

Quote from: zawawa on January 07, 2018, 06:19:35 AM

Gateless Gate Sharp 1.1.17 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.17-alpha

* Improved the watchdog process.

It should work so much better now.
If everything goes well, I will make this the next stable version.

do i need to do something to use this watchdog?

No.

treanski

full member

Activity: 364

Merit: 106

ONe Social Network.

Quote from: zawawa on January 07, 2018, 06:19:35 AM

Gateless Gate Sharp 1.1.17 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.17-alpha

* Improved the watchdog process.

It should work so much better now.
If everything goes well, I will make this the next stable version.

do i need to do something to use this watchdog?

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Gateless Gate Sharp 1.1.17 alpha
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.17-alpha

* Improved the watchdog process.

It should work so much better now.
If everything goes well, I will make this the next stable version.

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: UnclWish on January 07, 2018, 02:57:35 AM

GGS monitor didn't work for me... again... GGS just closes if any error and nothing. No program error, no GGS relaunch, no events...

This is very strange... I might need a different approach with GGS Monitor.

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: nerdralph on January 06, 2018, 11:14:07 PM

Quote from: zawawa on January 06, 2018, 06:47:06 PM

(1) LLVM's AMDGPU backend is compatible with Mesa 3D.
(2) Mesa 3D is compatible with Windows.

Does this mean that I could use AMDGPU without modifications!?

At least on the Linux side, as of amdgpu-pro 17, the default OpenCL compiler has been llvm (HSA/ROCm). As you know, it's optimizer for GCN is shit compared to the old gcc version, hence my reason for using -legacy. I haven't tried 17.50 yet, but 17.40 didn't recognize inline asm. There's lots of commits going on, so you'll probably just have to go with 7.0.0/trunk unless the amdgpu updates get back ported to 6.0. The docs are quite thorough now, if you haven't looked at them in the last few months:
https://llvm.org/docs/AMDGPUUsage.html

I think you'll have to write most of the code in straight asm, as it will be a long time before the optimizer gets good. Some instruction intrinsics like alignbit have only been added in the last few months, and as far as I can tell, there are no optimization hints to indicate different instruction timing.
https://reviews.llvm.org/D34046

I know for a fact that AMDGPU supports the inline GCN assembly for quite some time now. It is just so much easier for me to rewrite kernels in the inline GCN assembly. I might develop a Visual Studio extension for the GCN assembly and work with it. We will see...

treanski

full member

Activity: 364

Merit: 106

ONe Social Network.

Quote from: UnclWish on January 07, 2018, 02:57:35 AM

GGS monitor didn't work for me... again... GGS just closes if any error and nothing. No program error, no GGS relaunch, no events...

i have the same problem that is close from time to time

run it in a loop with batchfile

:start
GatelessGateSharp.exe
goto start

UnclWish

sr. member

Activity: 1484

Merit: 253

GGS monitor didn't work for me... again... GGS just closes if any error and nothing. No program error, no GGS relaunch, no events...

zawawa

sr. member

Activity: 728

Merit: 304

Miner Developer

Quote from: UnclWish on January 07, 2018, 12:49:07 AM

Quote from: OhMyHolyToken on January 06, 2018, 11:40:00 PM

Any support of SSL pool? if yes - how? "ssl://pool_address" - doesn't work.

No. At this moment GGS dosn't support ssl. But I don't know pools that have only ssl port...

Support for SSL pools does not seem all that difficult:

https://stackoverflow.com/questions/8375013/how-to-use-ssl-in-tcpclient-class

It might be a good idea to make connections to default pools more secure...

UnclWish

sr. member

Activity: 1484

Merit: 253

Quote from: spyder09 on January 07, 2018, 12:34:49 AM

I havent mined in quite a while, but how is this miner doing compared to claymore? I will be running 2 rx580s for now and since its open source I assume it has no fee? what is everyone averaging on eth with rx 580? thanks in advance

GGS gives speed that is not lee that Claymore, and in some cases faster. It depends on card and bios mod. Mod that gives good speed on Claymore, in GGS can give less or more speed. It's needs to try. GGS work another way than Claymore, and GGS consumpes less power in cryptonight.
Although GGS is open-source it have dev-fee. Author needs an incentive to develop GGS farther.

Topic: Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 95. (Read 214410 times)