Jump to content

Tomatodude95

Member
  • Posts

    150
  • Joined

  • Last visited

Everything posted by Tomatodude95

  1. I have looked at it during the competition and tried it out a bit, so I'm no expert. Also I only tested Google Cloud so this might differ significantly from other providers. If you are willing to put in the work to deal with preemptible instances in Google Cloud you can setup a Node with 8 vCores and 4 Nvidia T4s for about 0.5 USD an hour, which makes it about 12 USD a day. If you go non-preemptible it becomes quite a bit more expensive, it'll be something like 1.18 - 1.70 USD per hour depending on how long you run it (the longer you do the more discount you get, ie. the closer to the 1.18 USD you get). You might be able to go with lower configs for the CPU, if you aren't using the CPU. I tried the 4 vCore version as well, but if I folded on the CPU as well I had an issue where the system ran out of RAM. (You get about 1GB of RAM per core.) However, the difference between 4 and 8 vCores is less than 3 cents per hour (About 22 USD for 24/7 preemptible usage per month). AFAIK on AWS it's a bit different because you can more or less decide at what price you want you instances to run and as long as prices are lower than you predefined price and AWS doesn't need the machines otherwise they will run. Anyone who is more informed about Spot instances on AWS can correct me if I'm wrong please!! Auzure has similar offers and possibilities, but I'm not informed about those. So it really depends on your definition of reasonable and expensive. Might make sense if you need some power in the short term or have some free credit to try it but it's still quite a chunk of money, I mean the cheapest option I presented costs 12 USD/day or 360 USD/month. For me it was fun to play with and good to get some experience with but I wouldn't consider running it in the long term. Anyhow if you want to run hardware over a longer period of time it is usually cheaper to buy the hardware and run it at home/locally. (Yes, I know not always, there are benefits to having stuff in the cloud.) Additional Info about Performance: So if you don't know what the performance is like of a Nvidia T4, it's about 800k PPD, so one of the machines with 8 cores and 4x T4s got about 3.2M PPD (I measured values from 2.8M - 3.8M PPD depending on the units and specific workload.) Hope that answers your question, don't hesitate to ask if you want more details on something.
  2. I know, no worries, that's why I put the there. Probably wouldn't be in my system right now if it wasn't just to get that bit of extra PPD when folding.
  3. Ok, good to know and thanks for the clarification. But how dare you go after my 770 like that ... that was once my pride and joy.
  4. Yeah thats more or less what I meant, since your using a mining card I don't know what the drivers look like on linux (I've never had a mining card) but if your using standard drivers (GTX version I mean) for the mining card, there's probably some miss-match or inefficiency somewhere on there. Nvidia always can be tricky to get right under linux. Do you mean modded BIOS as in a different one from the regular GPUs using a TPU 106 die? (So that would be a 1060 if I'm not mistaken right?)
  5. Yeah that makes sense, I still have a windows drive in my system for the exact same reason. I use Linux on the daily and whenever a game runs on linux, but I'll reboot and switch to Windows when it doesn't run on linux (yet, hopefully).
  6. I'm guessing it's the drivers. PCIe bandwidth isn't really an issue for folding. I have an old GTX 770 running in a probably broken slot that only get 3.0 x1 bandwidth and it performs as it should/did before it was bumped by the upgrade.
  7. First of all, yes there probably still is some GPU WU shortage out there. I'm guessing your GPU is showing up on either the web-interface or the advance client right? If not you need to add another slot in the advanced config for your GPU. You should be able to do both CPU and GPU WU at the same time, so either be patient and wait to get a WU or you could set the client-type to advanced or even beta in the advanced client under the expert tab. Does that make sense? If not just tell me what you need clarification on.
  8. Ok, yes I know about WSL, but at least for me it's not really the same. But I get where your coming from and especially with the remark below, it makes a lot of sense though.
  9. Is that a wallpaper/backgroup promoting Linux on a Windows machine I see there ...
  10. Yep, that's how all the communication for FAH works, through telnet at what ever port you have configured, 36330 per default. That really should work, otherwise you have some other problem, it's never failed me and I've used it 10s to 100s of times.
  11. Yep, I have no clue about the actual billing right, now just going by my own calculations of usage time and price per hour. I've also had to make sure that I don't include the 30% price reduction you get after a number of hours cause that can bite you in the end.
  12. Ok, seems like your probably right. Are you gonna try it? I mean like in the next couple of days? I don't want to take the risk as everything is working right now and I don't want to waist PPD in trying to convert, when I don't know if it will work the way I want it to, but I'd be very interested in hearing from you once you've tried it.
  13. Yes, please keep us up to date, really want to know as well. However, I don't think it will work, so I'm not gonna risk it by trying it out. To me the Statement is pretty clear, but I could of course be completely wrong.
  14. Interesting I haven't had any problems getting both 2x P100 and 2x ( 4x T4 ), which I'm switching to at the moment. Never had a single request denied because of capacity, only had to make sure I had the correct quotas in the correct area. I got all of mine at US-west-1, so Oregon.
  15. Yes, it would be however sadly you can't use your Free Credits here. It's one of the limitations, they don't want you to use your credits too efficiently. https://cloud.google.com/compute/docs/instances/preemptible
  16. If you just use one of their Deep Learning Linux, you will be prompted if you wanna install the Nvidia drivers on the first boot and it does everything automatically. Then you just need to download the .deb file and install it. I think this is much simpler.
  17. You can set your client to advanced or beta mode to get WU that aren't (quite) validated yet and have a higher chance of failing, but there is more available work so you will most likely see a more consistent stream of WUs. You enable it by going into the advanced control application and then to the settings of you client, under advanced you can set an extra option called client-type and value of advanced. If you only want it for a certain slot you can also add it in your slot configuration at the bottom under extra options.
  18. Make sure you wait for the sheet to load completely and then click the document before trying to search with Ctrl + F otherwise the browsers search gets used, and only the beginning of the document gets loaded and can be searched by the browser, but Google Sheets search can search the whole document.
  19. Did you search by forum name? 2229 Flylead 278654 1261618 0.01% 0.06% It's by forum not FAH name!
  20. https://apps.foldingathome.org/serverstats Seems like the server is up, but no collection server so you might have problems submitting finished WUs.
  21. Yeah, as far as I can tell these don't seem to be used by FAH. However I can't be 100% certain, at least one of them seems to be running some sort of completely unrelated website, the other one doesn't have anything on it to support the conjecture that it's running for the FAH network. To me these seem to be unrelated IP blocks. Neither of them is mentioned on the FAH server list or their collection servers. In conclusion it's highly unlikely that these have anything to do with FAH, I would guess that the problem is somewhere else.
×