Posts by Greg

1) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107683) Posted 7 Nov 2022 by Greg_BE Post: Hey Sid, KMA...seriously. You think you own these boards and know it all. If you do, then why not go work for them? I'll come back here to pester your ass every now and then. But I'm settled on projects that have plenty of work, not the hodgepodge castoff stuff that comes here now.
2) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107621) Posted 1 Nov 2022 by Greg_BE Post: There really is nothing quite so tiresome as random people claiming the project is going to close down because they had a period with no tasks. Not sure as the front page and server status page haven't been updated quite yet, but another batch of tasks seem to be coming through Very small - maybe 50-60k. Not even enough to fill all cores at any one time. Ah well, something. The majority of the work goes to their 3 channel deep learning system. That's what I saw on Robetta at least. Robetta is the system that feeds all systems right? There is no separate system that feeds here? You never seem to get tired of whining about something, do you I thought you were going to leave The latter would be preferable if the former is the only thing you ever have to say Because you all say, oh theres work...yeah..where? I say theres work for their neural network and not so much for us. So I wonder where you guys get your data from that there is "work" for us. I go through 17 pages of data at random intervals and find only stuff for neural network. So what pages and what tags have the stuff for here? That's what I am pushing at. You might try to use some logic in that head of yours instead of whining all the time about what I say. I get tired of you just whining and not providing answers to show that I am wrong in what I say. Don't like what I have to say, ignore it and move on or just put me in your list of people you don't want to see. Computing status Work Tasks ready to send 0 So where in all those pages on Robetta is the next supposed batch of work?
3) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107614) Posted 29 Oct 2022 by Greg_BE Post: [snip] The majority of the work goes to their 3 channel deep learning system. That's what I saw on Robetta at least. Robetta is the system that feeds all systems right? There is no separate system that feeds here? Robetta is a system that allows people outside their campus to enter work. This creates workunits with names that start with rb_ . Work can also be created on campus, giving workunits with names that don't start with rb_ and at least used to be most of what was available. ok, so short of clicking every workunit or whatever...how can you tell at a glance what is going to come here?
4) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107612) Posted 29 Oct 2022 by Greg_BE Post: There really is nothing quite so tiresome as random people claiming the project is going to close down because they had a period with no tasks. Not sure as the front page and server status page haven't been updated quite yet, but another batch of tasks seem to be coming through Very small - maybe 50-60k. Not even enough to fill all cores at any one time. Ah well, something. The majority of the work goes to their 3 channel deep learning system. That's what I saw on Robetta at least. Robetta is the system that feeds all systems right? There is no separate system that feeds here?
5) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107608) Posted 28 Oct 2022 by Greg_BE Post: wow..17 tasks available..whoopie! Where do you see the so called upcoming RAH stuff? I looked randomly up to page 17 of Robetta and all I see is neural network stuff. So either I am looking in the wrong place or something does not add up.
6) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107594) Posted 27 Oct 2022 by Greg_BE Post: Rosetta is out of work, DENIS has a model problem, COSMO has a SSL problem, what else is going on out there with problems? Seems like you need 12 projects to get 8 of them with work at any given time.
7) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107591) Posted 26 Oct 2022 by Greg_BE Post: WCG now appears to be trying to get more useful work done by sending mostly tasks with small total sizes of the input files, such as tasks for the OPN1 subproject. Is that what others are also seeing? Not especially. My machine is set to process 4 WCG tasks at a time and is currently running 2 ARP1 and 2 OPN1 tasks. (and one more OPN1 ready to go.) True, I get fewer of the ARP1 tasks, sometimes none. Most of the time when my Boinc-client requests more work, it does not get any. It is also runninig 4 Rosetta 4.20 tasks. Three 14res_af and one SRS9mpnn. They're now back to tasks for at least three sub-projects, at the cost of much less total work. The ARP1 tasks now seem to be only retries for tasks originally sent out about a week ago. Download problems again. I'm not doing ARP. Just the others. Grinding through them just fine. Getting mostly OPN.
8) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107589) Posted 26 Oct 2022 by Greg_BE Post: Thanks, but it doesn't work for me. Also, I can't attach to any location except "school", even if I detach from the project and reattach. EDIT: I can change locations by "resetting" the project, instead of detach/reattach. But even that trick worked only on my Win10 machine, not my Ubuntu 20.04 machines. Weird, because I am using default profile and all I did was recheck rainfall and save and then uncheck it and save and the use BOINC to update.
9) Message boards : Cafe Rosetta : Off topic stuff from technical discussion (Message 107585) Posted 26 Oct 2022 by Greg_BE Post: I use my phone all the time but I've never been done for it, very easy to drop it in your lap out of sight when you see bacon. Anyway, they can't tell I'm holding a phone, I could just be resting my head on my hand. they hide in undercover cars parked among other cars or off to the side somewhere and look for you. You can't drop your phone unless you know which bright and shiny car is theirs and then you would not be paying attention to your driving or your phone. They can also work in teams. The undercover sits in the parking and then radios your car type and plate to a bike cop waiting further down the road or to the check point where they are doing an action and randomly checking cars paperwork. I've seen on the police show here, a bike cop in downtown nailed a person for using their phone with their hand. They even use trash cans with holes in the side for the camera to nail you on a residential street for speeding or other locations where a wheeled trash can will blend in. Undercover cars have infrared or ccd plate readers, so they don't need the big flash of the old days. They can get you in the dark without you knowing until a ticket shows up in your mailbox.
10) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107584) Posted 26 Oct 2022 by Greg_BE Post: After some conversation over in WCG, it seems there is a little trick to get work flowing again. Go into your profile and change your project settings, save that and then change back to what you had and send out a update via BOINC. I did that and I have work again. Transient errors were apparently fixed last week according to the tech wizard. So I guess they got the bugs fixed finally.
11) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107580) Posted 25 Oct 2022 by Greg_BE Post: I don't expect Krembil to like this, since it means little work for the MCM1 subproject they are especially interested in. Then why don't Krembil put some more money into their falling apart server? I think we should start calling them Crumble. I think things were a lot smoother when IBM ran WCG. They did. Why did IBM drop them like a rock anyway? I never read that part.
12) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107577) Posted 25 Oct 2022 by Greg_BE Post: Try to run cd /d c:\Program Files\BOINC :loop TIMEOUT /T 10 /nobreak boinccmd.exe --project http://www.worldcommunitygrid.org update TIMEOUT /T 256 /nobreak goto loop Better yet...remove WCG. A few tasks once a week or something. come on. Posted in their forum. Got 5 identical responses. So all this rah rah rah stuff WCG posts has nothing to do with their inability to generate work. Only that they can talk about the stuff they had run in the past.
13) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107574) Posted 24 Oct 2022 by Greg_BE Post: This is all I get these days 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for OpenPandemics - COVID 19 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for OpenPandemics - COVID-19 - GPU 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for Help Stop TB 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for Mapping Cancer Markers 10/24/2022 2:50:37 PM \| World Community Grid \| Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them 10/24/2022 2:50:37 PM \| World Community Grid \| Tasks for Intel GPU are available, but your preferences are set to not accept them I have no idea what the brain farts is going on over there. So you ask for CPU work and it says you can only have GPU work, yet I ask for GPU work and it says I can only have CPU work. Krembil are not the sharpest knives in the drawer. And no NVDIA work..just AMD/ATI and Intel? What the +-*/ ! Come on....
14) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107572) Posted 24 Oct 2022 by Greg_BE Post: This is all I get these days 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for OpenPandemics - COVID 19 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for OpenPandemics - COVID-19 - GPU 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for Help Stop TB 10/24/2022 2:50:37 PM \| World Community Grid \| No tasks are available for Mapping Cancer Markers 10/24/2022 2:50:37 PM \| World Community Grid \| Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them 10/24/2022 2:50:37 PM \| World Community Grid \| Tasks for Intel GPU are available, but your preferences are set to not accept them I have no idea what the brain farts is going on over there.
15) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107554) Posted 22 Oct 2022 by Greg_BE Post: Now you have seen every angle and data point as to how the GPU's operate with MOO and FAH working simultaneously. There is no OC on this. Forgot to turn that on. So this is default mode.
16) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107553) Posted 22 Oct 2022 by Greg_BE Post:
17) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107549) Posted 22 Oct 2022 by Greg_BE Post: Rule 1: If heat is pouring off the chip, it's doing a lot of work.- The fans get their workout everyday. At the moment they are just doing the easy stuff with FAH. Apparently TN needs everything cpu, thought I suppose I can write a script to put that down to 14 or something.
18) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107546) Posted 22 Oct 2022 by Greg_BE Post: But MOO is sharing the cards with FAH on my system. So maybe that downgrades the time a bit? And I don't have the fastest cards. A 1080 plain and a 1050TI
19) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107539) Posted 22 Oct 2022 by Greg_BE Post: So whats happening then in plain terms is that the MOO task is huge and complex and splits its self among the two GPU's and the CPU's to be more efficient in getting the work done?
20) Message boards : Number crunching : Problems and Technical Issues with Rosetta@home (Message 107534) Posted 21 Oct 2022 by Greg_BE Post: There are generally two ways to distribute computation across multiple devices: Data parallelism, where a single model gets replicated on multiple devices or multiple machines. Each of them processes different batches of data, then they merge their results. There exist many variants of this setup, that differ in how the different model replicas merge results, in whether they stay in sync at every batch or whether they are more loosely coupled, etc. Model parallelism, where different parts of a single model run on different devices, processing a single batch of data together. This works best with models that have a naturally-parallel architecture, such as models that feature multiple branches. This guide focuses on data parallelism, in particular synchronous data parallelism, where the different replicas of the model stay in sync after each batch they process. Synchronicity keeps the model convergence behavior identical to what you would see for single-device training. Specifically, this guide teaches you how to use the tf.distribute API to train Keras models on multiple GPUs, with minimal changes to your code, in the following two setups: On multiple GPUs (typically 2 to 8) installed on a single machine (single host, multi-device training). This is the most common setup for researchers and small-scale industry workflows. On a cluster of many machines, each hosting one or multiple GPUs (multi-worker distributed training). This is a good setup for large-scale industry workflows, e.g. training high-resolution image classification models on tens of millions of images using 20-100 GPUs. More at: https://keras.io/guides/distributed_training/

Next 20

Posts by Greg_BE