Topic: Claymore's CryptoNote AMD GPU Miner v11.3 - page 80. (Read 2145043 times)

To be able for 10.2 to run for a six card miner we need to set the virtual memory @40,000-50,000mb

What hash rate increase are you getting now?

It deos.

No difference, I have 3 cards rig and pagefile was set to 40gb.

v10.2 beta - bug

I also confirm drop of hashrate (~ -100h/s) when monitor goes to sleep or is turned off. Same happens when TeamViewer disconnects. Only solution is to have monitor always on or TeamViewer connected.
VGA: 1x290x (HDMI), 2x280x

In v9.7 everything works normally.

Make sure you have enough virtual memory (around 7500 MB per card).

I am running the 10.2 and it works smoothly (on 10X RX580)

I upgraded from 9.7 to 10.2 and I get the following.  Any help would be great.
C:\Mining\ClaymoreCryptoNoteGPU-XMR>timeout /t 60

Waiting for 45 seconds, press a key to continue ...
19:26:54.82 Tue 11/07/2017

SUCCESS: Specified value was saved.
SUCCESS: Specified value was saved.
SUCCESS: Specified value was saved.
SUCCESS: Specified value was saved.
SUCCESS: Specified value was saved.
+----------------------------------------------------------------+
¦         Claymore's CryptoNote AMD GPU Miner v10.2 Beta         ¦
+----------------------------------------------------------------+

XMR: 1 pool is specified
Main Monero pool is us-east.cryptonight-hub.miningpoolhub.com:20580
At least 16 GB of Virtual Memory is required for multi-GPU systems
Make sure you defined GPU_MAX_ALLOC_PERCENT 100
Be careful with overclocking, use default clocks for first tests
press "s" for current statistics, "0".."9" to turn on/off cards, "r" to reload pools
OpenCL initializing...
AMD Cards available: 10
GPU #0: Hawaii, 8192 MB available, 44 compute units
GPU #0 recognized as Radeon 390X
GPU #1: Ellesmere, 8192 MB available, 36 compute units
GPU #1 recognized as Radeon RX 480/580
GPU #2: Hawaii, 4096 MB available, 40 compute units
GPU #2 recognized as Radeon 290
GPU #3: Hawaii, 4096 MB available, 44 compute units
GPU #3 recognized as Radeon 290X
GPU #4: Hawaii, 4096 MB available, 44 compute units
GPU #4 recognized as Radeon 290X
GPU #5: Hawaii, 4096 MB available, 40 compute units
GPU #5 recognized as Radeon 290
GPU #6: Hawaii, 4096 MB available, 40 compute units
GPU #6 recognized as Radeon 290
GPU #7: Hawaii, 4096 MB available, 40 compute units
GPU #7 recognized as Radeon 290
GPU #8: Hawaii, 4096 MB available, 44 compute units
GPU #8 recognized as Radeon 290X
GPU #9: Ellesmere, 8192 MB available, 36 compute units
GPU #9 recognized as Radeon RX 480/580
POOL version
GPU #0 algorithm ASM, -h 1376, -dmem 1 (Memory used: 5778MB)
GPU #1 algorithm ASM, -h 1800, -dmem 1 (Memory used: 7560MB)
GPU #2 algorithm ASM, -h 1280, -dmem 0 (Memory used: 2688MB)
GPU #3 algorithm ASM, -h 1376, -dmem 0 (Memory used: 2889MB)
GPU #4 algorithm ASM, -h 1376, -dmem 0 (Memory used: 2889MB)
GPU #5 algorithm ASM, -h 1280, -dmem 0 (Memory used: 2688MB)
GPU #6 algorithm ASM, -h 1280, -dmem 0 (Memory used: 2688MB)
GPU #7 algorithm ASM, -h 1280, -dmem 0 (Memory used: 2688MB)
GPU #8 algorithm ASM, -h 1376, -dmem 0 (Memory used: 2889MB)
GPU #9 algorithm ASM, -h 1800, -dmem 1 (Memory used: 7560MB)
Total cards: 10
XMR: Stratum - connecting to 'us-east.cryptonight-hub.miningpoolhub.com' <172.104.76.21> port 20580 (SSL/TLS)
Watchdog enabled
Remote management (READ-ONLY MODE) is enabled on port 3333

SSL/TLS encryption is enabled
XMR: Stratum - Connected (us-east.cryptonight-hub.miningpoolhub.com:20580) (SSL/TLS)
GPU 1, GpuMiner dk2a failed -4
GPU 5, GpuMiner k1a failed -4
GPU 5, GpuMiner k1a failed -4
GPU 0, GpuMiner dk2a failed -4
GPU 0, GpuMiner dk2an0 failed -4
GPU 1, GpuMiner dk2an0 failed -4
GPU 5, GpuMiner k1a failed -4
GPU 0, GpuMiner dk2an0 failed -4
GPU 0, GpuMiner dk2an0 failed -4
GPU 5, GpuMiner k1a failed -4
GPU 1, GpuMiner dk2an0 failed -4
GPU 1, GpuMiner dk2an0 failed -4

It then restarts and gets the same results.

Thanks!

+++++++++++++++++++++++++++
Quick Update.....

Moving from a 25GB swap file to 40GB appears to have fixed the issue.
Your page is great!  Thanks for all the info.

Why 10.2 is so bad, very often crash, no problems with 9.6?

Could someone give me a more detailled explanation about:
-dcri
Decred/Siacoin/Lbry/Pascal intensity, or Ethereum fine-tuning value in ETH-only ASM mode. Default value is 30, you can adjust this value to get the best Decred/Siacoin/Lbry mining speed without reducing Ethereum mining speed.
   You can also specify values for every card, for example "-dcri 30,100,50".
   You can change the intensity in runtime with "+" and "-" keys and check current statistics with "s" key.
   For example, by default (-dcri 30) 390 card shows 29MH/s for Ethereum and 440MH/s for Decred. Setting -dcri 70 causes 24MH/s for Ethereum and 850MH/s for Decred.
-asm
(AMD cards only) enables assembler GPU kernels. In this mode some tuning is required even in ETH-only mode, use "-dcri" option or or "+/-" keys in runtime to set best speed.
   Currently ETH-LBRY mode is not supported in assembler.
   Specify "-asm 0" to disable this option. You can also specify values for every card, for example "-asm 0,1,0". Default value is "1".
   If ASM mode is enabled, miner must show "GPU #x: algorithm ASM" at startup.
   Check "FINE-TUNING" section below for additional notes.
   NEW: added alternative assembler kernels for Tonga and Polaris cards for ETH-only mode. Use them if you get best speed at "-dcri 1" (i.e. you cannot find speed peak), use "-asm 2" option to enable this mode.

Do I need to set both -asm and -dcri? When I set these, how do they relate to overclocking and undervolting with -cclock, -mclock, -cvddc and -mvddc ?

What is your gpu's settings?

Set the virtualmemory to 64000 - I had the same problem with my rig of 580s at it was solved like this.

Thank you!

I have 10 X RX580 and after increasing the Virtual Memory to 48000 I still got the errors. After I increased to 64000 it worked smoothly (it may also depend on how many cards you have).

Let me know when you get the RX580 8GB stable. Really want to run 10.2.   

Im running 9.7.    I noticed when it fires up it sets -Hashcnt 1024 for all cards (this is for RX580 8GB, core 1170 mem 2100 1000mv Samsung memory straps copied.)

I am going to test this number to see what it makes a difference in hash. Im assuming ill just get same speeds with 10.2 as 9.7.

Thanks for all the info btw. awesome stuff

Use 9.7 version, same 270x 1100/1500, 509H/s

so do i, my 580 x5 need almost 40GB pagefile (38,xxxxx)

i think i am not alone getting "WATCHDOG: GPU x hangs in OpenCL call, exit" message. Restart in every few hours or sometimes every minutes.. but mostly happened in gpu 0/1 (with long type pci-ex slot/pci-ex x16), maybe i'll try to only use x1 type pci-ex..

Just wanted to update. I tried lowering the -h value on my rigs. I have three different setups:

6x RX570 4G - these are at 1200/1750/1000mV -h 850 -> These have been stable for last 24h, no resets

11x RX570 4G - these are at 1200/1750/1000mV -h 850 -> These have reset 2-4 times over last 24h - Going to test -h 800 to see if I can get better stability

11x RX580 8G - these are at 1200/2000/1000mV -h 1700 -> These have reset 1-2 times over last 24h - Going to test -h 1600 to see if I can get better stability

For the RX570 4G cards, I'm hashing at 644, for the RX580 8G cards, I'm hashing at 733. I'm not sure why the RX570 4G have been stable but the RX580 8G have not. Hopefully this next test shows a better result.

I'm not sure if there's a way to increase the virtual memory on SMOS. My rigs use 8G of physical RAM (tho I just realized that both my 6x and 11x rigs both are using 8G of RAM so maybe I should increase my 11x rigs to 16G of RAM). I'm not using SSDs either, SMOS is being booted from a 16G 3.0 USB drive. I'm not sure if 16G of physical RAM would increase stability as I used to mine Ethereum with these rigs and they have been stable with far more aggressive settings. The RX570 4G were at 1150/1750/950mV and the RX580 8G were at 1150/2100/950mV and the systems were stable for 100+ hours with no resets.

Cheers,

That I don't see anymore. H/s 450-480 looks low

My 470 8GB x6 rigs need at least 48GB pagefile to mine at optimal speed (~940H/s).  Eth I would just do 16GB. I use all 120GB ssds so it's no big deal.

^^^ Do you see that message often? Have you tired lowering your -h value?

Optimal settings for a AMD Sapphire Dual-X R9 270X?

Core Clock oc @ 1100MHz
Mem Clock oc @ 1500MHz



Windows 10 64bit

Catalyst 15.12

Right now using Claymore's 10.2 default settings.

Any help greatly appreciated, thanks.