Hi all  I would like to get some feedbacks (if there are) of data system cooling engineers or pro miners that already thought or created a small GPU mine in a container ? I've looked at the Envion one ( https://www.facebook.com/envion.org/videos/envion-mobile-mining-operations/1930469803947689/). They have a clever passive cooling solution for asics as they plug tubes and pipes directly at the exhaust of each asic but they do no show their plans for GPU mining rigs. for 200 to 250 GPUs I was thinking of 3 cased fans at the bottom 50cm or 60cm (20 or 24 inches) each on one side of the container that would drag the outside air inside and 3 others at the roof that would suck the air out. Note that container might be in a sheltered place that have temperatures arrround 5 to 10 degrees C (41 to 50 F) . What is your opinion ? What would you do to cool such a massive rig ? Also what about the air condensation ? The cold is not a problem at all (Confirmed at temps above -27C outside). Except for you plan to place the container at far north. To keep GPU temps below 70C the incoming air temp should not exceed 35C under the corresponding air flow. There are enough information about how to calculate the airflow. It's generally depends on total rig heat producing. About air condensation: The condensate can appear on ONLY COLD surfaces. GPUs / ASICs are HOT when working Yeah, I dunno why condensation comes up so much. Cold surface, cools the hot, moist air, precipitation formeth. Hot chipset, no moisture. |
|
|
|
|
Win10 updates are a complete lottery - disable them via policy manager and only apply them when forced to. If it ain't broke...
|
|
|
|
Well my Vegas are running @ ~900Khs lool
And when I run cast xmr (last version) it says Gfx900(?) , can anyone confirm this ? it was supposed to say Gfx901(?)
A Vega lists as gfx901 when HBCC is on gfx900 when HBCC is off Hi dear glph3k! I find part answer in your previous post, but i have problem 6-gpu Vega64 in OLD mode (CN V7) work very good 12kHs stability, after update to cast xmr 1.5.0 and driver (this moment i use 18.5.2) two GPUs in my log example №4,5 not hashing... and i observe after command switch-radeon-gpu -G 0,1,2,3,4,5 --hbcc=on autorestart i have it: GPU0-3 it is ok gfx901 but GPU4,5 gfx900 --??? i confused [09:54:39] Initializing GPU, loading kernel ...
Detected OpenCL Platform: OpenCL 2.1 AMD-APP (2580.6)
Fast job switching mode enabled.
Driver Version OK.
GPU0: Radeon RX Vega (gfx901) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
HBCC Memory Segment is enabled.
It is recommended to turn HBCC off to run with higher stable intensity settings.
GPU1: Radeon RX Vega (gfx901) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
HBCC Memory Segment is enabled.
It is recommended to turn HBCC off to run with higher stable intensity settings.
GPU2: Radeon RX Vega (gfx901) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
HBCC Memory Segment is enabled.
It is recommended to turn HBCC off to run with higher stable intensity settings.
GPU3: Radeon RX Vega (gfx901) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
HBCC Memory Segment is enabled.
It is recommended to turn HBCC off to run with higher stable intensity settings.
GPU4: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU5: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10 i think that GPU 4,5 HBCC Memory Segment not switch, it is bug (soft, driver..e.t.c) or my two GPU (HBCC Memory) defective? half all hashrate [Hash Rate Avg: 4766.1 H/s]
922.9 H/s GPU0
913.3 H/s GPU1
893.3 H/s GPU2
879.3 H/s GPU3
730.9 H/s GPU4
426.4 H/s GPU5
[Shares Found: 233 | Avg Search Time: 14.2 sec]
68 ( 29%) Accepted
19 ( 8%) Rejected by pool
0 ( 0%) Invalid result computation failed
0 ( 0%) Could not be submitted because of network error
146 ( 63%) Outdated because of job change after command switch-radeon-gpu -G 0,1,2,3,4,5 --hbcc=off autorestart i have all gfx900, but GPU4,5 not hashing((( [10:15:10] Initializing GPU, loading kernel ...
Detected OpenCL Platform: OpenCL 2.1 AMD-APP (2580.6)
Fast job switching mode enabled.
Driver Version OK.
GPU0: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU1: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU2: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU3: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU4: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
GPU5: Radeon RX Vega (gfx900) | 64 Compute Units | Memory (MB): 8176 | Intensity: 9 / 10
[10:15:14] Connecting to pool.supportxmr.com:7777 ... [Pool: 'pool.supportxmr.com:7777' | Connected: 2018-10-25 10:15:14]
0:02:38 (100%) Online
0:00:00 ( 0%) Offline
[Job: #3 | Difficulty: 185704 | Running: 87.0 sec | Avg Job Time: 35.0 sec]
[Hash Rate Avg: 7912.9 H/s]
1853.2 H/s GPU0
1871.1 H/s GPU1
1868.7 H/s GPU2
1873.9 H/s GPU3
227.9 H/s GPU4
218.2 H/s GPU5
[Shares Found: 16 | Avg Search Time: 11.3 sec]
8 ( 50%) Accepted
0 ( 0%) Rejected by pool
0 ( 0%) Invalid result computation failed
0 ( 0%) Could not be submitted because of network error
8 ( 50%) Outdated because of job change GPU0: Radeon RX Vega (gfx900) ... no changes.
GPU1: Radeon RX Vega (gfx900) ... no changes.
GPU2: Radeon RX Vega (gfx900) ... no changes.
GPU3: Radeon RX Vega (gfx900) ... no changes.
GPU4: Radeon RX Vega (gfx900) ... no changes.
GPU5: Radeon RX Vega (gfx900) ... no changes.
GPU0: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF
GPU1: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF
GPU2: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF
GPU3: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF
GPU4: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF
GPU5: Radeon RX Vega (gfx900) | HBCC: OFF | Large Pages: OFF... Plz help me to solve this issue, thank very much! Me too. I gave up in the end, rolled back to blockchain drivers, mining other CN variants. Worth pointing out that on blockchain drivers the problem vanishes. Tried SRB miner too, same issue. I assume it's a driver bug (thanks AMD). |
|
|
|
|
Hullo folks, is there a powershell script or similar to automate the switch-radeon-gpu commands/overdriveN and run Cast? Many thanks in advance!
|
|
|
|
SO,my BCU1525 Board has been shipped.
It will arrive with SQRL mining software ?
where I can find other software that can be used besides Zetheron ?
DC has a range on fpga.guide too. There's a setup guide for the official BCU shell on all-mine.co, which currently has Lyra2Z, more imminent (according to the allmine folk). These are three different setups though, so I'd advise doing your research first... |
|
|
|
Mining Monero seems to be a PITA at the moment.
I cannot make myself bothering to troubleshoot and make this software works properly with the new fork and cannot bother to move to SBR miner or wtf is called.
I was doing a lot of tweaking and testing in my early days of mining, about 10/12 months ago, when the gig was much more profitable than these days.
Also considering the 7% performance hit you take moving to the new fork, IMO mining Monero does not make a lot of sense.
So, hopefully our Dev is gong to make, via an update, the transition to v.8 a little more straight forward than it actually is.
In the mean time, I moved to other coins that are slightly more profitable and liquid enough in exchanges. A couple of them may also have more upside potential than XMR.
I am just curious if other Cast users are doing the same.
Cheers.
There was a difficulty drop initially, but that's almost vanished now, and the new drivers have wrought havoc with my rigs. Deffo one for a rollback and some kind of CN-Heavy action, methinks. |
|
|
|
The original thread starter does not come to this thread very often because of cancer filled comments. Want something answered? ask it on discord. Too many trolls and toxic people here calling it a scam, why should he have to waste his time arguing with people like you?
Discord is a Circle Jerk where you get banned for not touting the party line. Freedom of speech comes with these drawbacks. Welcome to a place you can speak your mind and you won't be banned for it. I've been on their discord channel for quite some time, and the only folks I saw get banned were the jerks who added nothing but ad hominem attacks and insults to anyone and everyone they could. Phil asked essentially the same hard questions he asked here and got some answers but not banned. He surely was not following the party line. That is not true. I was banned for wanting to sell my 8x 1070ti rig for 1 fpga 1525. I thought it was a fair trade due to the high demand in refunds and I was willing to trade their risk. They got insulted and banned me. That is unacceptable and I expect more professional behavior if selling expensive products. That's not how you treat your future customers that want to buy in Did you think you were communicating with some car dealership and SQRL was going set you up with a good finance deal on a FPGA for your trade-in? Come on now, you were asking for a whoopin with that approach.  A car dealership equivalent would be trying to straight swap a used midrange Ford for a brand new Porsche. You'd have to be stark staring insane. |
|
|
|
Restored to blockchain drivers.
Managed to get pp tables and everything working in short order to restore things as they were. So, I haven't lost my touch yet. It's AMD who sucks... lol
Switching to ETH for now.
Enjoy my absence of hash power !
I might have to do that. two rigs now misbehaving, hashrate drops every hour or so, still missing two cards on one. Complete mess. Those beta blockchain drivers were at least stable. |
|
|
|
Me too. Driving me bonkers. |
|
|
|
Hashrate on cnv8 has cratered. Vega 64s/blockchain drivers/HBCC enabled. intensity 7. hashing at under 1k...
This is a known problem with the blockchain drivers. Update to 18.5.1, they are more stable and don't need the hardware reset trick and will fix performance with V8. HBCC should be turned off. Vega64 can go with intensity 9 if used headless (no desktop running on it), Hmmm. That seems to have worked for all but two of the gpus, which are still hashing in the low hundreds. I did DDU, and device manager says all vegas are using hte same driver... Like a bad nightmare, 'disable crossfire, disable upls' has floated into my head... Which hasn't worked. Bastard Vegas. Anybody else having a similar issue? 5card vega rig, updated drivers, 3 cards running as standard, 2 running at 120 and 300h/s. DDUed several times, crossfire etc disabled, HBCC on or off makes no difference. intensity set to 7... Arrrgh. The solution is to run a dummy plug. When I have a monitor plugged in I have the exact same issue. Put in an hdmi dummy plug. reboot, and its back to normal Fraid not, I run with a dummy plug as standard. Is this a known issue with 18.5.1? Someone else must have experienced this. Overdrive seems to be working, as both the low-hashing cards have fans set high. It's like some invisible crossfire setting...grrrrr... |
|
|
|
Hey guys, how's your v8 performance?
On Polarises (both 8GB and 4GB) I see a drop in hashrate of roughly 10%
However Vegas is a diffident story, it seems to be somewhat inconsistent.
My optimal v7 threads intensities were 2024 and 1968 and while it seems to be working fine on a single Vega56(18.6.1 driver), producing about 20% lower hash, just above 1600...
It gives me trouble on a rig of mixed ones (FE, 64, 56) it will still run on those high intensities, but the GPU would produce around 1000h/s
And after I lower it all to about 1800&1200 some of the GPUs(one 56 and one 64) would not go above 1000h/s, while the rest would hash at around 1300.
Any ideas?
Just built 2.5.1, yet it seems to make no difference.
Not sure if this is a related issue, but I'm getting a similar issue (one 64 and one 56 in a rig will not go above a few hundred h/s). Tried all the usual things, kinda stumped now. Running Cast. |
|
|
|
Hashrate on cnv8 has cratered. Vega 64s/blockchain drivers/HBCC enabled. intensity 7. hashing at under 1k...
This is a known problem with the blockchain drivers. Update to 18.5.1, they are more stable and don't need the hardware reset trick and will fix performance with V8. HBCC should be turned off. Vega64 can go with intensity 9 if used headless (no desktop running on it), Hmmm. That seems to have worked for all but two of the gpus, which are still hashing in the low hundreds. I did DDU, and device manager says all vegas are using hte same driver... Like a bad nightmare, 'disable crossfire, disable upls' has floated into my head... Which hasn't worked. Bastard Vegas. Anybody else having a similar issue? 5card vega rig, updated drivers, 3 cards running as standard, 2 running at 120 and 300h/s. DDUed several times, crossfire etc disabled, HBCC on or off makes no difference. intensity set to 7... Arrrgh. |
|
|
|
Hashrate on cnv8 has cratered. Vega 64s/blockchain drivers/HBCC enabled. intensity 7. hashing at under 1k...
This is a known problem with the blockchain drivers. Update to 18.5.1, they are more stable and don't need the hardware reset trick and will fix performance with V8. HBCC should be turned off. Vega64 can go with intensity 9 if used headless (no desktop running on it), Hmmm. That seems to have worked for all but two of the gpus, which are still hashing in the low hundreds. I did DDU, and device manager says all vegas are using hte same driver... Like a bad nightmare, 'disable crossfire, disable upls' has floated into my head... Which hasn't worked. Bastard Vegas. |
|
|
|
Just ran into the problems with hashrate/V8/blockchain driver
So we need to finally update drivers. For a Vega owner, that may result in endless pain and lifelong suffering but ok...
Do the new drivers work with powerplay tables and other tricks we are used to?
I know. It's all very well saying 'update your drivers' but as we all know, with Vegas that's just the beginning of a cosmic battle. |
|
|
|
Hashrate on cnv8 has cratered. Vega 64s/blockchain drivers/HBCC enabled. intensity 7. hashing at under 1k...
This is a known problem with the blockchain drivers. Update to 18.5.1, they are more stable and don't need the hardware reset trick and will fix performance with V8. HBCC should be turned off. Vega64 can go with intensity 9 if used headless (no desktop running on it), Hmmm. That seems to have worked for all but two of the gpus, which are still hashing in the low hundreds. I did DDU, and device manager says all vegas are using hte same driver... |
|
|
|
|
Hashrate on cnv8 has cratered. Vega 64s/blockchain drivers/HBCC enabled. intensity 7. hashing at under 1k...
|
|
|
|
|
the 580's are alright on cryptonight too, OK on Lyra, so quite versatile. On the other hand, all the AI offerings only accept Nvidia cards, so a few cheap 1070s might be worth investigating for that and X16.
I think if I were buying now on a budget I'd probably build two rigs, one 580s and one 1070s. eBay prices are very good right now, although profits are tiny too, so don't be expecting to make your fortune.
|
|
|
|
self promotion? I would fall on this trick...it's an old one and if you know how the thief as thinking... yea, this is BS.
Post some pictures of actual miners. I can make a photo of 20 boxes as well.
Aye. And the OP is surprisingly defensive about the company in question, having come all the way here to ask if they were legit, then quickly placing a sizeable order anyway. Not a timeline that makes much sense if you think about it for the smallest of microseconds, IMHO. |
|
|
|
I did actually follow advice on disabling windows update. There were a number of places to disable it including registry and these were all taken care of and windows still managed to force update. I will take your advice on cloning though, this was something I thought about but didnt follow through on. Going to do a fresh install on it this weekend. Is it worth trying to use one of the later windows builds or should i just stick with 1703 and blockchain drivers? @GordoLui I did try 2 other drivers other than blockchain and had no luck, but willing to try if i can get stable. @MahaBTC I dont think it was a miner issue, rather something with some of the settings somewhere but honestly will use any miner that works. I dunno if this helps, but I spent hours swearing at one Vega rig, only to discover that the card indexes had changed after an update. So if you're using overdriveN tool to OC, you might be overclocking the wrong card (I had a mix of 56/64, and used one on a lower clock as an output). You situation might be different, but once I'd figured that out mine has been rock solid since (touch wood)! Blockchain drivers, of course. |
|
|
|
|
Doubt we'll see more than mid-30s on Eth, and as everybody has pointed out, that'll be irrelevant soon with reducing block reward (at least for huge swathes of the world due to power costs).
Gamers already have a huge range of last-gen cards at rock bottom prices, or better firepower in Vegas/1080tis, so not sure who these are aimed at TBH. If they're cheap enough maybe...
|
|
|
|
|
Show Posts
