TAS Note: Quad Challenge

[dajeepster]

Originally posted by: Rudy Toody

Originally posted by: dajeepster

Originally posted by: Rudy Toody
I am hyperapping 12 threads as we speak. Ibercivis can't keep up!

Edit: biodoc--your credits yesterday are the second highest you have had. Maybe hyperapping does make a difference. My credits per day are going up.

Edit2: Because my hyperapping has raised the cpu temp a bit, I have implemented a program of occasional hyperventilating.

i wonder what the result would be if I hyperapped my "i"?
mmmm.. 16 threads :evil:

In cc_config.xml, in the <options> section, <ncpus>16</ncpus> will do it.

i'll do it as soon as my last class is over and I leave campus

 

[petrusbroder]

dajeepster, your comp is one of the most powerful:
Check here - you are #4! :thumbsup:

and I am #8!
Biodoc is #20,
Neurodog is #24,
Rudy Toody #26 :sun:
Tom Pilippart #40 :beer:
salvorhardin #62

Well done, TAS-mates!

 

[dajeepster]

Originally posted by: Rudy Toody

Originally posted by: dajeepster

Originally posted by: Rudy Toody
I am hyperapping 12 threads as we speak. Ibercivis can't keep up!

Edit: biodoc--your credits yesterday are the second highest you have had. Maybe hyperapping does make a difference. My credits per day are going up.

Edit2: Because my hyperapping has raised the cpu temp a bit, I have implemented a program of occasional hyperventilating.

i wonder what the result would be if I hyperapped my "i"?
mmmm.. 16 threads :evil:

In cc_config.xml, in the <options> section, <ncpus>16</ncpus> will do it.

i don't have the cc_config.xml file in my folder

 

[dajeepster]

Originally posted by: petrusbroder
dajeepster, your comp is one of the most powerful:
Check here - you are #4! :thumbsup:

and I am #8!
Biodoc is #20,
Neurodog is #24,
Rudy Toody #26 :sun:
Tom Pilippart #40 :beer:
salvorhardin #62

Well done, TAS-mates!

I thought your computer was doing better than mine... hmm... maybe it's time for me to overclock

 

[GLeeM]

Originally posted by: petrusbroder
dajeepster, your comp is one of the most powerful:
Check here - you are #4! :thumbsup:

LOL! The top computer on that list is a Pentium III @ 600 MHz
Were those good overclockers?

How long will it take to get a reading on how this i7 is doing without HT on?

How do we compare with HT on? Do I stop for a while (when done testing without HT) and then start again with HT?

Let me know what you want!

I started last night at about local 6 PM. It was without work for maybe 1-2 hours.

 

[biodoc]

Originally posted by: dajeepster

Originally posted by: Rudy Toody

Originally posted by: dajeepster

Originally posted by: Rudy Toody
I am hyperapping 12 threads as we speak. Ibercivis can't keep up!

Edit: biodoc--your credits yesterday are the second highest you have had. Maybe hyperapping does make a difference. My credits per day are going up.

Edit2: Because my hyperapping has raised the cpu temp a bit, I have implemented a program of occasional hyperventilating.

i wonder what the result would be if I hyperapped my "i"?
mmmm.. 16 threads :evil:

In cc_config.xml, in the <options> section, <ncpus>16</ncpus> will do it.

i don't have the cc_config.xml file in my folder

Use a text editor like notepad and create the file with this content:

<cc_config>
<options>
<ncpus>16</ncpus>
</options>
</cc_config>

 

[biodoc]

Originally posted by: GLeeM

Originally posted by: petrusbroder
dajeepster, your comp is one of the most powerful:
Check here - you are #4! :thumbsup:

LOL! The top computer on that list is a Pentium III @ 600 MHz
Were those good overclockers?

I read somewhere that this is the account where they put anonymous contributions.

They could use my account!!

How long will it take to get a reading on how this i7 is doing without HT on?

Maybe 6-12 hours? They don't have "quorum" on WUs so a few hours should be enough.

How do we compare with HT on? Do I stop for a while (when done testing without HT) and then start again with HT?

yes, sounds good to me!:thumbsup:


I started last night at about local 6 PM. It was without work for maybe 1-2 hours.

Hyperapping didn't work well for me but I only tested it for 5 hours.

Now I put "Monoapping" into play. :laugh: I edited my ibercivis preferences and picked one app to crunch. It'll take 1-2 days to flush my cache though.

It's too bad they don't have a message board. I like digging for info on every project i crunch.

 

[GLeeM]

Originally posted by: biodoc

How long will it take to get a reading on how this i7 is doing without HT on?

Maybe 6-12 hours? They don't have "quorum" on WUs so a few hours should be enough.

I don't understand how we will be able to compare.

How do we compare with HT on? Do I stop for a while (when done testing without HT) and then start again with HT?

yes, sounds good to me!:thumbsup:

Do I do "No new tasks" and then let the WUs finish and then after a day or two start again with HT on?

I started last night at about local 6 PM. It was without work for maybe 1-2 hours.

Hyperapping didn't work well for me but I only tested it for 5 hours.

Now I put "Monoapping" into play. :laugh: I edited my ibercivis preferences and picked one app to crunch. It'll take 1-2 days to flush my cache though.

It's too bad they don't have a message board. I like digging for info on every project i crunch.

I wonder if Rudy Toody knows which app gives the best ppd?

Sorry about the mess with the quote My comments are in bold.

 

[Rudy Toody]

Originally posted by: GLeeM
I wonder if Rudy Toody knows which app gives the best ppd?

They have the results display turned off, so we can't look at the WUs to determine the best ppd.

 

[dajeepster]

Originally posted by: biodoc

Originally posted by: dajeepster

Originally posted by: Rudy Toody

Originally posted by: dajeepster

Originally posted by: Rudy Toody
I am hyperapping 12 threads as we speak. Ibercivis can't keep up!

Edit: biodoc--your credits yesterday are the second highest you have had. Maybe hyperapping does make a difference. My credits per day are going up.

Edit2: Because my hyperapping has raised the cpu temp a bit, I have implemented a program of occasional hyperventilating.

i wonder what the result would be if I hyperapped my "i"?
mmmm.. 16 threads :evil:

In cc_config.xml, in the <options> section, <ncpus>16</ncpus> will do it.

i don't have the cc_config.xml file in my folder

Use a text editor like notepad and create the file with this content:

<cc_config>
<options>
<ncpus>16</ncpus>
</options>
</cc_config>

i would put that in my ibercivis folder i would presume?

 

[Rudy Toody]

cc_config.xml resides in the boinc-client folder. It controls all projects on that box.

 

[biodoc]

Originally posted by: GLeeM

Originally posted by: biodoc

How long will it take to get a reading on how this i7 is doing without HT on?

Maybe 6-12 hours? They don't have "quorum" on WUs so a few hours should be enough.

I don't understand how we will be able to compare.

How do we compare with HT on? Do I stop for a while (when done testing without HT) and then start again with HT?

yes, sounds good to me!:thumbsup:

Do I do "No new tasks" and then let the WUs finish and then after a day or two start again with HT on?

I started last night at about local 6 PM. It was without work for maybe 1-2 hours.

Hyperapping didn't work well for me but I only tested it for 5 hours.

Now I put "Monoapping" into play. :laugh: I edited my ibercivis preferences and picked one app to crunch. It'll take 1-2 days to flush my cache though.

It's too bad they don't have a message board. I like digging for info on every project i crunch.

I wonder if Rudy Toody knows which app gives the best ppd?

Sorry about the mess with the quote My comments are in bold.

I would just go to your account page and write down the point total and time of day and then 6 hours later check the point total again and then repeat with HT on. No need to flush the cache in between. This would give a rough idea if HT makes a big difference in ppd. If you could do 24 hours for each run, that would be better.

There is this chart here which suggests there is no benefit to running one app over the other but I did read somewhere that the "i" series with HT on is much better at crunching mixed apps/projects. I'm not sure I understand this but that's why I decided to try hyperapping and monoapping.

I'm just trying to figure out why the "i" series is 2-3x better than the other chips on this project.

If there is no difference with or without HT, then the impressive performance of the "i" series must be due to caching or something else I guess.

 

[Philippart]

they use autodock fairly often, the same app that for instance WCG uses for the childhood cancer project, so somebody with an "i" cpu should check if the speedup canbe seen there too. If not, there's a strange reason at Ibercivis...

 

[dajeepster]

Originally posted by: dajeepster

Originally posted by: biodoc

Originally posted by: dajeepster

Originally posted by: Rudy Toody

Originally posted by: dajeepster

Originally posted by: Rudy Toody
I am hyperapping 12 threads as we speak. Ibercivis can't keep up!

Edit: biodoc--your credits yesterday are the second highest you have had. Maybe hyperapping does make a difference. My credits per day are going up.

Edit2: Because my hyperapping has raised the cpu temp a bit, I have implemented a program of occasional hyperventilating.

i wonder what the result would be if I hyperapped my "i"?
mmmm.. 16 threads :evil:

In cc_config.xml, in the <options> section, <ncpus>16</ncpus> will do it.

i don't have the cc_config.xml file in my folder

Use a text editor like notepad and create the file with this content:

<cc_config>
<options>
<ncpus>16</ncpus>
</options>
</cc_config>

i would put that in my ibercivis folder i would presume?

I just did this for my dual xeon E5310 that i have on campus... it's running 16 threads on seti... I'll just leave this computer like this to see how it does over time.

the "i" is next once i get home.

 

[petrusbroder]

I have great difficulties uploading WUs - just now 184 WUs in the upload buffer ...
BOINC reports HTML-error ...

 

[Rudy Toody]

The Docking WUs have been sitting in the que for a day. I finally unchecked Docking in my projects tab at the site. That way they would stop sending more of them. They just came back online around the time you posted. I was out of work for about 3 hours.

 

[dajeepster]

ok... i'm now hyperapping on my "i"

 

[petrusbroder]

Still no upload --- "HTTP-Error"

 

[GLeeM]

Originally posted by: biodoc
I did read somewhere that the "i" series with HT on is much better at crunching mixed apps/projects. I'm not sure I understand this but that's why I decided to try hyperapping and monoapping.

I'm just trying to figure out why the "i" series is 2-3x better than the other chips on this project.

If there is no difference with or without HT, then the impressive performance of the "i" series must be due to caching or something else I guess.

Back in the P4 HT days on F@H I used to be able to get up to 35% better ppd with two certain different kinds of WUs. If I ran two of the same kind of WUs together I would get <10% better ppd out of HT. It was explained that the different WUs used different parts of the CPU (SSE, FPU, etc.).

So I think it is the mixed apps/projects that make part of the difference anyway.

 

[biodoc]

Originally posted by: GLeeM

Originally posted by: biodoc
I did read somewhere that the "i" series with HT on is much better at crunching mixed apps/projects. I'm not sure I understand this but that's why I decided to try hyperapping and monoapping.

I'm just trying to figure out why the "i" series is 2-3x better than the other chips on this project.

If there is no difference with or without HT, then the impressive performance of the "i" series must be due to caching or something else I guess.

Back in the P4 HT days on F@H I used to be able to get up to 35% better ppd with two certain different kinds of WUs. If I ran two of the same kind of WUs together I would get <10% better ppd out of HT. It was explained that the different WUs used different parts of the CPU (SSE, FPU, etc.).

So I think it is the mixed apps/projects that make part of the difference anyway.

I think you're right. I found an app (amiloide docking) that seem to run quite fast on 64 bit linux so I'll stick with that for the remainder of the race.

Now if I could get more work!!!

 

[petrusbroder]

I have - as said in the TAS-furom - collected the final stats. However I am out of town and can not access the data from away. I'll post the stats when I come back tomorrow night.

Please vote in the TAS-forums!

 

[petrusbroder]

Stats after the race:

Computer ___ Credits ___ Credits ____ Diff ___ TAS-mate
_________ __ Oct 20 ____ Oct 31 _______________________
_119045 ____ 12 304 ____ 39 669 __ 27 365 __ dajeepster :awe:
_119121 _____ 4 289 ____ 28 067 __ 23 778 __ Petrusbroder
_118825 ____ 52 976 ____ 71 795 __ 18 819 __ Rudy Toody :biggrin:
_119225 _____ 4 055 ____ 18 004 __ 13 949 __ biodoc ()
_119141 _____ 4 676 ____ 17 494 __ 12 818 __ Neurodog
_118989 _____ 4 210 ____ 15 163 __ 10 953 __ Philippart
_119497 ____ 17 401 ____ 24 849 ___ 7 448 __ Salvorhardin
Total _______ 99 911 ___ 215 071 _ 115 160 __ TAS-total

 

[biodoc]

Great job everyone!

I'm kinda wondering why peter gave me the ()

Maybe a would have been more appropriate!!!:biggrin:

 

[Rudy Toody]

Thanks for the stats, Peter

 

[petrusbroder]

Great job everyone!

I'm kinda wondering why peter gave me the ()

Maybe a would have been more appropriate!!!:biggrin:

Becuase you are the biodoc in a biology project which require at least a doc (postdoc, PhD, MD) ...


I also wanted to test the emoticons ... more tests comin.... D: