Posts by -= Vyper =-


log in
41) Message boards : Number crunching : CLOSED- SETI/BOINC Milestones [ v2.0 ] - XXII -CLOSED (Message 1087886)
Posted 17 Mar 2011 by Profile -= Vyper =-Project donor
Thank you everybody!

//Vyper
42) Message boards : Number crunching : CLOSED- SETI/BOINC Milestones [ v2.0 ] - XXII -CLOSED (Message 1087582)
Posted 16 Mar 2011 by Profile -= Vyper =-Project donor
OMGZ!

I finally passed 100M Credits..
Thought that i would never get there in my lifetime for four years ago and now i'm there..

*happy*

Kind regards Vyper
43) Message boards : Cafe SETI : Earthquake and Tsunami hits Japan. (Message 1087409)
Posted 15 Mar 2011 by Profile -= Vyper =-Project donor
Tragedy!

Prayers goes out to all of you people out there!
http://www.bspcn.com/2011/03/14/the-most-tragic-picture-i-have-seen-from-japan/

Kind regards Vyper
44) Message boards : Number crunching : Multi core greater than 80 core (Message 1085286)
Posted 9 Mar 2011 by Profile -= Vyper =-Project donor
3/9/2011 12:55:01 AM Processor: 60 GenuineIntel Intel(R) Xeon(R) CPU E7- 2870 @ 2.40GHz [Family 6 Model 47 Stepping 2]


Boinc only seems to see 60 (cores physical as virtual) cores and thus cannot adress more cores than this! Whatever change you do if that numbers show others then you would most ceratinly succeed in utilizing more cores in the system when finding that magical tweak.

//Vyper
45) Message boards : Number crunching : Multi core greater than 80 core (Message 1085285)
Posted 9 Mar 2011 by Profile -= Vyper =-Project donor
Hmm

I have an idea, perhaps it's a limit in boinc to only handle 64 subprocesses and it can't load more cores by itself.

If you take a read at my blog i have a workaround that could suit the need to fully utilize that system.

Unfortunately the hosts get separate "hosts id" though they are running in the same machine but you can easily merge the numbers you get from the system to see the total RAC of that system.

Look here for the principals to run more than one boinc installation in one system.
http://vyper.kafit.se/wp/index.php/2011/02/04/running-different-nvidia-architectures-most-optimal-at-setihome/

Kind regards Vyper
46) Message boards : Number crunching : My God, why hasn't nVidia noticed? (Message 1082876)
Posted 2 Mar 2011 by Profile -= Vyper =-Project donor
nVidia HAS noticed.

The first s@h app if i don't recall wrong nvidia helped Berkeley in incorporating and they were quite fast with that *hrm*
Only about two years after the first CUDA 8800 GT* generation arrived late 2006 :)

P.S Hmm with that beeing said, we need some sponsors for showing the real deal of GPUs , the electricity for running those isn't cheap at all in my country. The cost is nearly 2,5 GTX 580's in price / year just powering it on for my top rig D.S

//Vyper
47) Message boards : Number crunching : OMG......I did it. (Message 1082441)
Posted 28 Feb 2011 by Profile -= Vyper =-Project donor
Hehe, why not.. A ban counter :) ..
I wonder what the numbers would be at that counter.. :P
As we all point out, hurry back Mark.

Kind regards Vyper
48) Message boards : Cafe SETI : Say Something About The Person Before You (Message 1081694)
Posted 26 Feb 2011 by Profile -= Vyper =-Project donor
The person before me is a inspiration and driving source to keep us motivated to donate to our common beloved project.
Kudos to that.

P.S Yes i donated cash to Synergy but i don't got a star anyway. I know it in my heart. D.S
49) Message boards : Number crunching : ok, this is unusual (Message 1080584)
Posted 23 Feb 2011 by Profile -= Vyper =-Project donor
Well

For some reason the Cuda part of the driver and/or PhysX isn't enabled, the quickest way to fix this is to only download the newest driver from nvidias homepage and install it and you should be set to go again.
Why this has occured i don't know but this it what it seems like.
Start with that and report back if it solved your issue or not.
You could use the archived driver 197.3x to install over to maintain the highest RAC possible.

//Vyper
50) Message boards : Number crunching : And we're back!! (Message 1077606)
Posted 15 Feb 2011 by Profile -= Vyper =-Project donor
It's alive!

Well, lets rephrase this with a typical Swedish sentence: "Nu kör vi!!" :)

//Vyper
51) Message boards : Number crunching : 2 or 3 WU's On A GTX 470 (Message 1075780)
Posted 10 Feb 2011 by Profile -= Vyper =-Project donor
There is no definite answer to this one.

But we can presume this would be the most accurate, run 2 WU's if you run the card at stock clocks and if you OC your graphics card you can try to run 3 Wu's.

What you can test if you are into OC'ing is to see if PCI-E speed has a remarkable impact on crunching speed.

For instance if you run three WU's at once in a OC configuration on the 470 you would perhaps get around 23 to 24 minutes for a bunch of 0.4x AR wu's..
Then go to the bios and try to set PCI-E speed to 110 Mhz instead of 100 (Don't set it much higher than that) and see if you lower your times to perhaps 22,3 - 23,3 minutes for the same 0.4x AR wu's..

If the difference is negliable then set it back to 100 because then you've made a proof of concept test that shows that setting the PCI-E speed gives you a speed boost to the GPU operations or perhaps not.

Hope you get some ideas of direction for a better throughput.

Kind regards Vyper
52) Message boards : Number crunching : Combine gtx275 & gtx570 in one system? (Message 1074349)
Posted 5 Feb 2011 by Profile -= Vyper =-Project donor
I wouldn't dare to do it!

Atleast 750W to be on the safe side even if the powerdraw is about 500W when running full speed.
Go for a Corsair or Thermaltake PSU , choose one which have atleast 2 x PCI-E 8 connectors and 2 x PCI-E 6 connectors to be absolutely sure.

//Vyper
53) Message boards : Number crunching : Running mixed nvidia hardware most optimal on the same host (Message 1074045)
Posted 4 Feb 2011 by Profile -= Vyper =-Project donor
I had struggles getting the most out of my system when running a GTX285 and a GTX480 in the same system.
The 285 decreased it's productivity when running two WU's at once compared to the GTX480 which loved it.
In a guide on my blog i made two separate boinc work folders with separate executables to suite my needs to increase the throughput on that host.
Head over to see screenshoots and a small guide.

With this way i could set my GTX285 to use a specific executable and run only 1WU at once and the 480 ran another executable and two at once instead.

http://vyper.kafit.se

Kind regards Vyper

P.S In my example i had a 295 but most people know the drill anyhow D.S
54) Message boards : Number crunching : Virtual Machine and CUDA Questions??? (Message 1070784)
Posted 26 Jan 2011 by Profile -= Vyper =-Project donor
I'm sending a "Shutting Up" to myself.
Thanks for digging BeNt..

//Vyper
55) Message boards : Number crunching : Virtual Machine and CUDA Questions??? (Message 1070750)
Posted 26 Jan 2011 by Profile -= Vyper =-Project donor
Q1: I'm not really sure but i doubt it. Don't think noone has invented graphics bypass directly to PCI-e adresses.

Q2: No! Again if someone/company hasn't invented a way of running 64bit hosts on a 32bit system. It is possible if you have 64bit all the way.
example:

HOST-HYPERVISOR-VIRTUAL MACHINE:

You need to have a:
64-64-64 configuration to make it work.

In your case you have:
64-32-64 and that is not possible from what i know.

Kind regards Vyper

P.S Could you please answer personal messages that i have wrote to you? D.S
56) Message boards : Cafe SETI : A bit of better news for the kittyman. (Message 1068503)
Posted 20 Jan 2011 by Profile -= Vyper =-Project donor
Mark!

Do you know how often your powercompany "polls" your powerline to measure how much KWh you are using at that particular moment or do they simply collect the data once a month?

If so, you cannot choose a method of poweroptimizing your electricity in a particular timeframe more for one month to the next.
The only thing for certain is that you lower the amount of powerdraw for that particular timeframe that your powercompany measure you.

Example once a month is perhaps 1500KWh regular usage for you, if you turn it off during a special period that sums up to for instance a day a week that equals to about 1288KWh a month of consumed electricity.

When your powercompany then sends a bill they sum it up to 1288KWh equals that months average powercost if you have a non-fixed rate.

That means that how much you try to optimize and turn of during prime times the average for that month wouldn't change a single digit at all , and if you turn it off during the primetime the only thing you would do is reducing the cost for what your powercompany needs to pay for purchasing electricity for themselves to deliver that to you because their powerpolling is not once a month it perhaps polls every minute or so.
It would only increase their own revenues.

So if you want to lower a powercompanys revenues , turn the computer off during the cheaper period and only run it at the expensive peaks instead :P ..
Because as i said they got polled and billed more compared to you as customer for your powercompany.

This what it "seems" to be the way it works here in Sweden until i've heard otherwise from other safe sources.

Kind regards Vyper
57) Message boards : Number crunching : GPU Users Group Server Donation (Message 1065732)
Posted 11 Jan 2011 by Profile -= Vyper =-Project donor
Well, our server Synergy is now up and running.
Hope we all can make use of it and further stabilize the infrastructure further!

Thank you Todd for your time and effort.

Kind regards Vyper
58) Message boards : Number crunching : PCIe speed and CUDA performance (Message 1064992)
Posted 9 Jan 2011 by Profile -= Vyper =-Project donor
Thank you , thank you.

I'm not sure that i'm 100% correct in what i describe but atleast that gives a small hum of what is going on in terms of what happens when involving different parts of your system.
Everything that eventually can be precalculated or expanded to a easy to follow grid pointer system to make sure that least amount of data possible needed to travel through the slow PCI-E bus is almost certainly a win-win situation.
Cpu can do other work along with gpu not needing to be fed with "what now then" parameters, and if that is not avoidable it simply isn't.

Simply said i presume the system do it's best if as much preparation of data and code is made before the transfers occur to the gpu.

I just couldn't stop myself from making something human referable to what happens inside the computer system at that time of writing.

Kind regards Vyper
59) Message boards : Number crunching : PCIe speed and CUDA performance (Message 1064944)
Posted 9 Jan 2011 by Profile -= Vyper =-Project donor
Don't know man..

Perhaps around half an hour or so. I had it all in my mind.
No. I have not taken drugs :P

Kind regards Vyper
60) Message boards : Number crunching : PCIe speed and CUDA performance (Message 1064934)
Posted 9 Jan 2011 by Profile -= Vyper =-Project donor
Man!

I've should've been a novelist instead:

http://www.imdb.com/title/tt0060196/
=
The CPU, The GPU & The coder

:)

Think twice before you click and send the driver on it's way!

Regards Vyper


Previous 20 · Next 20

Copyright © 2014 University of California