Comments for A Random Walk http://larkin.nuclearwinter.com/blog From Disorder and Chaos, comes more Disorder and Chaos Sat, 09 Jul 2011 20:15:06 +0000 hourly 1 http://wordpress.org/?v=3.1 Comment on Droid 3 vs Droid 1 LCD Matrix by llowrey http://larkin.nuclearwinter.com/blog/index.php/2011/07/droid-3-vs-droid-1-lcd-matrix/comment-page-1/#comment-117 llowrey Sat, 09 Jul 2011 20:15:06 +0000 http://larkin.nuclearwinter.com/blog/?p=58#comment-117 I do feel that the Droid 3 LCD is inferior to the Droid 1/2 but not by much. I notice that solid colors appear grainy but not so much that I would return the phone. I think what makes PenTile survivable for the Droid 3 is that it is only 4" and not the 4.3" of the Droid X2. The 4.3" screen, which has the same "resolution," just stretches out the effect and makes it more obvious. I do feel that the Droid 3 LCD is inferior to the Droid 1/2 but not by much. I notice that solid colors appear grainy but not so much that I would return the phone. I think what makes PenTile survivable for the Droid 3 is that it is only 4″ and not the 4.3″ of the Droid X2. The 4.3″ screen, which has the same “resolution,” just stretches out the effect and makes it more obvious.

]]> Comment on Droid 3 vs Droid 1 LCD Matrix by Jerome http://larkin.nuclearwinter.com/blog/index.php/2011/07/droid-3-vs-droid-1-lcd-matrix/comment-page-1/#comment-116 Jerome Sat, 09 Jul 2011 19:39:11 +0000 http://larkin.nuclearwinter.com/blog/?p=58#comment-116 Do you feel that the screen of the Droid 3 is severely lacking then? Do you feel that the screen of the Droid 3 is severely lacking then?

]]> Comment on Powernow-k8 AMD Turbo Core Bug by Sordna http://larkin.nuclearwinter.com/blog/index.php/2010/05/powernow-k8-amd-turbo-core-bug/comment-page-1/#comment-78 Sordna Thu, 17 Mar 2011 16:53:27 +0000 http://larkin.nuclearwinter.com/blog/?p=34#comment-78 Thanks for the update! Also thanks for the tmpfs firefox link, very useful! Here are my results with the 2.5GHz 9850 Phenom X4 on 64bit xubuntu 10.10 (kernel 2.6.35) in mops/secs for asin/fibo respectively. I only ran it once and din't lock the cpu high, I left it kick up the speed on its own (it idles at 1.2Ghz). Multithread results across cores are averaged: 1 thread asin 22.2, fibo44 13.5 2 threads asin 22.2, fibo44 15.1 3 threads asin 22.0, fibo44 14.5 4 threads asin 21.7, fibo44 15.3 Thanks for the update! Also thanks for the tmpfs firefox link, very useful! Here are my results with the 2.5GHz 9850 Phenom X4 on 64bit xubuntu 10.10 (kernel 2.6.35) in mops/secs for asin/fibo respectively. I only ran it once and din’t lock the cpu high, I left it kick up the speed on its own (it idles at 1.2Ghz). Multithread results across cores are averaged:
1 thread asin 22.2, fibo44 13.5
2 threads asin 22.2, fibo44 15.1
3 threads asin 22.0, fibo44 14.5
4 threads asin 21.7, fibo44 15.3

]]> Comment on Powernow-k8 AMD Turbo Core Bug by llowrey http://larkin.nuclearwinter.com/blog/index.php/2010/05/powernow-k8-amd-turbo-core-bug/comment-page-1/#comment-77 llowrey Thu, 17 Mar 2011 15:41:29 +0000 http://larkin.nuclearwinter.com/blog/?p=34#comment-77 Ok, re-ran lbench. The 2 and 3 thread numbers look much better. Ok, re-ran lbench. The 2 and 3 thread numbers look much better.

]]> Comment on Powernow-k8 AMD Turbo Core Bug by Sordna http://larkin.nuclearwinter.com/blog/index.php/2010/05/powernow-k8-amd-turbo-core-bug/comment-page-1/#comment-76 Sordna Thu, 17 Mar 2011 14:45:11 +0000 http://larkin.nuclearwinter.com/blog/?p=34#comment-76 Well, I do see firefox and even flash player having multiple entries in the process table even with 1 tab open when a youtube video is playing (and a cpu applet shows all of them getting busy) which leads me to believe it does take advantage of multiple cores. Anyway, could you try lbench again? I'm wondering if the perceived speed for 2 to 3 threads has increased from the 3.3GHz you reported last year. Thanks! Well, I do see firefox and even flash player having multiple entries in the process table even with 1 tab open when a youtube video is playing (and a cpu applet shows all of them getting busy) which leads me to believe it does take advantage of multiple cores.
Anyway, could you try lbench again? I’m wondering if the perceived speed for 2 to 3 threads has increased from the 3.3GHz you reported last year.
Thanks!

]]> Comment on Powernow-k8 AMD Turbo Core Bug by llowrey http://larkin.nuclearwinter.com/blog/index.php/2010/05/powernow-k8-amd-turbo-core-bug/comment-page-1/#comment-75 llowrey Thu, 17 Mar 2011 13:52:32 +0000 http://larkin.nuclearwinter.com/blog/?p=34#comment-75 AMD's Turbo Core has been well supported for quite some time now. There is a new tool called cpupowerutil that purports to show Turbo Core statistics but I have not been able to validate it. I am running Fedora 14 x64. I'm not sure the extra two cores would help that much for firefox. For your particular usecase I would go for the faster cores of the X4 975. If you have memory to spare, I would first recommend <a href="https://wiki.archlinux.org/index.php/Speed-up_Firefox_using_tmpfs" rel="nofollow">running your firefox profile on tmpfs</a>. I would also give Google's Chrome a spin. I recently switched to Chrome from Firefox and haven't looked back. AMD’s Turbo Core has been well supported for quite some time now. There is a new tool called cpupowerutil that purports to show Turbo Core statistics but I have not been able to validate it.

I am running Fedora 14 x64. I’m not sure the extra two cores would help that much for firefox. For your particular usecase I would go for the faster cores of the X4 975. If you have memory to spare, I would first recommend running your firefox profile on tmpfs. I would also give Google’s Chrome a spin. I recently switched to Chrome from Firefox and haven’t looked back.

]]> Comment on Powernow-k8 AMD Turbo Core Bug by Sordna http://larkin.nuclearwinter.com/blog/index.php/2010/05/powernow-k8-amd-turbo-core-bug/comment-page-1/#comment-74 Sordna Thu, 17 Mar 2011 06:24:08 +0000 http://larkin.nuclearwinter.com/blog/?p=34#comment-74 Very nice article! What distro are you using? Also, since it's been a year, I'm wondering if you've tried lbench with a more recent distro / kernel version. Perhaps Turbo Core is handled better nowdays, please report! I'm actually debating whether to upgrade my old Phenom I 9850 (2.5 GHz) to a Phenom II X6 1090T (3.2 GHz) or the new X4 975 (3.6GHz). The application I'm mostly interested in speeding up is ... firefox (!) which kinda slows down with heavy multimedia pages, when playing some videos, or with lots of open tabs. I'm using 64bit Ubuntu 10.10 (soon to be 11.4) on an Asus M3N72-D mobo. What's your recommendation to my CPU dilemma? Very nice article!
What distro are you using? Also, since it’s been a year, I’m wondering if you’ve tried lbench with a more recent distro / kernel version. Perhaps Turbo Core is handled better nowdays, please report!
I’m actually debating whether to upgrade my old Phenom I 9850 (2.5 GHz) to a Phenom II X6 1090T (3.2 GHz) or the new X4 975 (3.6GHz). The application I’m mostly interested in speeding up is … firefox (!) which kinda slows down with heavy multimedia pages, when playing some videos, or with lots of open tabs. I’m using 64bit Ubuntu 10.10 (soon to be 11.4) on an Asus M3N72-D mobo. What’s your recommendation to my CPU dilemma?

]]>