Message boards :
Number crunching :
Bitcoin GPU-based Mining Machines good for BOINC / SETI?
Message board moderation
Previous · 1 . . . 51 · 52 · 53 · 54 · 55 · 56 · 57 . . . 66 · Next
Author | Message |
---|---|
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Have you tried the card in another system? Or in the same system but not using the riser, to verify it’s the card that’s the problem and not the riser itself. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
Have you tried the card in another system? Or in the same system but not using the riser, to verify it’s the card that’s the problem and not the riser itself. I tried in two different systems, and its doing the samething. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Sounds like a dead card then. Is it still under warranty? GPUs usually have a 3 year warranty. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
bloodrain Send message Joined: 8 Dec 08 Posts: 231 Credit: 28,112,547 RAC: 1 |
got mobo ,cpu,ram, psu. testing now. but dam hs giving me issue with install. |
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
Do you guys believe that the Supermicro boards are more stable then regular desktop boards? |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Do you guys believe that the Supermicro boards are more stable then regular desktop boards? Server MB's are more stable than regular desktop M boards. So are many mining specific MB's Nearly any higher end MB with top end VRM support will suffer through a lot of stones and slings and keep on crunching. I never checked to see if Supermicro makes consumer grade MB's which might be why its not Supermicro vs. Supermicro. Tom A proud member of the OFA (Old Farts Association). |
Spartana Send message Joined: 24 Apr 16 Posts: 99 Credit: 41,712,387 RAC: 25 |
Do you guys believe that the Supermicro boards are more stable then regular desktop boards? Server mobos can certainly come with their own challenges, though. Most notably for crunching, the way some server boards handle their PCI bus architecture can be an issue for GPU computing, and you also may be limited in your OS selection. I love using servers for their robustness and stability...just be aware that getting them up and running for crunching can sometimes be a challenge. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Do you guys believe that the Supermicro boards are more stable then regular desktop boards? The fastest system on the entire seti project is running on a supermicro server board with 10 GPUs ;) Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Spartana Send message Joined: 24 Apr 16 Posts: 99 Credit: 41,712,387 RAC: 25 |
Ha. I am certainly not knocking supermicro—I have used them, but never crunched with them. My post was simply a word of caution for any folks that may run out and start buying server hardware and expect it to act like a PC. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
I’ve used lots of supermicro boards over the years. And I haven’t noticed any of them working any differently than normal boards. At least as far as GPU support or OS support or anything like that. It’s essentially just a normal board with server features. Better reliability, ECC memory support, IPMI/LOM support, more server feature centric BIOS with knobs to turn not seen on consumer boards (especially in the memory and PCIe sections), and available in more proprietary form factors. Stuff like that. Where you see drawbacks is really only in overclocking support. Which shouldn’t be surprising, overclocking is rare in the server space due to the emphasis on reliability. Server boards from companies like Intel/HP/Dell, and others that seem to exist in their own proprietary bubble, I can definitely see weird issues you’re describing. But not really supermicro. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
If you are careful and do the research you can avoid server MB's with non-standard power supply connections (I'm looking at you HP/Dell). The other thing you need to be careful about is how a server is cooled. There are many servers on the market that when you boot them up they sound like a loud shop vac and keep getting louder :( In the Seti-Mining Machine world you need enough cpu power to be able to drive your gpus at full speed. Tbar has demonstrated this to my satisfaction. My problem is I want to crunch cpu tasks (perhaps from a non-gpu project) too. And that requires a minimum surplus of threads not dedicated or nearly dedicated to driving the gpus. So if you have a realistic budget in mind and want advice on "what to build" start a new thread and we will be happy to pile on..... errr respond :) Tom A proud member of the OFA (Old Farts Association). |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Price-wise I'd still recommend the 12 slot BioStar Mining board, even if all the open-box sales are gone. Now the TB250-BTC Pro will cost full price, $75. I haven't had any trouble at all with my TB250 Pro, and the times are just as fast as they are on 'normal' desktop board. Compare the 1060 times on the $33 BioStar board to the 1060 times on the $400+ ASUS Workstation, they are pretty much the same. The BioStar board is currently running 8 lower-end GPUs and it is up to #26, add a 1060 or two and it will be on the first page. |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Well, I almost ordered an i7-6700 ES for $85 from China to light up my Mining board but got distracted because I offered someone a lower price on another MB and he took me up on it. Maybe next month. Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
@Tbar, What kind of power draw is your top performing rig pulling? https://setiathome.berkeley.edu/show_host_detail.php?hostid=6813106 I am trying to get a handle on what kind of numbers I would be facing depending on the exact setup I am running if I want to put it on a UPS. Thank you. Tom A proud member of the OFA (Old Farts Association). |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
The numbers change every second. Take your best shot at it. The machine uses Three Power Supplies. Uptime is close to 18 days. Mon Oct 21 15:34:07 2019 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 410.104 Driver Version: 410.104 CUDA Version: 10.0 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce RTX 2070 On | 00000000:01:00.0 On | N/A | | 55% 69C P2 170W / 185W | 1545MiB / 7951MiB | 95% Default | +-------------------------------+----------------------+----------------------+ | 1 GeForce GTX 106... On | 00000000:02:00.0 Off | N/A | | 85% 69C P2 118W / 120W | 1269MiB / 3019MiB | 99% Default | +-------------------------------+----------------------+----------------------+ | 2 GeForce GTX 106... On | 00000000:03:00.0 Off | N/A | | 85% 70C P2 85W / 120W | 1269MiB / 3019MiB | 96% Default | +-------------------------------+----------------------+----------------------+ | 3 GeForce GTX 106... On | 00000000:06:00.0 Off | N/A | | 90% 66C P2 104W / 120W | 1269MiB / 3019MiB | 91% Default | +-------------------------------+----------------------+----------------------+ | 4 GeForce GTX 106... On | 00000000:07:00.0 Off | N/A | | 70% 65C P2 82W / 120W | 1269MiB / 3019MiB | 81% Default | +-------------------------------+----------------------+----------------------+ | 5 GeForce GTX 106... On | 00000000:0B:00.0 Off | N/A | | 81% 66C P0 37W / 120W | 12MiB / 3019MiB | 2% Default | +-------------------------------+----------------------+----------------------+ | 6 GeForce GTX 106... On | 00000000:0D:00.0 Off | N/A | | 60% 66C P2 110W / 150W | 1269MiB / 3019MiB | 99% Default | +-------------------------------+----------------------+----------------------+ | 7 GeForce GTX 1070 On | 00000000:0E:00.0 Off | N/A | | 49% 56C P2 154W / 151W | 1297MiB / 8119MiB | 94% Default | +-------------------------------+----------------------+----------------------+ | 8 GeForce GTX 1070 On | 00000000:0F:00.0 Off | N/A | | 60% 61C P2 98W / 151W | 1297MiB / 8119MiB | 88% Default | +-------------------------------+----------------------+----------------------+ | 9 GeForce GTX 1070 On | 00000000:10:00.0 Off | N/A | | 99% 75C P2 110W / 151W | 1297MiB / 8119MiB | 92% Default | +-------------------------------+----------------------+----------------------+ | 10 GeForce GTX 106... On | 00000000:11:00.0 Off | N/A | | 70% 63C P2 110W / 150W | 1269MiB / 3019MiB | 82% Default | +-------------------------------+----------------------+----------------------+ | 11 GeForce GTX 1070 On | 00000000:12:00.0 Off | N/A | | 76% 65C P2 46W / 151W | 1297MiB / 8119MiB | 0% Default | +-------------------------------+----------------------+----------------------+ | 12 GeForce GTX 1070 On | 00000000:13:00.0 Off | N/A | | 55% 62C P2 95W / 151W | 1297MiB / 8119MiB | 84% Default | +-------------------------------+----------------------+----------------------+ | 13 GeForce GTX 1070 On | 00000000:14:00.0 Off | N/A | | 55% 62C P2 79W / 151W | 1297MiB / 8119MiB | 93% Default | +-------------------------------+----------------------+----------------------+ |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
you could log that info to a file using nvidia-smi --query-gpu over say 10-20 mins, then take an average. add ~100W to that average for the CPU/MB/RAM/PSUeff and it should be a decent estimate of the average draw from the wall. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Or... use a kill-o-watt and directly measure the power usage. Mine shows 770 W with 2x2070+2x1070. In my case a UPS must be in the range of 2 KVA to be sure no surprises could happening. Long time ago someone old technician tell me something: If you go for a UPS, forget all you learn from your Electric Engineering school, measure the power used (W) and use at least 2X (in VA) that number, add 20% and you will be happy. I never had a UPS with problems by following that "unwritten rule". My 0.02 |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
If my calculator didn't miss anything you are running at least 1288 watts. And some of those draw numbers looked a bit low. I wouldn't surprise me if you were pulling near 1800 watts. But this is a lovely selection/distribution of different gpus. Thank you for giving me some idea of the power load. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Ian is on a 240V system and I don't think there is a 240V Kill-a-Watt meter version. He would have to get a standard multimeter with a current transformer pickup or amp-clamp to measure his draw on his 240V circuit. You have to be aware of the type of output that a UPS provides. Most consumer versions only approximate a sine wave with a stepped voltage output. The area under the curve needs to be derated against what a true RMS sinewave output provides for available power in watts. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
2 out of 3 are on 240V. the 7-GPU system is on 120V still. it pulls about 1250W (24hr avg measured with Kill-A-Watt) I could get a metered PDU for the 240v systems, but as with all things marketed for the server space, they are $$$$ I imagine Tbar's 14-GPU system is probably 1400-ish on average. just a guess. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.