DEAR DEVELOPER PLEASE STATE OFFICIALLY IN DESCRIPTION TO MINER WHICH SERIES OF CARDS HE SUPPORTS
It SUPPORTS every GCN card, if it works then it is supported, right? Don't be nervous, chill.
|
|
|
You should use that other miner that works on those old cards. I can't support and develop for cards i don't have, sorry.
|
|
|
1.7.2 best!!!!! heavy algo
rx570 8g card old version 1060 unstable
1.7.2 1130 stable 1150\2000-2030
[2018-12-04 12:18:35] GPU0 : 1130 H/s [T: 51c, RPM: 1502, CC: 1150 MHz, MC: 2030 MHz][BUS:1] [2018-12-04 12:18:35] GPU1 : 1130 H/s [T: 55c, RPM: 1780, CC: 1150 MHz, MC: 2030 MHz][BUS:2] [2018-12-04 12:18:35] GPU2 : 1128 H/s [T: 58c, RPM: 190, CC: 1150 MHz, MC: 2030 MHz][BUS:3] [2018-12-04 12:18:35] GPU3 : 1074 H/s [T: 53c, RPM: 2250, CC: 1150 MHz, MC: 1900 MHz][BUS:4] [2018-12-04 12:18:35] GPU4 : 1131 H/s [T: 53c, RPM: 1671, CC: 1150 MHz, MC: 2030 MHz][BUS:6] [2018-12-04 12:18:35] GPU5 : 1130 H/s [T: 55c, RPM: 1815, CC: 1150 MHz, MC: 2030 MHz][BUS:7] [2018-12-04 12:18:35] GPU6 : 1132 H/s [T: 52c, RPM: 1576, CC: 1150 MHz, MC: 2030 MHz][BUS:8] [2018-12-04 12:18:35] GPU7 : 1172 H/s [T: 58c, RPM: 2591, CC: 1150 MHz, MC: 2000 MHz][BUS:9]
570 4bg also add rate on heavy +30-60 per card
now try heavy_mode 3
|
|
|
V1.7.2
- Ellesmere 8Gb cards can now reach the full hashing speed on heavy algos - Rx550 and Rx560 should now be faster on heavy algos by ~8%, and ~5% on V7 algos - Rewritten a lot of stuff so kernel can now be compiled on newer video drivers too, so no binary kernels provided anymore - Much faster kernel compilation time - Removed bralock, aes and kernel parameters - Added value 9997 and 9998 for fragments which could improve stability on some cards - heavy_mode 3 (experimental) is now even faster, but still can produce more compute errors than heavy_mode 1, so 1 is still the default
+ Good news, finally got that hashrate stable for the 8GB ellesmere cards on heavy algos (rx470, rx480, rx570, rx580)
+ Miner can now compile working binaries on newer video drivers too, but i don't know about the speeds, i only tested on the latest AMD driver and it is slower than 18.6.1 Maybe there is a driver version that is faster, if someone finds it please share
+ Kernel compilation times are now reduced by a few times
+ Bralock, Aes and Kernel parameters removed, as they are not needed anymore
+ You can now use values 9997 and 9998 for fragments. If you have hashrate instability issues, try them.
+ If you want to see even more speed on heavy algos use heavy_mode 3, but be prepared to get more compute errors than on 1. Test it and see if the hashrate / compute errors ratio satisfies you
|
|
|
Ok, so looks like i finally resolved the ellesmere 8gb heavy algos thing. Now they easily hash over 1.1khs.
New version in a few days
|
|
|
Da je to tako lako, vec bi odavno uradio
|
|
|
for mining stellite I stopped at version 1.6.4 the hashrate is somewhat lower, but without failures and without errors
ok, in global which version worked the best on those R7 370 pitcairns ? Im asking because i can try to replace kernel 2 with that old one you say worked the best for you, and add all new algos without trying to mess up too much. Also do you get compute errors on new SRB version on V7 too ? Because V8 can make more compute errors, it's more 'sensitive' to overclocking and the strap you are using. If i remember good you had a lot of compute errors even on 1.6.4
|
|
|
heavy algo (haven) hashing at 1800HR superb doing V8 algo i get like 1300 to 1400HR
what's up?
something on your side is up tried multiple rigs I: 56 (heavy loves this) WS: 8 --resetvega --ccryptonighttype normalv8 --cgpuid 0,1,2,3,4 --cgpuworksize 8,8,8,8,8 --cgpuintensity 56,56,56,56,56 --cgputhreads 2,2,2,2,2 --sendallstales --enableduplicategpuid --gpureorder this is extremly good for heavy, but what is going on V8, if i switch back to heavy everything goes like charm.. Intensity 56 is too low for V8, if im right you have a vega, so set it to : intensity : 112 worksize: 16
|
|
|
heavy algo (haven) hashing at 1800HR superb doing V8 algo i get like 1300 to 1400HR
what's up?
something on your side is up
|
|
|
[2018-11-24 13:10:06] hashrate: GPU0: 263 H/s [BUS:8] [2018-11-24 13:10:06] hashrate: GPU1: 262 H/s [BUS:10] [2018-11-24 13:10:06] hashrate: GPU2: 438 H/s [BUS:4] [2018-11-24 13:10:06] ADL2_Overdrive5_FanSpeed_Get for ADL index 79 failed. [2018-11-24 13:10:06] hashrate: GPU3: 238 H/s [BUS:11] [2018-11-24 13:10:06] hashrate: GPU4: 429 H/s [BUS:1] [2018-11-24 13:10:06] hashrate: GPU5: 569 H/s [BUS:9] [2018-11-24 13:10:06] hashrate: Total: 2199 H/s
Do you get the same results with kernel 2? which SRB version was best for your pitcairns?
|
|
|
stellite v4 srb 1.6.9 --- 6 cards
[2018-11-24 11:02:01] GPU0: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 1][K: 1][BUS: 8] [2018-11-24 11:02:01] GPU1: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 1][K: 1][BUS: 10] [2018-11-24 11:02:01] GPU2: Radeon(TM) RX 460 Graphics [baffin] [2048 MB][I: 26.0][W: 8][T: 2][F: 1][K: 1][BUS: 4] [2018-11-24 11:02:01] GPU3: AMD Radeon R7 200 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 1][K: 1][BUS: 11] [2018-11-24 11:02:01] GPU4: Radeon(TM) RX 460 Graphics [baffin] [2048 MB][I: 14.0][W: 8][T: 2][F: 1][K: 1][BUS: 1] [2018-11-24 11:02:01] GPU5: AMD Radeon R9 200 Series [tonga] [2048 MB][I: 28.0][W: 8][T: 2][F: 16][K: 1][BUS: 9]
тeпepь вepcия srb 1.7.1
[2018-11-24 11:07:56] GPU0: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 0][BUS: 8] [2018-11-24 11:07:56] GPU1: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 0][BUS: 10] [2018-11-24 11:07:56] GPU2: Radeon(TM) RX 460 Graphics [baffin] [2048 MB][I: 26.0][W: 8][T: 2][F: 1][BUS: 4] [2018-11-24 11:07:56] GPU3: AMD Radeon R7 200 Series [pitcairn] [2048 MB][I: 28.0][W: 8][T: 2][F: 0][BUS: 11] [2018-11-24 11:07:56] GPU4: Radeon(TM) RX 460 Graphics [baffin] [2048 MB][I: 14.0][W: 8][T: 2][F: 1][BUS: 1] [2018-11-24 11:07:56] GPU5: AMD Radeon R9 200 Series [tonga] [2048 MB][I: 28.0][W: 8][T: 2][F: 16][BUS: 9]
I agree the settings for each version are different and can change the results but not so much, 2 times less this rig is my solyanka assembled from different series of cards, old and new
Look at the fragments in 1.6.9 and 1.7.1. Try using these in 1.7.1 : GPU0 : F1 GPU1 : F1 GPU2 : F1 GPU3 : F1 GPU4 : F1 GPU5 : F16 Set in gpu_conf, you can see on the first page how. Then come back here and tell me is it better.
|
|
|
test lot of cards on tube
rx470\480\570\580 4gb +50-60 h\s in heavy 3 mode 1.7.1 820-860 h\s
and old trouble with 8gb 470\480\570\580 card. test with different config. -70-80 h\s back to 1.6.7
test 560 4 gb card 1.6.9 508 h\s 1.7.1 440 h\s
560 2 gb 1.7.0 380 h\s 1.7.1 406 h\s
550 2 gb and 4 gb same rate
For 560 4g try different settings for fragments, try 0,1,2,4,8, 9999 one of these should bring back the hash
|
|
|
Freehaven is about 10-12% slower than xmrig its not if you configure it yourself and not use the auto settings
|
|
|
STELLITE V4 [2018-11-23 22:11:17] Miner version: 1.7.1 [2018-11-23 22:11:17] Windows version: 10.0 build 16299 [2018-11-23 22:11:17] Video driver version: 24.20.11021.1000 [2018-11-23 22:11:17] Startup monitor attached [2018-11-23 22:11:25] AMD Platform ID: 0 [2018-11-23 22:11:25] AMD platform FOUND [2018-11-23 22:11:25] Found 6 AMD devices [2018-11-23 22:11:25] CPU AES-NI: FALSE [2018-11-23 22:11:25] GPU0: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 25.0][W: 8][T: 2][F: 0][BUS: 8] [2018-11-23 22:11:25] GPU1: AMD Radeon (TM) R7 370 Series [pitcairn] [2048 MB][I: 25.0][W: 8][T: 2][F: 0][BUS: 10] [2018-11-23 22:11:25] GPU2: AMD Radeon R7 200 Series [pitcairn] [2048 MB][I: 25.0][W: 8][T: 2][F: 0][BUS: 11] [2018-11-23 22:11:25] ADL is enabled [2018-11-23 22:11:25] Cryptonight StelliteV4 mode enabled [2018-11-23 22:11:25] Starting init of mining threads [2018-11-23 22:11:25] Created OCL context [2018-11-23 22:11:25] Using fragments 0 for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Using aes 0 for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Using bralock 8 for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Using heavy_mode 1 for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Using kernel 2 for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Created OCL command queue for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Created OCL input buffer for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Created OCL output buffer for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Loading stellitev4 kernel for DEVICE BUS_ID[8] ... [2018-11-23 22:11:25] Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx->Program] for DeviceID 0 (Thread 0) [2018-11-23 22:11:25] Error initing GPU's. Stopping miner process [2018-11-23 22:11:25] Stopping miner process
I could create a binary for pitcairn for kernel1 without problems. (stellite) Here it is, try it : https://ufile.io/33ldwCompiled with your settings : Worksize : 8 Fragments : 0 You must set exact settings so the miner uses the binary i linked. I am more curious why it reports 6 GPU's. You are not using every GPU in the machine in the config file?
|
|
|
Having tested version 1.7.1 I will say with confidence for old pitcairn cards benchmark shows a huge drop on heavy, normalv7, normalv8, stellitev4 test interrupted by message Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx-> Program] for DeviceID 0 (Thread 0) [2018-11-23 17:25:04] GPU's Error initing. Stopping miner process [2018-11-23 17:25:04] Stopping miner process transition to kernel 2 did not give results I am completely disappointed with the miner great job
ok, ok, you remember i connected a few times with teamviewer trying to help you even before v8. you don't need to be sarcastic every time. Better give me more info, something that can help me, so i can help you. Upload part of log file where the build fails. Also i want to know what hashrates you get on which card/algo. I ran an offline compile of kernels and basically it could compile on every GCN, that means pitcairn also.
|
|
|
And now tell me do i need to fix anything, or not change anything ??
|
|
|
Hey Dok.
Thanks for the constant updates on the miner.
Tried the last one and it seems I've lost 150 hs on each vega 56 on haven compared to the last one.
My config is as simple as possible. I did try to play with the values but it doesn't affect it. The miner starts above 6700 for 4 vega, then drops slowly to 6300 / 6200. Does seem watts are down a bit...
Any tips?
Thanks
If you don't really want to play try : "intensity" : 61, "double_threads" : true, "worksize" : 9, "fragments" : 1, "aes" : 0, "bralock" : 32, "heavy_mode" : 2
|
|
|
Having tested version 1.7.1 I will say with confidence for old pitcairn cards benchmark shows a huge drop on heavy, normalv7, normalv8, stellitev4 test interrupted by message Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx-> Program] for DeviceID 0 (Thread 0) [2018-11-23 17:25:04] GPU's Error initing. Stopping miner process [2018-11-23 17:25:04] Stopping miner process transition to kernel 2 did not give results I am completely disappointed with the miner great job
you got 370 2gb right?
|
|
|
Auto setup for freehaven is sh*t, i forgot to tune it, so it uses settings like for heaven, which is obviously bad. For Vega try intensity 110, for other i dunno
|
|
|
|