Show Posts
|
Pages: [1] 2 3 »
|
Ok I got some help and found out there was an issue with HiveOS and Vegas using TRM and amdmemtweak. It got fixed on the back end for me and supposedly goes out in an update soon. Changed hash rate significantly so it's much more in line with expectations. Figured I'd give my ending values in case they help anyone else.
amdmemtweak –CL 20 --RAS 26 --RCDRD 12 --RCDWR 12 --RC 38 --RP 12 --RRDS 3 --RRDL 5 --RTP 6 --FAW 12 --CWL 8 --WTRS 4 --WTRL 9 --WR 13 --REF 12000 --RFC 239
I'm at 46 Mh/s on my RX Vega 56 Samsung with stock bios and down clocked 940mhz core, 850mv, 950mhz mem. Pulling 1020 watts from the wall with a 6 GPU rig.
|
|
|
I just started using HiveOS and I have a couple rigs of RX Vega 56. I used PhoenixMiner at first and OC would take effect but not good hashrates. I switched to TRM and OC wouldn't take effect at all and only about 38Mh/s. admmemtweaks would work via shell but still only upto 43Mh/s. Now hoping it was related to the version I downloaded. Also didn't know RX Vegas didn't need the specific version for them and seems they can use the latest stable image. So I will have to pull and flash the drive and try again.
|
|
|
Can anyone advise what line I should be using for the following?
root@Rig1:~# amdmemtweak --current GPU 0: Vega 10 XL/XT [Radeon RX Vega 56/64] pci:0000:06:00.0 Memory state: 800MHz Memory: Samsung HBM2 Memory vendor and model 61AB1A9C Timing 1: CL: 17 RAS: 27 RCDRD: 13 RCDWR: 8 Timing 2: RCAb (RC): 39 RCPb (RC): 39 RPAb (RP): 12 RPPb (RP): 12 Timing 3: RRDS: 4 RRDL: 5 RTP: 5 Timing 4: FAW: 13 Timing 5: CWL: 6 WTRS: 3 WTRL: 7 Timing 6: WR: 13 Timing 7: RREFD: 7 Timing 8: RDRDDD: 3 RDRDSD: 2 RDRDSC: 1 RDRDSCL: 3 Timing 9: WRWRDD: 3 WRWRSD: 2 WRWRSC: 1 WRWRSCL: 3 Timing 10: WRRD: 1 RDWR: 17 Timing 12: REF: 3120 Timing 13: MRD: 8 MOD: 12 Timing 14: XS: 216 Timing 16: XSMRS: 6 Timing 17: PD: 6 CKSRE: 8 CKSRX: 8 Timing 20: RFCPB: 128 STAG: 7 Timing 21: XP: 6 CPDED: 2 CKE: 6 Timing 22: RDDATA: 16 WRLAT: 4 RDLAT: 14 WRDATA: 1 Timing 23: CKESTAG: 2 RFC Timing: RFC: 208
So far the best I've attained is 43.5 Mh/s using: amdmemtweak --CL 20 --RAS 20 --RCDRD 12 --RCDWR 12 --RC 37 --RP 11 --RRDS 3 --RRDL 6 --RTP 5 --FAW 12 --CWL 8 --WTRS 4 --WTRL 9 --WR 14 --REF 65535 --RFC 248
BIOS is stock. I'm on HiveOS. Just running the line above through the shell. Not really sure what I'm doing yet. Also the core and memory clocks aren't seeming to be changed even though I specify it in HiveOS. So perhaps it is the lack of overclock and not the timings that is causing it to have low hashrate???
|
|
|
@greatauror28 - We appreciate your enthusiastic replies on clock speeds and voltage but can you please post your BIOS too. That makes all the difference. Thanks in advance.
Sorry but I did all the work myself without heeding any help from someone. There's tons of information all over the internet and you need very basic skills to get the information what you need if you know where to look and how to look. A great man once said "if you're good at something, never do it for free". I guess that's clear enough. Unreal. Tons of people here share stuff to help each other out. Nobody knows everything. The whole point of the board is to come together and be better than we would be individually. If that isn’t your philosophy then why are you here? Your own statements in the same post are mutually exclusive too. You say you did it all yourself. You also say the info is all over the internet. Did you write your own BIOS editor? Did you write your own overclocking and hardware monitoring tools? Did you invent how to modify timings or create the base BIOS you started with? Do as you please but don’t expect people to be impressed with you here when you are boasting of your own “success”with nothing to provide to the greater community. If the developers of all the cryptocoin were that way none of us would have any of it. As a developer myself I have a great deal of respect for the elite ones working in this space. The very best are also a great deal more humble than you seem to be. Maybe you’ll reconsider but I certainly won’t beg you or bother anymore.
|
|
|
@greatauror28 - We appreciate your enthusiastic replies on clock speeds and voltage but can you please post your BIOS too. That makes all the difference. Thanks in advance.
|
|
|
So you're having issues mining? It sounds like the system banned you for some reason. Email me with your ip address and I'll fix it. Or contact me on Telegram with it.
Pm sent
|
|
|
All afternoon I’ve had connection issues. Socket closed remotely by pool.
|
|
|
That's been an issue for iPhone users. We have an android app on the google play store and and ios app coming out very soon, that should help with the issue.
What’s the problem exactly? I didn’t look into it further.
|
|
|
Speaking of connections, I noticed I can never get to the pool when on AT&T cellular network. Would be nice to be able to monitor workers on the pool and their status via cell phone data plan. Works when I jump onto Wi-Fi. DNS???
|
|
|
Are you using Sapphire Nitro+ Rx580 4gb with elpida memory as well?
If yes, send me a PM we'll work something out.
Yes. Pm sent.
|
|
|
@greatauror28 - What BIOS are you using or BIOS tweaks? What voltage and clocks?
Modded bios, custom timing strap 1150/2025/875 30.3Mh eth 750 dcr @ 145w at the wall Can you share your BIOS?
|
|
|
No connection issues here
|
|
|
@greatauror28 - What BIOS are you using or BIOS tweaks? What voltage and clocks?
|
|
|
Current Round Variance: 102%. The miner with 170 GH dropped down to a minuscule 60 MH sometime between the 30%-45% variance (wasn't watching exactly when he dropped). So he technically has 59% of round share right now and also has technically continued mining the whole block. Did this miner find the loop hole and is he circumventing the "tax"? Not complaining. Just observing and being transparent. I think miners on the pool would like to know.
|
|
|
Well we finally got over the hump but a 150-170 GH/s single miner is now tearing it up. Nothing like a single miner adding 8x the hashrate of the whole pool. Curious the payouts don't reflect a tax. Is it because that miner stayed past 30%? If so just curious why 30% was chosen as the cut off point if the white paper on the subject matter says 43.5% is the magic number?
|
|
|
MSI X370 Gaming Plus Above 4g decoding enabled
What am I doing wrong?
Try without the 4g decoding enabled. Especially since you don't have 6 or 8GB gpus. I had nothing but trouble with this. Mine won't even post with that option enabled. I'm using the similar MSI X370 GAMING PRO CARBON AM4 AMD X370 and have multiple of these.
|
|
|
Quantity 6 AMD RX Vega 56 confirmed working on ASRock X370 Killer SLI/ac AM4 AMD Promontory X370. More trouble than other motherboards I listed before but nothing terrible. Had to DDU/reinstall drivers a few times before it took with all 6 GPUs on my base image SSD. Worked once I "express" installed the AMD driver back on top of itself again after an initial DDU and re-install. I've frequently had troubles with Asrock boards in the past (bad voltage regulators in old ones) but don't want to assume it more than a fluke. We all know how AMD drivers can be. Working fine for a couple days now. Nice to have wifi onboard if needed and the price is right for the features.
|
|
|
Seriously bad luck this past week but not bailing yet and even throwing a bit more at it. Glad all those other > 1 GH/s miners who jumped off real early should be getting taxed.
|
|
|
The pool hoppers I am talking about are the people who jump on the instant we hit a block with hash power equal the the amount of the rest of the pool combined, stay till 30% variance and leave. They always leave at exactly 30% and always jump on as soon as a block is hit.
It doesn’t take long watching this pool to see what noobpool said is indeed true. In the past couple weeks I’ve watched it closely. I’ve even seen a single miner jump on with greater hash rate than everyone else on the pool combined. The hash rate at least doubles or triples immediately after a solved block. Also as he said they are gone at 30% even though statistically they could stay as long as 43.5%. But they leave beforehand so they can do the same things to more other pools. Examine the block chain and see how many pools they are getting funds from. However, many obfuscate further using multiple wallets I’m sure. I’m glad to see noobpool’s oversight. It’s the only chance PPS has to function as intended. It’s a good thing to decentralize some hash rate to a well run smaller pool. No complaints by me so far. Hopefully stays that way.
|
|
|
|