Hi doktor83. After upgrade on version 1.8.7 because i catch error "- Fixed a bug with pool reconnect on job timeout" on version 1.8.6 Now after miner start i always have in log mesage. [2019-05-04 10:33:56] hashrate: GPU0: 0 H/s [T:48c][BUS:1] [2019-05-04 10:33:56] hashrate: GPU1: 0 H/s [T:46c][BUS:4] [2019-05-04 10:33:56] hashrate: GPU2: 0 H/s [T:41c][BUS:6] [2019-05-04 10:33:56] hashrate: GPU3: 0 H/s [T:49c][BUS:8] [2019-05-04 10:33:56] hashrate: GPU4: 0 H/s [T:48c][BUS:9] [2019-05-04 10:33:56] hashrate: GPU5: 0 H/s [T:46c][BUS:10] [2019-05-04 10:33:56] hashrate: GPU6: 0 H/s [T:43c][BUS:11] [2019-05-04 10:33:56] hashrate: GPU7: 0 H/s [T:50c][BUS:12] [2019-05-04 10:33:56] hashrate: GPU8: 0 H/s [T:42c][BUS:13] [2019-05-04 10:33:56] hashrate: GPU9: 0 H/s [T:40c][BUS:14] [2019-05-04 10:33:56] hashrate: GPU10: 0 H/s [T:45c][BUS:15] [2019-05-04 10:33:56] hashrate: GPU11: 0 H/s [T:37c][BUS:16] [2019-05-04 10:33:56] hashrate: GPU12: 0 H/s [T:39c][BUS:17] [2019-05-04 10:33:56] hashrate: Total: 0 H/s After this, the speed is normalized and working several minutes. And then message with all card is 0H/s repeat and computer restart. This behaviour on 2 my rig. Log here https://pastebin.com/uyxj7wmqHappens because kernel precompilation takes a lot of time on those slow cards so watchdog triggers in the mean time. This v4 precompilation It also fu**s up the min_rig_speed. Disable watchdog or set it to disable mode - temporary fix. I will try to do something with this.. |
|
|
|
The latest version 1.8.7 works perfectly for upx2. CPU is not overloaded. On the contrary on the same rig running at the same time srbminer and xmrig.
Will try it out, i used the 1.8.6 version yesterday. Otherwise i will use the "old_mode" solution. Thanks a lot:) I tested the 1.8.7 version and the "old_mode" method and in my case none of them helped. Still the cpu is at 99% and it's throttling my gpu's. With "old_mode" : true the hashrate drops even further, around 22kh/s per card. Is there anything else i can do? Doctor, did you found out what's causing the cpu overload? Nah, not really possible that with old mode enabled still 99%. Yes, hashrate is little lower on some algos with it enabled because more work is done on the GPU, and almost none on the CPU. I exactly know why it's giving more load on the cpu, because upx job round is done in less than 100ms, and a 1,2,4mb scratchpaded job is done in ~2s. Rewriting some stuff to use intrinsics, so cpu load will be lower. But still old mode enabled should decrease your cpu load a lot. Fixed the issue, my config was not properly structured and it didn't initialize old_mode. Now the hashrate is great and the cpu workload is around 55%. Great job doctor. Currently i achieved to lower the cpu usage by around 40% without the old_mode enabled, still working on it, maybe i can lower it more. Weak CPU + lot of GPU's in rig + low difficulty can really make your CPU sweat. I am surprised that your cpu usage is 55% with old_mode, that must be some really weak CPU in there  |
|
|
|
Hi Doktor!
Noriced when using gpureorder hashrate table and found share list are not equal.
Hashrate table is correct.
2019-05-02 13:33:42] GPU1[BUS:18] result 0x2AC35BCD accepted for job[7] [195ms]
[2019-05-02 13:33:42] GPU0 : 56304 H/s [T: 64c, RPM: 3302, CC: 1408 MHz, MC: 970 MHz][BUS:3]
[2019-05-02 13:33:42] GPU1 : 61957 H/s [T: 65c, RPM: 3499, CC: 1408 MHz, MC: 1100 MHz][BUS:8]
[2019-05-02 13:33:42] GPU2 : 61483 H/s [T: 63c, RPM: 3313, CC: 1408 MHz, MC: 1095 MHz][BUS:11]
[2019-05-02 13:33:42] GPU3 : 56613 H/s [T: 57c, RPM: 3635, CC: 1408 MHz, MC: 990 MHz][BUS:18]
[2019-05-02 13:33:42] GPU4 : 59985 H/s [T: 64c, RPM: 3724, CC: 1408 MHz, MC: 1075 MHz][BUS:22]
[2019-05-02 13:33:42] GPU5 : 67075 H/s [T: 60c, RPM: 3295, CC: 1408 MHz, MC: 1075 MHz][BUS:25]
[2019-05-02 13:33:42] Total: 363417 H/s
[2019-05-02 13:33:46] GPU0[BUS:3] result 0x156FBE0E accepted for job[7] [201ms]
[2019-05-02 13:33:47] GPU5[BUS:11] result 0xD5726E9D accepted for job[7] [195ms]
[2019-05-02 13:33:47] GPU3[BUS:8] result 0x801D62F8 accepted for job[7] [194ms]
[2019-05-02 13:33:54] GPU3[BUS:8] result 0x9575D679 accepted for job[7] [200ms]
[2019-05-02 13:33:56] GPU4[BUS:25] result 0xAACF5A44 accepted for job[7] [201ms]
[2019-05-02 13:33:58] GPU1[BUS:18] result 0x401FDA9E accepted for job[7] [201ms]
[2019-05-02 13:34:08] GPU3[BUS:8] result 0x80274AAD accepted for job[7] [195ms]
[2019-05-02 13:34:10] GPU2[BUS:22] result 0x557C5F8A accepted for job[7] [194ms]
[2019-05-02 13:34:12] GPU2[BUS:22] result 0x6AD29592 accepted for job[7] [195ms]
[2019-05-02 13:34:15] GPU2[BUS:22] result 0x557EBD5A accepted for job[7] [195ms]
[2019-05-02 13:34:17] GPU5[BUS:11] result 0xD5802A03 accepted for job[7] [194ms]
[2019-05-02 13:34:17] GPU3[BUS:8] result 0x95807C9D accepted for job[7] [201ms]
[2019-05-02 13:34:34] GPU2[BUS:22] result 0x6ADC6630 accepted for job[7] [194ms]
[2019-05-02 13:34:35] GPU1[BUS:18] result 0x2AD9FDCB accepted for job[7] [194ms]
Also, parameters configured in gpu config applies to unexpected GPUs.
ver. 1.8.7
Going to check this out, thanks for reporting. Bus is always the right one, i adjust the gpu number related to bus id. |
|
|
|
hi Dok,
if i have ElioVP tool open in windows, SRB miner starts without tweak_profiles and i cant enable them. Vice versa if i start SRB miner i cant open ElioVP GUI tool, gives error and it crashes. Is that intended as tweak profiles are basically those suggested timings on the elioVP topic on bitcointalk?
Intended - you even get a message about it in the log, and no, the timings are not predefined, the values are calculated based on your values found at the moment of miner start. |
|
|
|
The latest version 1.8.7 works perfectly for upx2. CPU is not overloaded. On the contrary on the same rig running at the same time srbminer and xmrig.
Will try it out, i used the 1.8.6 version yesterday. Otherwise i will use the "old_mode" solution. Thanks a lot:) I tested the 1.8.7 version and the "old_mode" method and in my case none of them helped. Still the cpu is at 99% and it's throttling my gpu's. With "old_mode" : true the hashrate drops even further, around 22kh/s per card. Is there anything else i can do? Doctor, did you found out what's causing the cpu overload? Nah, not really possible that with old mode enabled still 99%. Yes, hashrate is little lower on some algos with it enabled because more work is done on the GPU, and almost none on the CPU. I exactly know why it's giving more load on the cpu, because upx job round is done in less than 100ms, and a 1,2,4mb scratchpaded job is done in ~2s. Rewriting some stuff to use intrinsics, so cpu load will be lower. But still old mode enabled should decrease your cpu load a lot. |
|
|
|
Hi, guys!
Who can explain what means compute_errors in SRB json reports?
One of my gpus (rx580_hynix8, Win10, 18.6.1) have a one error every 1 min. I reduced an overclocking and intensity, but the errors in srbminer still comes.
HWINFO no have memory errors.
Maybe somebody knows?
I think it is some kind of bug from version 1.8.4 All my polaris cards have 10-20 hw errors per day on heavy algo now according to srbminer. But they are perfectly stable and pool hash is good. All these cards had no hw errors before with previous versions of srbminer and other miner software with the same clock/voltage settings. It's not a bug. You noticed the hw errors, but did not notice the hashrate increase ? https://bitcointalk.org/index.php?topic=3167363.msg50571884#msg50571884- heavy_mode 3 is now the default
- heavy_mode 3 (experimental) is now even faster, but can produce more compute errors than heavy_mode 1
So if you have too much compute errors on heavy derived algos switch to heavy_mode 1 or 2. They are slower tho. But the guy having a compute error every minute isn't related to this. That might be who-knows-what, maybe driver issues. |
|
|
|
Thank for suggesion, but I tried drivers till the latest 19
Situation the same.
I suspect the porblem that I dont have no WinIO64.sys nor WinIO64.dll.
If it is the problem - how I can instal them?
You are lost here I answered you in my thread. ElioVP's tweaker does not use WinIO driver btw. |
|
|
|
Hi Dok and mates!
Cant make tweaking work:(
I get error in log:
Tweaking disabled - WinIO could not be loaded
I suspect the porblem that I dont have no WinIO64.sys nor WinIO64.dll.
If it is the problem - how I can instal them?
Tried different drivers till latest, the same error.
Please, help!
I don't know where you downloaded SRB from but download from https://www.srbminer.com/download.htmlWinIO64.sys is bundled with the miner. After that you need to give miner administrator privileges, without that you can't use tweaking profiles. Here's a how-to video : https://www.youtube.com/watch?v=QyJzAhdDnPg&feature=youtu.be |
|
|
|
"If your Vega64 with Samsung mem. can handle tweak profile 10, then you will get some hashrate increase on it (thanks goes to Rednow, Gkumaran)"
Tried on 2 rigs - both with only samsung memory - on all 24 vega56 with 64 bios cards profile 10 not working - maybe a bug? on 1.8.6 not a problem with profile 5
What do you mean not working, is it crashing the rig, or no change in hash? It does not start, the miner windows closes when gets to enumerate the gpus. Nothing inside logs. That's not related to tweaking profiles then, they get applied after all the gpu's got inited but right before connecting to the pool. |
|
|
|
I am mining Monero coin with the both versions on the same rig . 8.6 and 8.7 . Both are keeping rr my rig after a few minutes ... All gpus in this rig are rx570 4GB . i have another rig with rx580 8G and your miner doesnt restart ... Do you know why i have this issues ? Thanks .
rr = restart ? SRB should work fine with all Polaris cards, so rx580 8g should work too. Maybe you can upload your log file from both rigs somewhere and share, that way i can get more info on what is happening. |
|
|
|
"If your Vega64 with Samsung mem. can handle tweak profile 10, then you will get some hashrate increase on it (thanks goes to Rednow, Gkumaran)"
Tried on 2 rigs - both with only samsung memory - on all 24 vega56 with 64 bios cards profile 10 not working - maybe a bug? on 1.8.6 not a problem with profile 5
What do you mean not working, is it crashing the rig, or no change in hash? |
|
|
|
for new card old mode too? Yes, it utilizes the cpu less, so that will help. |
|
|
|
No cpu mining. The new Upx2 algo is the case.
I just tested it with Mox and the cpu is at 15-20% on the same rig.
Yes it looks like something is killing the cpu with upx2, going to check it out maybe i can do something about it. Great, thanks a lot:) Ok i found a temporary fix/workaround until i fix it. If you use "old_mode" the cpu usage will be normal. Only drawback is that you can't use auto settings if you want to use "old_mode", you need to set up things manually, or in gpu_conf. I use a batch start file for single coins that take values from text files(pools, config and log). Does that mean that i have to use manual values for intensity, fragments,...? So i can use the "old_mode" in my gpu config text file right? If all gpu's use the same settings, you can do a simple as this in config file : {
"cryptonight_type" : "upx2",
"intensity" : 105,
"worksize" : 8,
"double_threads" : true,
"old_mode" : true,
}
Or if your gpu's don't have the same settings you can do : {
"cryptonight_type" : "upx2",
"intensity" : 0,
"double_threads" : true,
"gpu_conf" :
[
{
"id" : 0,
"intensity" : 105,
"worksize" : 8,
"threads" : 2,
"old_mode" : true
},
{
"id" : 1,
"intensity" : 104,
"worksize" : 8,
"threads" : 2,
"old_mode" : true
},
{
"id" : 2,
"intensity" : 106,
"worksize" : 8,
"threads" : 2,
"old_mode" : true
},
{
"id" : 3,
"intensity" : 105,
"worksize" : 8,
"threads" : 2,
"old_mode" : true
}
]
}
|
|
|
|
No cpu mining. The new Upx2 algo is the case.
I just tested it with Mox and the cpu is at 15-20% on the same rig.
Yes it looks like something is killing the cpu with upx2, going to check it out maybe i can do something about it. Great, thanks a lot:) Ok i found a temporary fix/workaround until i fix it. If you use "old_mode" the cpu usage will be normal. Only drawback is that you can't use auto settings if you want to use "old_mode", you need to set up things manually, or in gpu_conf. |
|
|
|
No cpu mining. The new Upx2 algo is the case.
I just tested it with Mox and the cpu is at 15-20% on the same rig.
Yes it looks like something is killing the cpu with upx2, going to check it out maybe i can do something about it. |
|
|
|
I checked the cpu and you are right. It's around 98% all the time.
Any solutions for that?
If you are not doing any cpu mining then it shouldnt be. Which algo? |
|
|
|
V1.8.7- Improvements with tweak profile 10 on Vega64 with Samsung memory - Fixed a bug with pool reconnect on job timeout - job_timeout is now disabled by default, to enable it set a value > 0 - Fixed min_rig_speed bug that triggered restart with V4 algos- Added info in api/web stats about tweak profile - Made hashrate display on web stats page bigger - In guided setup mode changed max wallet and password length to 200 so address+payment id can be used + If your Vega64 with Samsung mem. can handle tweak profile 10, then you will get some hashrate increase on it (thanks goes to Rednow, Gkumaran) + Disabled job_timeout check by default, so if you want to use it set a value that is > 0 + Min_rig_speed *should* now work ok with V4 algos, and shouldn't trigger because of the kernel precompilationI think i pushed 3 new versions in one week, so it's time to slow down. 
|
|
|
|
trouble with old 2gb card on upx2 i=354 but need 36 posible auto setup gone mad, which GPU is that? |
|
|
|
the result vega56samsung(v64bios) srb 1.8.6 slower 5% & higher power usage 6% compared teamred, hope next iteration will better
Srb is 100% faster on Litev7, Tube, Haven, Gpu, Hycon, Mox, Wownero, than teamred. Just because TRM not supporting them But it's still true. I still use more SRB and the day you short those watts i will never use other miner. Thanks doctor P.S.: Statistic for one of the most used algo in last days: 12 GPU rx 570/80 8GB Micron SRB Miner - 97.3kH,1420W power consumption TRM Miner - 92.5kH,1200W power consumption Same settings for both such nice words, thank you cas333  |
|
|
|
|
Show Posts
