BigAdvPPD chart

[Patriot]

In regards to the BigAdv PPD/ TPF charts...

Please only post your average TPF.
You can post your average on a fast wu if you want...but please do not post your fastest frame.
(perhaps we can add a column for that if you want but I want accurate data and for that we all need to post the same information.)
We can add a epeen stroking section for the benchmarker in you but I would like to keep our data valid.

Also...if you apply any filters...release them when you are finished.
Otherwise it looks like data has been deleted...

https://docs.google.com/spreadsheet/ccc?key=0Aspl6-RkaxtFdHdTdUdmUjhWalpXWVZ2S2xvejBDcHc

I try and keep it as accurate as possible...Please aid me in this.


I will also be starting a refresh of the SMP TPF database... I will need smp data in boatloads.

 

[Schro]

Are we just adding TPF and WU along with system data and the points and whatnot will be handled by you? Or do we need to dig all of it up?

Ninja edit - added my rig, tpf and Wu number for you.

 

[Grandpa_01]

The spread sheet is much appreciated Patriot. (Thank you) I am assuming you are just going to add a sheet for SMP to the current Bigadv spread sheet.

As far as the epeen column goes that is virtually useless information to big of a variation between wu for any type of ligament comparison there is no standard what so ever in a within a given wu.

 

[musky]

Just out of curiosity, what are you trying to accomplish with this?

 

[Patriot]

The spread sheet is much appreciated Patriot. (Thank you) I am assuming you are just going to add a sheet for SMP to the current Bigadv spread sheet.

As far as the epeen column goes that is virtually useless information to big of a variation between wu for any type of ligament comparison there is no standard what so ever in a within a given wu.

I am open to recommendations...

I was thinking of keeping it all together and adding a column to allow filtering between the different types of wu's.... Column (wu type) smp, bigadv , uni...

I would like to grow it into a database of more than just bigadv... But I might just start a new spreadsheet all together.

Just out of curiosity, what are you trying to accomplish with this?

Lack of sleep saw some fastest frame numbers... grumble fix post... lol.

And awareness ... More data the merrier ... just a few helpful guidelines on what would help keep it the most accurate. As Grandpa noted there is already a ton of variation...

 

[Schro]

I am open to recommendations...

I was thinking of keeping it all together and adding a column to allow filtering between the different types of wu's.... Column (wu type) smp, bigadv , uni...

I would like to grow it into a database of more than just bigadv... But I might just start a new spreadsheet all together.



Lack of sleep saw some fastest frame numbers... grumble fix post... lol.

And awareness ... More data the merrier ... just a few helpful guidelines on what would help keep it the most accurate. As Grandpa noted there is already a ton of variation...

You can use the Form editor to only allow input via a form, rather than opening it up to the internet masses for editing. You can enforce some rules on data entry with it,then make the resultant spreadsheet read only...

 

[Patriot]

You can use the Form editor to only allow input via a form, rather than opening it up to the internet masses for editing. You can enforce some rules on data entry with it,then make the resultant spreadsheet read only...

I do have version control... but yes people can currently mess it up and make me roll it back.
I do want to do a form but have not taken the time to set it up.

Any volunteers ... in due time...ty...solid suggestion.

My other thought/concern... we have other databases... I really hate minimizing the work put into them... there is an older smp tpf database out there... I would like to pull into this one.

 

[sbinh]

Don't know how to get other info to input ...

 

[sirmonkey1985]

if you click tools then benchmark viewer. it lists out every WU you've ever run by the WU number. you can then click on it and it'll list your average. just remember that it averages out over all the times you run the WU, not just the last time you ran the WU.

also in the same menu if you click work unit history viewer, it'll show every single WU you ran + average frame time + PPD + credit you received for it.

those are the two options for finding your average PPD.

 

[quickz]

Great work! It gives us lots of useful informations.
And I know a web-site with the same purpose:
http://www.fahinfo.org/
This web-site contains ten thousands of FAH PPD scores, but all of the data is too old. Seems no one updates for it now.

 

[Patriot]

Great work! It gives us lots of useful informations.
And I know a web-site with the same purpose:
http://www.fahinfo.org/
This web-site contains ten thousands of FAH PPD scores, but all of the data is too old. Seems no one updates for it now.

Lol...it was very much and is a team effort...

I think 5-10 of us built up 25% of that over a weekend to try and get the 8101 points adjusted when it was still in beta.

 

[Toconator]

if you click tools then benchmark viewer. it lists out every WU you've ever run by the WU number. you can then click on it and it'll list your average. just remember that it averages out over all the times you run the WU, not just the last time you ran the WU.

also in the same menu if you click work unit history viewer, it'll show every single WU you ran + average frame time + PPD + credit you received for it.

those are the two options for finding your average PPD.

Excellent, never tried playing around with HFM too much. Great for comparing ppd on diff CPU's. Thanks Monkey! - Patriot: when do you want smp data?

 

[rhavern]

It may be worth mentioning that there are formula embedded in some of the columns, so in order for the rest of the columns to populate, you need to copy a functioning row. With that, I have added my details and fixed some rows that weren't populated.

 

[rhavern]

Also, when filling in my details, I noted that on 6903 section that many entries have the k-factor listed as 38.5 instead of 38.05 that HFM currently reports. Was the k-factor actually this in the past or is this a copy/paste issue?

 

[402blownstroker]

You know, a simple Ruby on Rails app would be a lot nicer for capturing this information. Plus it can easily have many different searching and sorting options. If there is any interest in this, it would be a simple project to hammer out over the weekend.

 

[rhavern]

You know, a simple Ruby on Rails app would be a lot nicer for capturing this information. Plus it can easily have many different searching and sorting options. If there is any interest in this, it would be a simple project to hammer out over the weekend.

That being said, and I am a network guy, so I may be talking rubbish, would it be feasible to have a chat with Harlam357 the HFM author and see if there is a way to automate this process? I did chat with him regarding this possibility 18-24 months ago and he indicated that it was something that he wanted to do, however the v7 rewrite took precedence. FWIW I'm a happy HFM user and beta tester.

 

[tkam]

If someone wanted to put together a web app for this (php, ruby, whatever) I can host it for free no problems.

 

[402blownstroker]

If someone wanted to put together a web app for this (php, ruby, whatever) I can host it for free no problems.

I will create a RoR app this weekend if you want to host it. As initial setup for it, you will need to install ruby-1.9.3 and mysql 5.5.25. If you need help doing that, just let me know.

 

[402blownstroker]

Does anyone know of a good way to determine the following based on a project number:

• Perferred Deadline
• Final Deadline
• K Factor
• Bonus Points

I found a site that has a static list of these, but it would be nice if there was something/somewhere that would have a current/updated listing.

I think I can scrap this site daily right? PS Summary

 

[W.Feather]

I don't see why not, hfm basically does it on a per client basis when updating psummary

 

[Tobit]

I think I can scrap this site daily right? PS Summary

Scrape psummaryC.html if you are interested in beta units as well.

psummary.html = public units and those released to -advmethods
psummaryC.html = combined public and beta units

 

[Grandpa_01]

Also...if you apply any filters...release them when you are finished.
Otherwise it looks like data has been deleted...

Classic I just looked at the page, I had to turn the filters of that somebody left on lol

 

[rhavern]

It may be worth looking at FCI's code if you speak the language it is written in. It does all that stuff already. The developer, smoking2000, has always been helpful and responsive when I have had issues or feature requests. Have a look here: http://foldingforum.org/viewtopic.php?f=14&t=8109. I've been using FCI for a couple of years now.

 

[402blownstroker]

I got most of the infrastructure done last night and mis morning. I should get the rest done here tonight and tomorrow morning to have it ready for beta testing this weekend. For the time being I think I will host it on my web server.

 

[Thomas R]

Added three results!
Think my system is not bad.

 

[knopflerbruce]

4P 16c Interlagos at 4GHz, what sort of PPD would that give on P8101/P8102?

 

[402blownstroker]

4P 16c Interlagos at 4GHz, what sort of PPD would that give on P8101/P8102?

Try it and see what happens

 

[Core32]

4P 16c Interlagos at 4GHz, what sort of PPD would that give on P8101/P8102?

In my estimation, somewhere between a "bunch" and a "sh*t load"

 

[knopflerbruce]

Try it and see what happens

I ran a few benchmarks at 3.8 (wprime 1024m, cinebench and ucbench) with a 2P rig, using some beefy air coolers, but one just with it's own weight as pressure (no mounting), and only needed 1.2v for 3.8GHz. I'm hoping I can do 4GHz at 1.3v on water cooling Not sure if the supermicro 4P VRMs are up to the task, though With some heatsinks and stufff, i think this KGPE-D16 can do that, however.

 

[402blownstroker]

I got the app working with all the major funcationality done. Starting to add some 'fluff' and 'bling-bling'. I am looking for a couple of beta testers to try things out and give feeback on usage. If you are interested in beta testing, drop me a PM here. Hopefully by the end of the day, the app will be ready to start recording live data.

I will also be creating a simple script to import the existing spreadsheet data.

 

[Patonb]

Can you add a column for the number of pocessers, and maybe speed. I know its there in words, but being able to filter it via # of proc, and Core speed would be great now that the 4p's can overclock.

 

[402blownstroker]

In the works

 

[Tobit]

Anyone have 4 6218's OC'd to 2.5 GHz? What's the power draw like? Thanks.

 

[Patonb]

Just thought i'd let you know the chart doesn't filter past row 156.

 

[402blownstroker]

How does one determine the type of a Project: uni vs smp vs ba vs bb? I have seen/found these two pages, but there is no clear indicator of the project type.

http://fah-web.stanford.edu/psummaryC.html

http://fah-web.stanford.edu/psummary.html

 

[Tobit]

How does one determine the type of a Project: uni vs smp vs ba vs bb? I have seen/found these two pages, but there is no clear indicator of the project type.

http://fah-web.stanford.edu/psummaryC.html

On psummaryC, greater than 8500 base credit = BA. You can scrape psummaryB for BB but it isn't always accurate. Projects slip through the cracks into advmethods or genpop. psummaryB is supposed to be beta and advmethods projects only but it hasn't been accurate in a LONG time (if ever truly)

UP vs. SMP is no longer possible to detect because many A4 units can run on either. There are still a few UP only projects out there, they are shown on psummary as "GROMACS" and "PROTOMOL".

 

[Patriot]

On psummaryC, greater than 8500 base credit = BA. You can scrape psummaryB for BB but it isn't always accurate. Projects slip through the cracks into advmethods or genpop. psummaryB is supposed to be beta and advmethods projects only but it hasn't been accurate in a LONG time (if ever truly)

UP vs. SMP is no longer possible to detect because many A4 units can run on either. There are still a few UP only projects out there, they are shown on psummary as "GROMACS" and "PROTOMOL".

A5 core is going to be Big-adv... They were a3 and a5 but they have since relabeled them since they upgraded the core.

 

[Tobit]

Ahh, good Patriot. So 402..

BA = GRO-A5
UP = GROMACS + PROTOMOL + GRO-A4 (some but not all)
SMP = GRO-A3 + GRO-A4

UP only is going to be a challenge to detect as, I've stated already, A4 is a core that can run on both UP and SMP and some units are coded to run on both architectures.

 

[402blownstroker]

Hmmmm..... maybe a good distinction would be BA/BB and everything else?