Topic: Gateless Gate Sharp 1.3.8: 30Mh/s (Ethash) on RX 480! - page 125. (Read 214431 times)
how many mhs arrives on eth the latest version do you know? 470/480/570/580
This loop does not work with NVIDIA cards...
Code:
ulong8 Skein512Block(ulong8 p, ulong8 h, ulong h8, const ulong *t)
{
#pragma unroll
for(int i = 0; i < 18; ++i)
{
p = SkeinEvenRound(p, h, t, i);
++i;
ulong tmp = h.s0;
h = shuffle(h, (ulong8)(1, 2, 3, 4, 5, 6, 7, 0));
h.s7 = h8;
h8 = tmp;
p = SkeinOddRound(p, h, t, i);
tmp = h.s0;
h = shuffle(h, (ulong8)(1, 2, 3, 4, 5, 6, 7, 0));
h.s7 = h8;
h8 = tmp;
}
SKEIN_INJECT_KEY(p, 18);
return(p);
}
Gateless Gate Sharp 0.0.14 beta
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.14-beta
* Implemented workarounds for bugs in NVIDIA drivers and Cloo.
* Fixed a critical bug that reduces the number of shares for cryptonight.
Feedback is much appreciated. Thanks!
Thanks for update!https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.14-beta
* Implemented workarounds for bugs in NVIDIA drivers and Cloo.
* Fixed a critical bug that reduces the number of shares for cryptonight.
Feedback is much appreciated. Thanks!
But I think that in 10-12 beta versions GSS gives more accepted shares about 5-10% on my RX 580 cards. But I can be wrong...
On RX480 speed comparable to the speed of the Claymore - 22,3Mh/s. Other things being equal - a modified BIOS but no setting of frequencies of a core/memory and voltages.
The interface is nice. But a bit buggy - it seems that the miner uses up all available memory.
The interface is nice. But a bit buggy - it seems that the miner uses up all available memory.
Gateless Gate Sharp 0.0.14 beta
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.14-beta
* Implemented workarounds for bugs in NVIDIA drivers and Cloo.
* Fixed a critical bug that reduces the number of shares for cryptonight.
Feedback is much appreciated. Thanks!
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.14-beta
* Implemented workarounds for bugs in NVIDIA drivers and Cloo.
* Fixed a critical bug that reduces the number of shares for cryptonight.
Feedback is much appreciated. Thanks!
I'm not sure it's really "Sharp" ... feels more like marketing (Nicehash sponsored) targeting mass adoption with convenience of usage... however I wasn't able to make it work proper rates/all cards
though... I was testing it on a rig where I have two Vega FEs and a bunch of 1060... lol.
Seem to be very persistent, as the miner should be. but it does stuff in background, looks heavily multithreaded.
Probably you should add small "sticky" status window that will show complete summary of current activities/brief history
though... I was testing it on a rig where I have two Vega FEs and a bunch of 1060... lol.
Seem to be very persistent, as the miner should be. but it does stuff in background, looks heavily multithreaded.
Probably you should add small "sticky" status window that will show complete summary of current activities/brief history
Interesting, but buggy...any reason main window is so slow?
It says my XMR address is bad...but claymore uses it with no problem
It says my XMR address is bad...but claymore uses it with no problem
can we add our own preferred pool?
Gateless Gate Sharp 0.0.13 beta
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.13-beta
* Added missing DLL's.
Phew!
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.13-beta
* Added missing DLL's.
Phew!
No changes.
The same as previous screen
This is really odd... You are not using a 32-bit Windows, right?
Could you show me the content of the following folders?
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp\x64
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp\x86
https://photos.app.goo.gl/WKNUv7g9ixOxEBVp1
https://photos.app.goo.gl/b07Dzep0uZktcvzw1
Oh, I almost forgot to mention this. I did find a fix for this problem. We will see...
I ended up reorganizing the class hierarchy quite a bit.
It seems that building the cryptonight kernel is rather resource intensive and that repeated calls to clBuildProgram() trigger bugs in NVIDIA drivers and result in corrupted memory.
The miner now reuses program objects, so kernel launches are almost instantaneous after the first build.
It seems that building the cryptonight kernel is rather resource intensive and that repeated calls to clBuildProgram() trigger bugs in NVIDIA drivers and result in corrupted memory.
The miner now reuses program objects, so kernel launches are almost instantaneous after the first build.
very well done i like this daily fixes keep up
I ended up reorganizing the class hierarchy quite a bit.
It seems that building the cryptonight kernel is rather resource intensive and that repeated calls to clBuildProgram() trigger bugs in NVIDIA drivers and result in corrupted memory.
The miner now reuses program objects, so kernel launches are almost instantaneous after the first build.
It seems that building the cryptonight kernel is rather resource intensive and that repeated calls to clBuildProgram() trigger bugs in NVIDIA drivers and result in corrupted memory.
The miner now reuses program objects, so kernel launches are almost instantaneous after the first build.
I just found a terrible bug in cryptonight. So much for "beta" versions...
I will upload a new version shortly.
Is this why it keeps crashing ran great for like 30mins now it wont stay runningI will upload a new version shortly.
This is a different issue altogether.
It's pretty difficult to tip toe around bugs in drivers, you know...
We will see how stable the next version is. So far, so good...
I just found a terrible bug in cryptonight. So much for "beta" versions...
I will upload a new version shortly.
Is this why it keeps crashing ran great for like 30mins now it wont stay running I will upload a new version shortly.
I just found a terrible bug in cryptonight. So much for "beta" versions...
I will upload a new version shortly.
I will upload a new version shortly.
can you add rawintensity for cryptonight?
Gateless Gate Sharp 0.0.13 beta
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.13-beta
* Added missing DLL's.
Phew!
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v0.0.13-beta
* Added missing DLL's.
Phew!
No changes.
The same as previous screen
This is really odd... You are not using a 32-bit Windows, right?
Could you show me the content of the following folders?
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp\x64
C:\Program Files\Zawawa Software LLC\Gateless Gate Sharp\x86
https://photos.app.goo.gl/WKNUv7g9ixOxEBVp1
https://photos.app.goo.gl/b07Dzep0uZktcvzw1
Just downloaded 0.0.13beta and looks good I'm mining monero and getting better rates than the original GG but I have one question how do I switch from the eu to us server on nicehash
Zawawa enabling phymem should i get a better perfomance?
As zawawa wrote before it's not need in phymem at this moment. It's will be need in future.GSS 0.0.13 beta works stable on cryptonight with AMD cards. Only annoying warning 10 sec message on start. Please make it disablebale in settings...
Zawawa enabling phymem should i get a better perfomance?
Really want to start using this miner regularly and I appreciate the work, but on my Mini ITX 1060s I get less than 15-17Mhs max on ethereum when I usually get 19.7 on ethminer and I cannot raise the threads, or I get constant errors. Maybe I am doing something wrong?
Right now I am at intensity 1000 and Local worksize at 128 with max speeds at 17.2
Edit: upped the intensity to 2000 and the Local worksize to 256 and got 18.4.
Right now I am at intensity 1000 and Local worksize at 128 with max speeds at 17.2
Edit: upped the intensity to 2000 and the Local worksize to 256 and got 18.4.
512 may give you a better performance. I will adjust the maximum value.
Jump to: