Recent DAB topics

[ihrsetrdr]

1. Recently it was decided by informal vote to make the DAB forum accessible to FoldingForum mods, with -rw privileges.


2. A thread regarding P7200 inconsistencies started by Torbit, mostly a dialogue between Dr. Pande & Torbit, which led to the points for Project 7200 to be recalculated using the updated benchmark scheme.


3. Yesterday Dr. Pande posted the following:

DAB reboot, ground rules

There's been a lot of activity here and some changes, so I thought it would be a good time to have a "DAB Reboot" thread, reminding of the changes, and also getting people on the same page. From now on, the Mods and PG can also read and write in these threads. My hope is that the DAB can bring donor feedback and representation from the big teams as well as report back. The Mods do a good job of representing a broader range of donors, especially those who post on our forum.

I also wanted to remind everyone of the general ground rules here. In order to have a frank discussion, please do not relay information to your teams until the discussion is over and we have all agreed on the plan going forward. In particular, give me a chance to make a formal announcement of any new changes, rather than making any announcements yourself. Also, we expect everyone to behave in a professional manner.

Finally, I think we need to have some ground rules expectations for how often we expect people to read and take to reply to posts here. It's clear that not everyone in the DAB is reading this section daily, and that's ok (we all have other jobs, etc). How about we agree to the expectation that we will give people a few days to read and reply and not expect immediate responses to posts. However, please do try to get here at least a few times a week, as we do not want to keep important issues delated unnecessarily.

If anyone (DAB member, Mods, PG member) cannot keep to these rules, we'll have no choice but to find an alternate person to serve in this capacity. I personally really hate interpersonal drama and would like to keep that to a minimum here. However, that doesn't mean we all have to mindlessly agree, as this is explicitly a place to debate concerns and issues, just that we discuss these issues in a professional, open-minded manner. I bet that everyone else feels similarly.

I'll post the short term agenda items in another thread.

4. Another thread started by Torbit suggesting that the Benchmark System needs upgrading:

If donors would prefer, we can possibly turn off HT on that machine to make it closer to an i5-750.

You'd have to turn it back on though to benchmark bigadv WUs. It's a tough call but I think I prefer benchmarking against real cores. What I'd really like to see is PG invest in a new benchmark system if you can find the budget for it. Something with 8 or 16 real cores.

We have a budget from NIH which includes a new benchmark machine, but it's slated for a several years in the future, since we did just buy this one about a year ago. Hopefully we can get by for a year or so. However, if people think that this is a disaster and worth spending ~$10K (we have to buy many of them in order to avoid benchmarking becoming a bottleneck), I can always take money away from something else (although we need new servers, new CS, etc ...)

One question for the DAB would be if you could change the benchmark machine to something else, what would you want to see?

Something with 8 or 16 real cores.

+1

One question for the DAB would be if you could change the benchmark machine to something else, what would you want to see?

Intel motherboard with TWO Westmere class Xeons (possibly E5640) and triple channel memory.

In response, Dr. Pande solicits input with the following:

How about you guys suggest a *specific* machine and we can go from there? Dr. McCord had some specific suggestions I think.

 

[ChasR]

2. A thread regarding P7200 inconsistencies started by Torbit, mostly a dialogue between Dr. Pande & Torbit, which led to the points for Project 7200 to be recalculated using the updated benchmark scheme.

what updated benchmark scheme?

 

[ihrsetrdr]

Not sure, that statement originated in this post.

 

[harlam357]

Posting that I have read.

Based on Dr. Pande's comments, I guess I don't understand about how Tim (our representative) isn't supposed to report back to us before decisions are made. I thought that we, the T32 committee, were to make a collective vote and Tim as the chair simply represented that vote at the DAB table. Am I missing something?

 

[ihrsetrdr]

There have been no decisions made by DAB representatives in foldingforum.org's DAB section, just discussion of suggestions made, either by users, or ideas that Dr. Pande wants feedback on.

As mentioned, the "updated benchmark scheme" was only referenced to in the one recent DAB post that I quoted, and not something that had ever been discussed in-depth in the DAB; I hadn't done a thorough search, but if in fact the benchmark system had been updated, it would have been done in the 'regular' hidden section of ff.org.

As for Torbit's suggestion to Dr. Pande that the benchmark machine be updated, Dr. Pande is open to suggestions, so if anyone here would like to pass a suggestion to Dr. Pande regarding the Benchmark equipment, I'll be happy to facilitate that.