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

[YourMom420]

I stalled newest version and its still displaying the number of gpus wrong. Atleast its down to 9 from 17 lol. If i run it they all mine but crash pretty quickly

[hksfrank]

Current Roadmap (Subject to Change)
3. On-the-fly memory timing mods and other optimizations (two weeks?)

I've been thinking of doing that since I wrote my strapmod uitilites several months ago.
https://github.com/nerdralph/strapread

strapmod.py makes a few changes that improve performance, and seem to be safe on Samsung, Elpida/Micron and Hynix RAM.  Modifying other parameters in a generic way is much more problematic.
I'd be interested if you find anything that can identify the specific RAM type.  Utilities like amdmeminfo can tell you a card has H5GC4H24AJR memory on it, but not if it is -ROC or -T2C.

yes i am using your strap mod (well it was on the web tools before server gone away and i did write it down)
for 570 it gives me 880H at 1100/1950 which was kinda good
more core or memory clock did give little improvement about 7-10H per 25 core / 50 memory  clock

much thanks =D
and people say there are better strap for cryptonight and i stilll unable to figure it out , how actually a card or strap bottleneck the hash rate =D

also GGS was a good stuff too
lots of update and bug fixes , i will be happy if the update button was added in the miner +__+

[zawawa]

I stalled newest version and its still displaying the number of gpus wrong. Atleast its down to 9 from 17 lol. If i run it they all mine but crash pretty quickly

There must be something different about your system... Could I see the output of clinfo?

[ayiphelmy]

Gateless Gate Sharp 1.1.21 beta
https://github.com/zawawawa/GatelessGateSharp/releases/tag/v1.1.21-beta

* Fixed fan control.

Here it is!

what about neoscrypt? i still can't use your miner in hawaii and 1070ti... i anly tested in 1 rig... 40 to go

See I don't have 1070 Ti right now... I have 1080 and 1070, though. Let me see if they work...

take your time... i know u can do it...

[zawawa]

also GGS was a good stuff too
lots of update and bug fixes , i will be happy if the update button was added in the miner +__+

I know! I think I will have to add this feature very soon, or I would go crazy myself...

[nerdralph]

Tighter timings can't be pushed as far, so I suspect that is really the issue.  i.e. I think it is safe to apply the strapmod changes (RRD=5, FAW&32AW=0, ACTRD=16), but you won't be able to overclock the memory as much.  If 2150 is stable without the strapmod changes, then I bet 2050-2100 should still be stable with the changes.

But will strapmod give more speed than 100MHz overclock? And for wich algos strapmod gives more speed?

Probably for ethash.  And probably for equihash too, though there is room for more tweaking when it comes to equihash.  I haven't tested cryptonight.
Memory will consume 5% less power at 2050 vs 2150, and so the memory will also run cooler.

See, I did months of experiments with memory timing mods, and I didn't find these strapmod changes particularly safe, at least with RX 470/480. 32AW=0 was especially dangerous, and I also found that ACTRD=16 to be too low in some cases. With Equihash, anything below ACTRD=23 hurts performance IIRC. I cannot wait to get back to these mods. We will see...

Thanks for the info.  I'm quite surprised about 32AW, as I have tried it on Polaris cards with Hynix and Samsung, and on Tonga cards with Elpida and Hynix RAM.  I guess I'll have to try higher clocks with non-zero 32AW to see.
I also beg to differ that ACTRD<23 hurts performance, at least with Samsung RAM.  Here's the stock 1500 strap for a card with Samsung K4G41325FE, which has ACTRD=21.
555000000000000022CC1C00AD515A3EC0570E142D4A64080048C700030114207A8900A00300000 0150E2A3186272E16

[ayiphelmy]

i have good news for you....


a little more tweak for mining speed... this is 25 intensity but the mining speed is way to low...
by the way good job so far

[singlass]

also GGS was a good stuff too
lots of update and bug fixes , i will be happy if the update button was added in the miner +__+

I know! I think I will have to add this feature very soon, or I would go crazy myself...

also thank you for considering making a command line version,till then i'll be rocking your old work GG which is still amazing

[YourMom420]

I stalled newest version and its still displaying the number of gpus wrong. Atleast its down to 9 from 17 lol. If i run it they all mine but crash pretty quickly

There must be something different about your system... Could I see the output of clinfo?

ill post it as soon as im home. Should be a little over an hour

[d0tcould]

Hi there

I have been mining cryptonight with xmr-stak-amd but I'm wanting to mine a different coin now which is neoscrypt and I came across this miner. I have 4x rx 570 4gb (bios modded) so I'm just wondering what some good settings would be to get started? Currently using threads 1, intensity 16, worksize 256 (this was from a config file I came across in a tutorial for GatelessGate but it might not be right for my gpu and or the newer version)

Any help would be greatly appreciated, thanks!

[zawawa]

Hi there

I have been mining cryptonight with xmr-stak-amd but I'm wanting to mine a different coin now which is neoscrypt and I came across this miner. I have 4x rx 570 4gb (bios modded) so I'm just wondering what some good settings would be to get started? Currently using threads 1, intensity 16, worksize 256 (this was from a config file I came across in a tutorial for GatelessGate but it might not be right for my gpu and or the newer version)

Any help would be greatly appreciated, thanks!

If you are using Gateless Gate Sharp, please just hit the "Reset All" button, and you should be good to go.
These numbers don't apply to the new version and actually way off. You can then slowly increase the intensity and see how that would affect the hashrate.

[zawawa]

Overclocking is finally coming! I hope to get it done today. This is exciting~

[YourMom420]

I stalled newest version and its still displaying the number of gpus wrong. Atleast its down to 9 from 17 lol. If i run it they all mine but crash pretty quickly

There must be something different about your system... Could I see the output of clinfo?

ill post it as soon as im home. Should be a little over an hour

Im trying to copy the log but its not working. Is that the info you need? It also looks like shutting off half the gpus helped and thats ok with me lo

[zawawa]

I stalled newest version and its still displaying the number of gpus wrong. Atleast its down to 9 from 17 lol. If i run it they all mine but crash pretty quickly

There must be something different about your system... Could I see the output of clinfo?

ill post it as soon as im home. Should be a little over an hour

Im trying to copy the log but its not working. Is that the info you need?

No, no. Please run this command from a command prompt and copy and paste the output here:

clinfo

Your output should look something like this:

https://i.imgur.com/sz6zRAL.png

[YourMom420]

ill have to sign up for imgur later. its saying too many characters and wont post
heres the start though

C:\Users\name>clinfo
Number of platforms:                             3
  Platform Profile:                              FULL_PROFILE
  Platform Version:                              OpenCL 2.0 AMD-APP (2482.5)
  Platform Name:                                 AMD Accelerated Parallel Processing
  Platform Vendor:                               Advanced Micro Devices, Inc.
  Platform Extensions:                           cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
  Platform Profile:                              FULL_PROFILE
  Platform Version:                              OpenCL 2.0 AMD-APP (2482.5)
  Platform Name:                                 AMD Accelerated Parallel Processing
  Platform Vendor:                               Advanced Micro Devices, Inc.
  Platform Extensions:                           cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices
  Platform Profile:                              FULL_PROFILE
  Platform Version:                              OpenCL 1.2 CUDA 9.1.104
  Platform Name:                                 NVIDIA CUDA
  Platform Vendor:                               NVIDIA Corporation
  Platform Extensions:                           cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_copy_opts cl_nv_create_buffer

[tigr72]

How to configure remote monitoring? 

[agus_sna]

Thank you very much for publishing your repo! Appreciated.

I'm not a C programmer (or OpenCL for this matter) but I'm a DRY fan; so when I'm reading input.cl I found the get_row () function and I think we can make it a little DRYer by doing something like this:

Code:

uint get_row (uint round, uint xi0)
{
uint row;
uint swp;
uint num;
#if NR_ROWS_LOG == 14
swp = 0;
#elif NR_ROWS_LOG == 15
swp = 1;
#elif NR_ROWS_LOG == 16
swp = 2;
# else
#error "not supported NR_ROWS_LOG"
# endif
num = (40 << SWP) - 1);
if (! (round% 2))
row = (xi0 & ((num << 8 | 0xff));
else
row = ((xi0 & (num << 16 | 0xf00)) >> | ((xi0 & 0xf0000000) >> 24);
return line;
}

[d0tcould]

Hi there

I have been mining cryptonight with xmr-stak-amd but I'm wanting to mine a different coin now which is neoscrypt and I came across this miner. I have 4x rx 570 4gb (bios modded) so I'm just wondering what some good settings would be to get started? Currently using threads 1, intensity 16, worksize 256 (this was from a config file I came across in a tutorial for GatelessGate but it might not be right for my gpu and or the newer version)

Any help would be greatly appreciated, thanks!

If you are using Gateless Gate Sharp, please just hit the "Reset All" button, and you should be good to go.
These numbers don't apply to the new version and actually way off. You can then slowly increase the intensity and see how that would affect the hashrate.

thanks for the reply, i will give it a shot!

[zawawa]

Not sure, but I understood that we should connect to the closest mining pool.
Currently I am using the default pool Nicehash.

HK or IN seems best from SE Asia but GGS mostly connects to USA sometimes to JP.

This is the same on v1.1.14 and v1.1.21

These are the ping result from my location:

ping cryptonight.hk.nicehash.com -n 1
Pinging cryptonight.hk.nicehash.com [119.81.240.194] with 32 bytes of data:
Reply from 119.81.240.194: bytes=32 time=76ms TTL=45

ping cryptonight.in.nicehash.com -n 1
Pinging cryptonight.in.nicehash.com [169.38.78.149] with 32 bytes of data:
Reply from 169.38.78.149: bytes=32 time=79ms TTL=46

ping cryptonight.jp.nicehash.com -n 1
Pinging cryptonight.jp.nicehash.com [161.202.120.197] with 32 bytes of data:
Reply from 161.202.120.197: bytes=32 time=118ms TTL=45

ping cryptonight.usa.nicehash.com -n 1
Pinging cryptonight.usa.nicehash.com [169.62.79.78] with 32 bytes of data:
Reply from 169.62.79.78: bytes=32 time=219ms TTL=42

I am using a modded vbios for my RX580 without any further OC/undervolting

Also, when I try to get statistics from www.nicehash.com I often get "Slow it down a bit ... You have reached your API request quota limit. You can try again in x seconds".

Any chance to opt-out at GGS to pull stats from nicehash or to adjust the frequency?

GGS chooses the best server based on NiceHash’es recommendations, so no worries.
I can add an option to disable updates of stats.

[treanski]

also GGS was a good stuff too
lots of update and bug fixes , i will be happy if the update button was added in the miner +__+

I know! I think I will have to add this feature very soon, or I would go crazy myself...

will you fix the gpu order problem too?
still not working for me