Haha, a few days back, I tried installing the 2.6.28 kernel. It went well, except that the XServer wouldn't start up.
It crashed with the following report:
(EE) NVIDIA(0): Failed to load the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***
(EE) Screen(s) found, but none have a usable configuration
But then, I noticed that the nvidia-kernel I installed was 2.6.26-1 and so, needed to update that. Well, a kernel installed _not_ from the repository would obviously need a nvidia-kernel install _not_ from the repo. ;)
A google search a couple of clicks later and some information from here, I landed at the-> NVIDIA's UNIX Drivers Portal
Grabbed the 180.22 version, the latest version for Linux AMD64/EM64. Follow instructions (essentially, chmod +x and run as root). It said my kernel was not new enough and tries to download the latest by ftp. In my case, I refused and it continued fine. Btw, you can't install the driver while you're running X.
Note: You should have the nvidia-xconfig program for the installation to go on smoothly, otherwise you'll end up with another problem
Installation and a reboot later, this is how it looks. :)
Saturday, January 24, 2009
The 2.6.28 kernel and the missing NVIDIA module
Monday, January 19, 2009
NVIDIA on Linux, an update
Another in the sequence of "update" posts.
I must say I was wrong when I said "bloody Nvidia drivers for Linux" here
Well, after three more Debian installations, I found out that the problem was not with the Nvidia drivers but with what they do when they are installed.
Err, let me clarify. On your first boot, Debian has a particular xorg.conf that it follows till you install, let's say nvidia-kernel-173.14.09. When you install this, the xorg.conf is replaced by an effectively blank xorg.conf
A way out, you ask? Well, install nvidia-xconfig and run it as superuser (as root, or use su) it will rewrite your xorg.conf. Reboot.
To further fine tune your control over your Nvidia graphics card/chipset, install nvidia-settings. Incase of Ubuntu, it will show up in Applications>System Tools, but somehow doesn't do so on Debian. Well, sudo nvidia-settings in the terminal does the job. So, no more complaints.
Sorry, Nvidia.
Sunday, December 07, 2008
NVIDIA on Linux
:X or :@ I'm angry with the bloody Nvidia drivers for Linux. Anyway, this post is supposed to be helpful, not a rant.
I upgraded from Ubuntu 8.04 to Ubuntu 8.10 when it came out. Then, I installed nvidia-kernel-2.6.26 which was my default nvidia stuff on Hardy and lo, my X crashed. It tried restarting some n times, with no success. So well, what was the problem. I still dunno. But the solution was to install the package nvidia-kernel-173.14.09 The last two numbers will vary, though. Lucky that I found this while random browsing on a friend's computer.
That did it.
A few weeks later, I wanted to use Debian, so, well, I removed Intrepid and installed Etch (Debian 40r5). This time, I upgraded, and installed nvidia-kernel-173. And guess what, it crashed again. I tried removing it and installing nvidia-kernel-2.6.26 which worked on Ubuntu 8.04. Well, it did not work this time. The solution? nvidia-kernel-2.6.26-1-openvz-amd64
I have absolutely no idea what is going on. But this worked. It installed an additional kernel though. I now boot into a kernel which is identified by Debian GNU/Linux, kernel 2.6.26-1-openvz-amd64
Hope this will remain a note for me in the future, for further X crashes.
Update: Well, this was helpful, but a better realization was this
Wednesday, February 20, 2008
A Google blooper ,yet again
Wow, not even in my dreams did I imagine that my article on MRAM would turn out to be a search result for...
Ladies and Gentlemen, hold your breadth ...only Google can relate MRAM with the search term...
"find out who i will be with in the future"
C'mon Mr.GOOG, that's way too funny.
Long time ago, I thought that this was the problem only with Google Ads alongside the search results.(Refer: here)
Look out for more blooper updates, and comments from friends :)
Posted by
[NpoWEr]
at
10:24 PM
View Comments
Labels: google, report, webmasters
Wednesday, April 04, 2007
Intel's processors and the confusion: 32-bit or 64-bit
Hi.It has been long since a post had been made in this blog. Forgive my negligence, those few loyal readers I have (At least, I have).These days were hectic and I too many jobs on hand to be completed and completed well.
This post, was to just comment on Intel’s, well, negligence (!)?
Hmm, I shall start off then.
As we all know pretty well, AMD had released the first 64-bit processor for desktops and Intel, for a long long time laid its bet on its very-successful 32-bit architecture, well known as IA32. However, Intel, of late (well, at least I have come to know it very late) has realized its fallacy and now all (read ALL) of its processors are 32-bit, including those which were released as 32bit initially which includes Pentium 4,Pentium D, Pentium Extreme Edition, Celeron D ,Xeon and Core Duo line of processors.
AMD has long copied Intel’s designs in that passed 32-bit era of computing (you are allowed to disagree, as usual :D).But, it seems history has its irony and Intel HAD TO adopt AMD’s x86-64 architecture, which was the name AMD gave to its 64-bit processors. Intel renamed this Intel64 and careful note is to be taken to the fact that Intel64 is a completely different architecture from IA-64 and all other implementations of x86-64. IA-64 is the architecture used for the Itanium series of processors from Intel.
Intel has lost the crown of having the fastest processor model since the introduction of AMD Athlon X2 series of processors. However, on July 27, 2006, Intel released its Core 2 Duo series of processors based on the Conroe core(covered on October 27th,click here), and reclaimed its crown and further establishing its lead when it released its Core 2 Quad series(covered on November 2nd,click here).
However, I have stated that I would be talking about Intel’s negligence, although that was a word used rashly above. What I am cribbing about is that Intel, on the Technical Specification sheet does not mention explicitly, the much wanted ’64-bit’ word .All the sheet says is that the Core 2 Duo supports the Intel Extended Memory 64 technology what it calls Intel EM64.
If it takes a person a reading in a online encyclopedia to determine whether a processor is 32-bit or not, and the company’s website says nothing, then should I presume that Intel’s tying to hide something from its customers? Especially that there are not many 64-bit applications around for common use? Any other reasons?
And ,as usual ,my fascination with codenames continues.
Here are some for trivia:
Conroe: The core with which the initial Core 2 series of processors came out.
Conroe XE: The core for Core 2 Extreme Edition.
Allendale: The cores which the present Core 2 processors had.
Kentsfield: Codename for Core 2 Quad.
Woodcrest: Codename for the server side version of Core 2, marketed as the well known Xeon.
Merom : The mobile version of Core 2.
Penryn: The successor to Merom (this will have a 45nm core.All others have 65nm cores)
Posted by
[NpoWEr]
at
1:16 AM
View Comments
Labels: core 2, intel, processors, report
Monday, March 19, 2007
The new quilt and a late post :)
My customization-of-the-template spree continues and a new widget gives its entry - The Quilt.
First,an introduction to the concept of scrobbling, or more precisely as audioscrobbling.Each and every song you listen to is noted down an a database of it is maintained online.This is called 'scrobbling'.And your 'charts' are generated.They might be weekly,monthly or even yearly.This is a way of analyzing your music habits.As time passes, you come to know of your tastes and how they have undergone changes as you discover new artists and songs.Like how Enter Sandman could make Linkin Park go down to Metallica in three days in my case ;) .
And, here is where Last.fm enters the scene.Last.fm© is primarily a Internet radio station and music recommendation system.
Last.fm provides a software that you can install on your computer that would scrobble all the songs that you play on your media player.Ofcourse you also have to install a Audioscrobbler© plugin for that media player.The best part is that it will scrobble even when you are offline.It will write all data to a file and then upload it when you are connected.
Now,the extras:
Last.fm provides nice goodies in the form of code snippets that you can embed on your webpage/blog. This includes live RSS feed of your played songs, and a nice widget called 'quilt' which consists of small thumbnails of all the albums/artists you've hears to.And again ,the choice between albums or artists is yours.
This was just intended to give you a little introduction to the world of audioscrobbling and the world's largest social music platform, Last.fm.
Find out more about Last.fm at Wikipedia and the site itself.
And,in case you join Last.fm ,add me as a friend :).
My profile at Last.fm
Posted by
[NpoWEr]
at
5:43 PM
View Comments
Labels: audioscrobbler, customize, last.fm, report, review, template
Sunday, March 18, 2007
Suzanne Shell and the Spiders on the Web
A woman by the name Suzanne Shell has sued Archive.org for the spiders crawling her site. The Colorado woman says she posted a notice on the site saying it she doesn't allow the site to be crawled.She sued for civil theft, breach of contract, and violations of the Racketeering Influence and Corrupt Organizations act(hell, what are they, never heard of them, after all she is an American ...phew).However, this is a late report by me and a court ruling last month dismissed all charges on Archive.com , except for the breach of contract claim. If Shell prevails on that claim, sites like Google will have to get online publishers to 'opt-in' before it can crawl their sites.
It seems it is more of her mistake.As a web-designer or an owner ,she shud've known all the ways to prevent crawling of the site.One of the important thing is to put up a 'robots.txt' file which she didn't. And it is too foolish to argue that the notice shudve served the purpose when it is well known that spiders don't understand the text which they crawl.
The case seems to be completely stupid.If she wanted to protect the page from being crawled , why didn't hse obtain a security certificate for it and why didnt she password protect it giving IDs to people who want to visit the site.Once a page is put up on the Web as a 'public' page ,which a page is by default, there is no use saying it is not intended for public.It is akin to a book once published and sold and then banned.
And what was the Internet for? When she launched the site without password protection, SHE ,infact had ignored the rules of the Net.
If I owned Archive.com , I wud've looked up the rules and sued her instead(ummm,they already know the rules...I'm waiting for the lawsuit to come out. American style :P). And if Miss Shell (or her lawyer,btw) was any more sensible ,they wudve withdrawn the case for the simple reason that there was no robots.txt instead of dropping 'shell's .It is simply the way the system works. Ignorance is no excuse.
I know it too, the hard way though.If she wins the case, it would only result in the emergence of new 'shell's to throw more 'shell's at such companies which crawl the Net.
Posted by
[NpoWEr]
at
4:03 AM
View Comments
Labels: Archive.com, report, slashdot, Suzanne Shell