Posts by Snow Crash
log in
1) Message boards : Cafe : The Last One to Post Wins! XXII (Message 22479)
Posted 603 days ago by Snow Crash
<--- intruder alert
<--- winning :-)
2) Message boards : Cafe : The Last One To Post Wins! V15 (Message 17918)
Posted 1517 days ago by Snow Crash
winning with many 'pending validation' - like money in the bank $
3) Message boards : Cafe : The Last One To Post Wins! V15 (Message 17912)
Posted 1517 days ago by Snow Crash
<--- winning ... my wife's baking cookies!!!
4) Message boards : News : Up and Running Again (Message 17204)
Posted 1638 days ago by Snow Crash
Slicker - the following where marked as "permanent error"

/collatz/download/110/solo_collatz_2380256540524033059176_824633720832
/collatz/download/255/solo_collatz_2380257177965899262312_824633720832
/collatz/download/116/solo_collatz_2380264727487613479272_824633720832
/collatz/download/19c/solo_collatz_2380264734909316966760_824633720832
/collatz/download/1ba/solo_collatz_2380264748412694145384_824633720832
/collatz/download/15/solo_collatz_2380277795870504364392_824633720832
/collatz/download/69/solo_collatz_2380291013718257150312_824633720832
5) Message boards : News : Up and Running Again (Message 17201)
Posted 1638 days ago by Snow Crash
it would really help if you could add the following to the log options in the cc_config.xml file so that you can report the entire path of each file as well as the specific errors occurring.

<http_debug>1</http_debug>
<http_xfer_debug>1</http_xfer_debug>


<cc_config>
<log_flags>
<http_debug>1</http_debug>
<http_xfer_debug>1</http_xfer_debug>
</log_flags>
</cc_config>


I apologize in advance if this is more than you need, I suspect it is on the GET line that is important.

GET /collatz/download/110/solo_collatz_2380256540524033059176_824633720832 HTTP/1.1

7/29/2013 2:45:48 AM | Collatz Conjecture | Started download of solo_collatz_2380256540524033059176_824633720832
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Info: Re-using existing connection! (#1) with host (nil)
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Info: Connected to (nil) (67.163.87.64) port 80 (#1)
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: GET /collatz/download/110/solo_collatz_2380256540524033059176_824633720832 HTTP/1.1
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: User-Agent: BOINC client (windows_x86_64 7.2.5)
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: Host: boinc.thesonntags.com
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: Accept: */*
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: Accept-Encoding: deflate, gzip
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server: Content-Type: application/x-www-form-urlencoded
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Sent header to server:
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: HTTP/1.1 404 Not Found
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Date: Mon, 29 Jul 2013 06:47:38 GMT
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Server: Apache/2.2.22 (Debian)
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Vary: Accept-Encoding
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Content-Encoding: gzip
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Content-Length: 289
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server: Content-Type: text/html; charset=iso-8859-1
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Received header from server:
7/29/2013 2:45:49 AM | | [http_xfer] [ID#9455] HTTP: wrote 354 bytes
7/29/2013 2:45:49 AM | Collatz Conjecture | [http] [ID#9455] Info: Connection #1 to host (nil) left intact
7/29/2013 2:45:50 AM | Collatz Conjecture | Giving up on download of solo_collatz_2380256540524033059176_824633720832: permanent HTTP error
6) Message boards : News : Up and Running Again (Message 17162)
Posted 1639 days ago by Snow Crash
If this persists I'm guessing WUs will start being server dropped due to "too many errors"
7) Message boards : Number crunching : What's your bet for GTX470/480 performance at Collatz? (Message 7568)
Posted 2827 days ago by Snow Crash
As the overwhelming majority of yours are running either side of 7 minutes with a 480, are you saying that you are finding a 30%+ speed increase from +/- 7mins to 5mins per WU with a Fermi 480 by turning off the HT?

Moving on from that, you obviously found a severe limitation in turning off HT to ignore such a huge GPU advantage of 30%+, as you currently run either side of 7mins per WU, what was the issue that made you back off?

Regards
Zy



Honestly, I participate in a cpu centric project that is my personal top priority. In the spirit of disclosure, I am only using this project to get a feel for my gpu computing capabilities and the limits of my card (volts, freq, fan, temps) while I wait for my gpu project of choice to get up to speed with fermi. That being said, while I am here I decided to participate as best I can. It looked like there was limited real crunching results being posted so I thought I would add my experiences to the general body of knowledge.
8) Message boards : Number crunching : What's your bet for GTX470/480 performance at Collatz? (Message 7543)
Posted 2827 days ago by Snow Crash
just over 5 minutes
9) Message boards : Number crunching : What's your bet for GTX470/480 performance at Collatz? (Message 7529)
Posted 2828 days ago by Snow Crash
Does the ATI version of the app use almost 1 to 1 ratio with CPU usage or am I reading other people's results incorrectly?
10) Message boards : Number crunching : What's your bet for GTX470/480 performance at Collatz? (Message 7520)
Posted 2828 days ago by Snow Crash
I see about a two minute difference per WU on tasks that take less than ten minutes ... I would classify that as a CHANGE :-)

With HT on, CPU loaded, GPU usage is ~85%
With HT on, CPU unloaded, GPU usage is 92-93%
With HT off, CPU unloaded, GPU usage is 95-96% (pretty amazing, really tight code!!!)

This is about the same GPU usage percentage behavior shifts I found testing GPUGrid (I used a GTX295 on GPUGrid) so it looks to me that once you get to the high end GPUs you really are CPU bound.

While CPU usage is very low (more kudos to Gispel) but switching contexts, threads, etc. takes time and it really does slow down your GPU processing.

Anyone else see a CHANGE between HT on/ off?
How about for CPU loaded/ unloaded CPU?

Steve
11) Message boards : Number crunching : What's your bet for GTX470/480 performance at Collatz? (Message 7498)
Posted 2829 days ago by Snow Crash
GTX480: Shaders at 1712 (stock volts)
i7-920 HT ON @4.2 GHz
Win7 64 Home Premium
BOINC 10.6.43
Fan at 70%, temp at 70c, ambient 22-23c

One WU takes about 7 minutes 30 seconds.


Yesterday I had the CPU with HT off, cranked up to 4.5 GHz, shaders at 1732 (1803 a little shakey, 1823 crashes) and was processing in a hair over 5 minutes.

Has anyone else notcied we are CPU bound?
12) Message boards : Number crunching : nvidia cuda (Message 7496)
Posted 2829 days ago by Snow Crash
sorry, wrong thread




Main page · Your account · Message boards


Copyright © 2018 Jon Sonntag; All rights reserved.