Any Core 17 WUs Without 'advanced' Flag? [YES]

[Kurtis200200]

I noticed on the Pande blog that Core_17 has been rolled out to full FAH (that is, Core_17 is ostensibly accessible without the client-type advanced flag), however aren't projects 8900, 7810, and 7811 (the only Zeta / Core_17 work units I know of as of this post) all still 'advanced' work units?

Put differently, I removed my 'advanced' flag and haven't gotten any Core_17 work units yet (about 22 hours now). Are they out there, or do I need to keep the 'advanced' flag for now?

[PantherX]

I would say that it's the case of supply and demand. When F@H released FahCore_17 to Full, there was an increase in demand. Since there's a limited supply of FahCore_17 WUs, it is a first-come-first-serve situation. If your GPU requested FahCore_17 WU when all FahCore_17 WUs have been assigned, you will either not get a WU or will get an older FahCore WU (depends on your GPU and the configuration of the Servers).

You can try the advanced flag but since the Projects aren't on advanced on the Servers, it might not make much difference.

[Kurtis200200]

PantherX,

Thanks for the quick reply; I'm happy to stand in line for Core_17s, I just wanted to make sure I was standing in the right one

[bruce]

There's never a guarantee you're standing in the right line. Stanford retains control of the final decisions in the assignment process, not you.

As a project moves from beta to advanced or advanced to full FAH, it is exposed to a multitude of new systems with different hardware, different software, different configurations, and many new opportunities to encounter unforeseen problems. If it all happens without difficulties, good.

If something unforeseen happens, it may take a few hours or a few days to adjust a server setting to fix whatever was encountered. In the interest of minimizing exposure to newly discovered issues, a project might be temporarily moved back to advanced. There also might be a temporary shortage of specific project(s). There's never a guarantee that you'll always get a particular project or group of projects. Their goal is to assign you work that your hardware can process in accordance with current priorities and a project that exposes people to a known problem is automatically a low priority assignment, even if your system doesn't happen to share the characteristics that provoke the problem.

If you were running core_17 on advanced previously there's noting wrong with setting the flag again and seeing what happens.

[Kurtis200200]

Bruce,

Solid advice; Is there a resource that I can consult to see whether a certain project (say, 8900 or 7810 or 7811) is being assigned as an advanced project or as a full project? I've looked through the Projects Summary page but can't seem to decipher that information from what I'm seeing.

[bruce]

No such resource.

I often wish I knew, but it can only be determined heuristically and is subject to change without notice. Stanford's priority setting is also subject to change without notice ... or even availability of a specific server.

[PantherX]

If you want, you can keep an eye out for posts/threads posted in this forum (viewforum.php?f=24). This is the where the researcher can announce the status of the project and also post if the project is moving between F@H stages. However, as bruce said, the information is subject to change so, YMMV.

[7im]

As Bruce said, there is no easy way to tell. However, if you want to dig in a little deeper, look up qd tools.

[davidcoton]

Even with the "advanced" flag, I've found Core17 projects rare on Windows in recent weeks, possibly since before the announced shift to Full.

On Linux, where only core17 is supported (by the v7 client), there's been no shortage yet.

David

[mlpickerel]

I'm not familiar with qd tools, but by going to F@H Projects Summary page http://fah-web.stanford.edu/psummary.html you can get the server ip numbers for the projects you are interested in.
Then go to server status and find the server ip on that page. The OS_WEIGHT_PROGRAM_PORT column holds the information on what stages the computer is assigning to. For zeta projects 7810 & 7811, both hosted on ip 171.64.65.98 the server is only issuing to beta and advanced. For project 8900 on ip 171.64.65.69 the server is assigning to beta, advanced, full, and internal.

This shows that projects 7810 and 7811 are currently not full stage, and that 8900 *might* be full stage. It is also possible that another project on that server is full, and project 8900 remains advanced.

Hope this helps,

Mark

[bruce]

The OS_WEIGHT_PROGRAM_PORT column holds the information on what stages the computer is assigning to. For zeta projects 7810 & 7811, both hosted on ip 171.64.65.98 the server is only issuing to beta and advanced. For project 8900 on ip 171.64.65.69 the server is assigning to beta, advanced, full, and internal.

Yes, this information is available, but sometimes it's useful, but rarely so. The original question was whether any of the three projects was on full or was restricted to advanced. (and the next question is, if they're not all the same, which ones) You can't really answer either of those questions because there's no way to know which of the F/A/B/I flags of the server happen to be populated with WUs. Moreover, if there's more than one project (e.g.- p7810 and 7811) it's not fair to assume that they're both the same. One could be beta and the other could be advanced and you'd have no way to determine that except from reports of people who recently were assigned a new WU.

Pick a server like 128.143.199.96 or 128.143.199.97 each of which have dozens and dozens of WUs and tell me which projects are assigning as Full, which might be restricted to Advanced, and which projects are no longer assigning at all because they have run out of new WUs. I wish I knew a way to determine those answers.

[Kurtis200200]

Mark,

As Bruce notes, your information doesn't tell me exactly what's being assigned under what flags, but it does helpfully let me know that I'm not going to get a 7810/7811 without an advanced or beta flag, which is I think as much information as Stanford currently makes available, so thanks for helping me play this self-serving point-maximizing logic game haha.

It would be nice to know more about assignments so I could plan my flagging to do the most good, maybe as a Christmas present Pande group? :):)

[7im]

Those projects can be moved to full fah at any time, and new projects are on the way if not here already. And if those WUs run low, you will be assigned Normal WUs instead of Advanced WUs, so it makes very little difference in the long run.

When core17 comes up to full speed, it won't matter at all.

[Blasko9]

Those projects can be moved to full fah at any time, and new projects are on the way if not here already. And if those WUs run low, you will be assigned Normal WUs instead of Advanced WUs, so it makes very little difference in the long run.

When core17 comes up to full speed, it won't matter at all.

Best guess when core17 will be up to full speed?

I am slogging through 8018s. Did I do something bad?

[7im]

Nothing wrong.

The advanced setting is a request for advanced work units, not a request for core_17 work units. Core_17 is one of many types of advanced work units. You will always get a mix of WUs when there are multipe options available.