PCLinuxOS-Forums
News: ...FLASH!!! ...New PCLinuxOS Testing board now open. Register today! Be an active contributor to the PCLinuxOS future! ... Read all about it now, on THIS forum!!!..
 
*
Welcome, Guest. Please login or register. May 27, 2012, 02:54:57 PM


Login with username, password and session length


Pages: [1]   Go Down
  Print  
Author Topic: Thanks for the kernel  (Read 316 times)
djohnston
PCLinuxOS Tester
Hero Member
*******
Offline Offline

Posts: 5710


I don't do Windows


« on: July 07, 2011, 03:57:35 AM »

Some people may know that I've been running folding@home on all my home computers for a few years now. I see it as a way to further disease research, as well as a quiet way of promoting the abilities and stability of PCLinuxOS. The team advertisement doesn't hurt one bit! But, this is about a kernel, isn't it?

I have two AMD Athlon64-equipped PCs, purchased from LinPC. On both, I opted for the 4GB RAM upgrade. When the BIOS is starting up, one can see the 4GB of memory displayed prominently. Until recently, I had been using one of Tex's a64 kernels. Along with the bfs kernel and Redmond's OS, each can only "see" 3GB of RAM when the OS is running. Recently, Tex built (someone correct me if I'm wrong) the first pae.bfs kernel. This kernel, due to the physical address extension, allows the OS to "see" all 4GB of RAM on both machines. The bfs portion is Con Kolivas's contribution to kernel task scheduling, tailored for today's desktops.

Tex brought the magic of both enhancements to the Linux kernel into one package. The runtime parameters for each of my folding@home clients has remained unchanged. I've always requested advanced computing units, and the largest ones available. Because of the "advanced" request, I sometimes receive experimental work units that will crash and burn at some point in the run. When they do, they will most often restart the same unit over again, then crash and burn at the same point, in an endless loop. That is, until I discover one and force the submittal of another unit to work on. All of the time used on the failed units count for nothing, in terms of points. Worse still, time is wasted where I could have been accumulating more points. For a couple of months, my points per day average would yo-yo back and forth between a high and low level.

Here's the "magic" I speak of. When running the a64 kernel, I was sometimes in the mid 800s for a points per day average high. Since switching to the pae.bfs kernel, my current points per day average is exactly 2,522. And it keeps climbing higher every day. That's a 200% increase! And all I did was change to a different kernel! Now I admit that the increase in available RAM may be bringing me different work units than before. But, I have also run just the "plain Jane" pae kernel before, without this sudden jump in productivity. And, the work units I'm currently receiving are worth more points per dhrystone than others are. Most are 1.5 million steps or more. So, the combination of RAM increase, more complicated work units and the kernel task scheduler are giving me three times as many points in a 24 hour period than before. Just sayin'.

Texstar, you've still got the magic touch!
Logged

Bare metal                           VBox
AMD Athlon 7750 Dual-Core    Single core
4GiB RAM                              1GiB RAM
nVidia GeForce FX 5200          64MB video
LXDE 32bit                            KDE 64bit

Registered Linux User #416378
OMSkates
Hero Member
*****
Offline Offline

Posts: 2265



« Reply #1 on: July 07, 2011, 04:13:30 AM »

That's a tremendous increase due to just those few changes.  You say something about furthering research in first paragraph, not sure if I understand unless that's a typo. Thanks for sharing Grin
Logged
djohnston
PCLinuxOS Tester
Hero Member
*******
Offline Offline

Posts: 5710


I don't do Windows


« Reply #2 on: July 07, 2011, 04:28:10 AM »


That's a tremendous increase due to just those few changes.  You say something about furthering research in first paragraph, not sure if I understand unless that's a typo. Thanks for sharing Grin


Quote

I see it as a way to further disease research


From the folding@home website's main page: http://folding.stanford.edu/

Protein folding is linked to disease, such as Alzheimer's, ALS, Huntington's, Parkinson's disease, and many Cancers. Moreover, when proteins do not fold correctly (i.e. "misfold"), there can be serious consequences, including many well known diseases, such as Alzheimer's, Mad Cow (BSE), CJD, ALS, Huntington's, Parkinson's disease, and many Cancers and cancer-related syndromes.
Logged

Bare metal                           VBox
AMD Athlon 7750 Dual-Core    Single core
4GiB RAM                              1GiB RAM
nVidia GeForce FX 5200          64MB video
LXDE 32bit                            KDE 64bit

Registered Linux User #416378
OMSkates
Hero Member
*****
Offline Offline

Posts: 2265



« Reply #3 on: July 07, 2011, 04:42:55 AM »

From the folding@home website's main page: http://folding.stanford.edu/

Protein folding is linked to disease, such as Alzheimer's, ALS, Huntington's, Parkinson's disease, and many Cancers. Moreover, when proteins do not fold correctly (i.e. "misfold"), there can be serious consequences, including many well known diseases, such as Alzheimer's, Mad Cow (BSE), CJD, ALS, Huntington's, Parkinson's disease, and many Cancers and cancer-related syndromes.


Ah, I see, thank you.  So its a simulation to understand the mis-folding that can cause these diseases.   I will have to look into this further.
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.16 | SMF © 2011, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM