Optimize your GPU. Find the value the easy way.


log in

Advanced search

Message boards : Number crunching : Optimize your GPU. Find the value the easy way.

Previous · 1 · 2 · 3 · 4 · 5 · 6 . . . 13 · Next
Author Message
Profile S@NL - eFMer - efmer.com/boincProject donor
Volunteer tester
Avatar
Send message
Joined: 7 Jun 99
Posts: 512
Credit: 130,864,061
RAC: 14,961
United States
Message 1276314 - Posted: 27 Aug 2012, 17:33:33 UTC - in response to Message 1276275.

Thanks Fred! Runs great with _g. Won't run with _z. Program failed to start on slot: 0

V 1.2 will be out very soon, tested with z.
____________
TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking.

Russ Sheldon
Avatar
Send message
Joined: 18 Jan 01
Posts: 379
Credit: 473,320,049
RAC: 185,186
United States
Message 1276330 - Posted: 27 Aug 2012, 17:48:36 UTC - in response to Message 1276314.

Thanks, will look for update.
Could you translate for me?
GTX570 btw

Starting automatic test: (lunatics_x41g_win32_cuda32_perf.exe)
27 August 2012 - 13:32:23 Start, devices: 2, device count: 1
27 August 2012 - 13:34:39 Runtime: Device: 0, count: 0, 134 seconds
27 August 2012 - 13:34:39 Device: 0, finished
27 August 2012 - 13:34:41 Runtime: Device: 1, count: 0, 136 seconds
27 August 2012 - 13:34:41 Device: 1, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 1, average time / count: 134, average time on device: 134 Seconds (2 Minutes, 14 Seconds)
Device: 1, device count: 1, average time / count: 136, average time on device: 136 Seconds (2 Minutes, 16 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 13:34:43 Start, devices: 2, device count: 2
27 August 2012 - 13:38:13 Runtime: Device: 0, count: 0, 207 seconds
27 August 2012 - 13:38:13 Device: 0, finished
27 August 2012 - 13:38:14 Runtime: Device: 0, count: 1, 208 seconds
27 August 2012 - 13:38:14 Device: 0, finished
27 August 2012 - 13:38:14 Runtime: Device: 1, count: 1, 208 seconds
27 August 2012 - 13:38:14 Device: 1, finished
27 August 2012 - 13:38:15 Runtime: Device: 1, count: 0, 209 seconds
27 August 2012 - 13:38:15 Device: 1, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 2, average time / count: 207, average time on device: 103 Seconds (1 Minutes, 43 Seconds)
Device: 1, device count: 2, average time / count: 208, average time on device: 104 Seconds (1 Minutes, 44 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 13:38:16 Start, devices: 2, device count: 3
27 August 2012 - 13:43:24 Runtime: Device: 0, count: 1, 304 seconds
27 August 2012 - 13:43:24 Device: 0, finished
27 August 2012 - 13:43:24 Runtime: Device: 0, count: 2, 304 seconds
27 August 2012 - 13:43:24 Device: 0, finished
27 August 2012 - 13:43:24 Runtime: Device: 1, count: 0, 304 seconds
27 August 2012 - 13:43:24 Runtime: Device: 0, count: 0, 304 seconds
27 August 2012 - 13:43:24 Device: 1, finished
27 August 2012 - 13:43:24 Device: 0, finished
27 August 2012 - 13:43:25 Runtime: Device: 1, count: 1, 305 seconds
27 August 2012 - 13:43:25 Device: 1, finished
27 August 2012 - 13:43:26 Runtime: Device: 1, count: 2, 306 seconds
27 August 2012 - 13:43:26 Device: 1, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 3, average time / count: 304, average time on device: 101 Seconds (1 Minutes, 41 Seconds)
Device: 1, device count: 3, average time / count: 305, average time on device: 101 Seconds (1 Minutes, 41 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 13:43:27 Start, devices: 2, device count: 4

Is the difference between 2 and 3 enough to run 3?

Thanks Fred!

____________
“Grandpa, I heard a story about a guy named Russ who crunched over 700,000 RAC for an unappreciative SETI in Oct. 2012. Is that legend true?”
“Yes Billy, and some people say he still crunches. All we can do is watch the skies... and hope.”

Horacio
Send message
Joined: 14 Jan 00
Posts: 536
Credit: 75,843,180
RAC: 19,688
Argentina
Message 1276352 - Posted: 27 Aug 2012, 18:13:39 UTC - in response to Message 1276313.


Weird thing... my gtx680 seems to be slower than others, even OC'ed... And nothing else running on the host other than MSI afterburner and Process Laso to set the apps at normal priority...

Yep seems to be on the low side.
NVIDIA 3d setting->power management->max performance.

Tried that, but it gaves me the same results...
I guess I should investigate why is so slow... work for the weekend...
____________

juan BFBProject donor
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 5471
Credit: 313,434,305
RAC: 116,118
Brazil
Message 1276354 - Posted: 27 Aug 2012, 18:15:12 UTC - in response to Message 1276275.
Last modified: 27 Aug 2012, 18:16:08 UTC

Thanks Fred! Runs great with _g. Won't run with _z. Program failed to start on slot: 0


Russ

You need to go to the config.xml and change to:

<config>
<data>
<exe> Lunatics_x41z_win32_cuda42.exe </exe>
</data>
</config>


<!--

Make sure you open this file in a simple text editor, like notepad.exe only.
This file has to be stored in ascii format.

-->

Then go to the data directory (where you put the setiperformance program) and copy the exe and dll of the x41z to that directory.

Then it will run OK (i test here in 7 diferent hosts/GPUs configurations all running with x41z)
____________

Russ Sheldon
Avatar
Send message
Joined: 18 Jan 01
Posts: 379
Credit: 473,320,049
RAC: 185,186
United States
Message 1276384 - Posted: 27 Aug 2012, 19:10:53 UTC - in response to Message 1276354.

Juan,

Thanks, have already tried that. Still no workee.

____________
“Grandpa, I heard a story about a guy named Russ who crunched over 700,000 RAC for an unappreciative SETI in Oct. 2012. Is that legend true?”
“Yes Billy, and some people say he still crunches. All we can do is watch the skies... and hope.”

Bojan
Send message
Joined: 28 Feb 04
Posts: 14
Credit: 10,585,098
RAC: 0
Slovenia
Message 1276386 - Posted: 27 Aug 2012, 19:15:41 UTC - in response to Message 1276354.

i just want to see that anyone paste test with 660 ti.

thanks to that person

juan BFBProject donor
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 5471
Credit: 313,434,305
RAC: 116,118
Brazil
Message 1276403 - Posted: 27 Aug 2012, 19:47:42 UTC - in response to Message 1276384.

Juan,

Thanks, have already tried that. Still no workee.


Do you rememeber to put the right dll and exe (the ones from x41z) on the setiperformance directory? not in the boinc directory itself.

____________

Russ Sheldon
Avatar
Send message
Joined: 18 Jan 01
Posts: 379
Credit: 473,320,049
RAC: 185,186
United States
Message 1276407 - Posted: 27 Aug 2012, 19:59:48 UTC - in response to Message 1276403.

Sim.
____________
“Grandpa, I heard a story about a guy named Russ who crunched over 700,000 RAC for an unappreciative SETI in Oct. 2012. Is that legend true?”
“Yes Billy, and some people say he still crunches. All we can do is watch the skies... and hope.”

Russ Sheldon
Avatar
Send message
Joined: 18 Jan 01
Posts: 379
Credit: 473,320,049
RAC: 185,186
United States
Message 1276411 - Posted: 27 Aug 2012, 20:05:57 UTC - in response to Message 1276407.
Last modified: 27 Aug 2012, 20:07:38 UTC

Why would _g work fine and _z not?
ie:
<config>
<data>
<exe> Lunatics_x41z_win32_cuda42_perf.exe </exe>
</data>
</config>


<!--

Make sure you open this file in a simple text editor, like notepad.exe only.
This file has to be stored in ascii format.

-->
____________
“Grandpa, I heard a story about a guy named Russ who crunched over 700,000 RAC for an unappreciative SETI in Oct. 2012. Is that legend true?”
“Yes Billy, and some people say he still crunches. All we can do is watch the skies... and hope.”

Horacio
Send message
Joined: 14 Jan 00
Posts: 536
Credit: 75,843,180
RAC: 19,688
Argentina
Message 1276416 - Posted: 27 Aug 2012, 20:14:22 UTC - in response to Message 1276411.

Why would _g work fine and _z not?
ie:
<config>
<data>
<exe> Lunatics_x41z_win32_cuda42_perf.exe </exe>
</data>
</config>


<!--

Make sure you open this file in a simple text editor, like notepad.exe only.
This file has to be stored in ascii format.

-->

Its that the name of the executable file?, or have you renamed it to include the _perf part to the file name?
____________

juan BFBProject donor
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 5471
Credit: 313,434,305
RAC: 116,118
Brazil
Message 1276424 - Posted: 27 Aug 2012, 20:30:58 UTC - in response to Message 1276411.
Last modified: 27 Aug 2012, 20:32:49 UTC

Why would _g work fine and _z not?
ie:
<config>
<data>
<exe> Lunatics_x41z_win32_cuda42_perf.exe </exe>
</data>
</config>


<!--

Make sure you open this file in a simple text editor, like notepad.exe only.
This file has to be stored in ascii format.

-->


Russ

Check the name of the file, it must match with your windows version, the name i give is for XP 32 Only, i belive you have Win 7 64, so the name and dll must be different.

Sorry i don´t have any Win 7 64 here to help.

Thanks you have no problem with cuban cigars... hehehehe
____________

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4213
Credit: 34,474,732
RAC: 16,131
United Kingdom
Message 1276432 - Posted: 27 Aug 2012, 20:50:05 UTC - in response to Message 1276424.
Last modified: 27 Aug 2012, 21:09:18 UTC

Russ

Check the name of the file, it must match with your windows version, the name i give is for XP 32 Only, i belive you have Win 7 64, so the name and dll must be different.

Sorry i don´t have any Win 7 64 here to help.

There isn't a 64bit x41g or x41z app, just the 32bit variants,

I couldn't get it to work on my GTX460 with x41z app eithier, until i removed the x41g app and dll's, then it did work:

Starting automatic test: (lunatics_x41g_win32_cuda32_perf.exe)
27 August 2012 - 21:15:46 Start, devices: 1, device count: 1
27 August 2012 - 21:18:35 Runtime: Device: 0, count: 0, 168 seconds
27 August 2012 - 21:18:35 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 1, average time / count: 168, average time on device: 168 Seconds (2 Minutes, 48 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 21:18:35 Start, devices: 1, device count: 2
27 August 2012 - 21:23:49 Runtime: Device: 0, count: 1, 313 seconds
27 August 2012 - 21:23:49 Device: 0, finished
27 August 2012 - 21:23:53 Runtime: Device: 0, count: 0, 317 seconds
27 August 2012 - 21:23:53 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 2, average time / count: 315, average time on device: 157 Seconds (2 Minutes, 37 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 21:23:53 Start, devices: 1, device count: 3
27 August 2012 - 21:31:41 Runtime: Device: 0, count: 2, 467 seconds
27 August 2012 - 21:31:41 Device: 0, finished
27 August 2012 - 21:31:46 Runtime: Device: 0, count: 1, 472 seconds
27 August 2012 - 21:31:46 Device: 0, finished
27 August 2012 - 21:31:46 Runtime: Device: 0, count: 0, 471 seconds
27 August 2012 - 21:31:46 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 3, average time / count: 470, average time on device: 156 Seconds (2 Minutes, 36 Seconds)
The best average time found: 157 Seconds (2 Minutes, 37 Seconds), with count: 0.50 (2)

Starting automatic test: (lunatics_x41z_win32_cuda42.exe)
27 August 2012 - 21:36:23 Start, devices: 1, device count: 1
27 August 2012 - 21:38:42 Runtime: Device: 0, count: 0, 138 seconds
27 August 2012 - 21:38:42 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 1, average time / count: 138, average time on device: 138 Seconds (2 Minutes, 18 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 21:38:43 Start, devices: 1, device count: 2
27 August 2012 - 21:43:13 Runtime: Device: 0, count: 0, 269 seconds
27 August 2012 - 21:43:13 Runtime: Device: 0, count: 1, 269 seconds
27 August 2012 - 21:43:13 Device: 0, finished
27 August 2012 - 21:43:13 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 2, average time / count: 269, average time on device: 134 Seconds (2 Minutes, 14 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 21:43:15 Start, devices: 1, device count: 3
27 August 2012 - 21:49:54 Runtime: Device: 0, count: 0, 398 seconds
27 August 2012 - 21:49:54 Device: 0, finished
27 August 2012 - 21:49:57 Runtime: Device: 0, count: 2, 401 seconds
27 August 2012 - 21:49:57 Device: 0, finished
27 August 2012 - 21:49:57 Runtime: Device: 0, count: 1, 401 seconds
27 August 2012 - 21:49:57 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 3, average time / count: 400, average time on device: 133 Seconds (2 Minutes, 13 Seconds)
The best average time found: 134 Seconds (2 Minutes, 14 Seconds), with count: 0.50 (2)


Claggy

John
Send message
Joined: 21 May 99
Posts: 49
Credit: 5,135,827
RAC: 440
United States
Message 1276474 - Posted: 27 Aug 2012, 23:25:20 UTC

Finally got both the GTX 670 and Gtx 470 working
added cc_config file to C:\ProgramData\BOINC\cc_Config

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
</cc_config>

just copy the one in C:\ProgramData\BOINC\projects\setiathome.berkeley.edu
and paste to C:\ProgramData\BOINC\cc_Config

Will see what results i get
Gtx 670 reportsw 4 tasks is better but will need to the Gtx 470 to 2 or 3

Thanks for the great app

____________

Profile arkaynProject donor
Volunteer tester
Avatar
Send message
Joined: 14 May 99
Posts: 3725
Credit: 48,768,260
RAC: 1,737
United States
Message 1276505 - Posted: 28 Aug 2012, 1:11:07 UTC

EVGA GTX-670 at +80/+300

Starting automatic test: (lunatics_x41z_win32_cuda42.exe)
27 August 2012 - 17:52:35 Start, devices: 1, device count: 1
27 August 2012 - 17:54:33 Runtime: Device: 0, count: 0, 117 seconds
27 August 2012 - 17:54:33 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 1, average time / count: 117, average time on device: 117 Seconds (1 Minutes, 57 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 17:54:33 Start, devices: 1, device count: 2
27 August 2012 - 17:57:42 Runtime: Device: 0, count: 0, 188 seconds
27 August 2012 - 17:57:42 Device: 0, finished
27 August 2012 - 17:57:43 Runtime: Device: 0, count: 1, 189 seconds
27 August 2012 - 17:57:43 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 2, average time / count: 188, average time on device: 94 Seconds (1 Minutes, 34 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 17:57:44 Start, devices: 1, device count: 3
27 August 2012 - 18:02:19 Runtime: Device: 0, count: 0, 274 seconds
27 August 2012 - 18:02:19 Device: 0, finished
27 August 2012 - 18:02:21 Runtime: Device: 0, count: 1, 276 seconds
27 August 2012 - 18:02:21 Device: 0, finished
27 August 2012 - 18:02:21 Runtime: Device: 0, count: 2, 276 seconds
27 August 2012 - 18:02:21 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 3, average time / count: 275, average time on device: 91 Seconds (1 Minutes, 31 Seconds)
Next ---------------------------------------------------------------------
27 August 2012 - 18:02:22 Start, devices: 1, device count: 4
27 August 2012 - 18:08:29 Runtime: Device: 0, count: 3, 366 seconds
27 August 2012 - 18:08:29 Device: 0, finished
27 August 2012 - 18:08:29 Runtime: Device: 0, count: 0, 366 seconds
27 August 2012 - 18:08:29 Device: 0, finished
27 August 2012 - 18:08:30 Runtime: Device: 0, count: 2, 367 seconds
27 August 2012 - 18:08:30 Device: 0, finished
27 August 2012 - 18:08:31 Runtime: Device: 0, count: 1, 368 seconds
27 August 2012 - 18:08:31 Device: 0, finished
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 4, average time / count: 366, average time on device: 91 Seconds (1 Minutes, 31 Seconds)
The best average time found: 91 Seconds (1 Minutes, 31 Seconds), with count: 0.33 (3)
____________

Russ Sheldon
Avatar
Send message
Joined: 18 Jan 01
Posts: 379
Credit: 473,320,049
RAC: 185,186
United States
Message 1276538 - Posted: 28 Aug 2012, 4:59:03 UTC - in response to Message 1276411.

Why would _g work fine and _z not?
ie:
<config>
<data>
<exe> Lunatics_x41z_win32_cuda42_perf.exe </exe>
</data>
</config>


My point is that if _g perf.exe will run just fine on my machine running _z, why will not _z perf.exe work after I change the xml to _z etc.
_x41z, cudart32_42_9.dll and cufft32_42_9.dll are in the same directory. Still no likey?
____________
“Grandpa, I heard a story about a guy named Russ who crunched over 700,000 RAC for an unappreciative SETI in Oct. 2012. Is that legend true?”
“Yes Billy, and some people say he still crunches. All we can do is watch the skies... and hope.”

Profile arkaynProject donor
Volunteer tester
Avatar
Send message
Joined: 14 May 99
Posts: 3725
Credit: 48,768,260
RAC: 1,737
United States
Message 1276547 - Posted: 28 Aug 2012, 5:30:50 UTC - in response to Message 1276538.

Why would _g work fine and _z not?
ie:
<config>
<data>
<exe> Lunatics_x41z_win32_cuda42_perf.exe </exe>
</data>
</config>


My point is that if _g perf.exe will run just fine on my machine running _z, why will not _z perf.exe work after I change the xml to _z etc.
_x41z, cudart32_42_9.dll and cufft32_42_9.dll are in the same directory. Still no likey?


The app name should be Lunatics_x41z_win32_cuda42.exe, also make certain that only the wu, the app and the dll files are in that directory. Anything else will cause it to fail.
____________

Profile S@NL - eFMer - efmer.com/boincProject donor
Volunteer tester
Avatar
Send message
Joined: 7 Jun 99
Posts: 512
Credit: 130,864,061
RAC: 14,961
United States
Message 1276557 - Posted: 28 Aug 2012, 6:25:19 UTC - in response to Message 1276547.

A new version 1.1, to solve some bug and to add all the requests.

http://www.efmer.eu/forum_tt/index.php?topic=974.msg5946#msg5946
____________
TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking.

tbretProject donor
Volunteer tester
Avatar
Send message
Joined: 28 May 99
Posts: 2897
Credit: 218,381,475
RAC: 29,826
United States
Message 1276586 - Posted: 28 Aug 2012, 8:06:01 UTC
Last modified: 28 Aug 2012, 8:21:00 UTC

GTX660Ti (reference cards) Driver 305.68

Starting automatic test: (x41g)
28 August 2012 - 02:45:56 Start, devices: 2, device count: 1 (1.00)
28 August 2012 - 02:50:00 Runtime: Device: 1, count: 0, 242 seconds
28 August 2012 - 02:50:00 Device: 1, Count: 0, finished.
28 August 2012 - 02:50:09 Runtime: Device: 0, count: 0, 251 seconds
28 August 2012 - 02:50:09 Device: 0, Count: 0, finished.
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 1, average time / count: 251, average time on device: 251 Seconds (4 Minutes, 11 Seconds)
Device: 1, device count: 1, average time / count: 242, average time on device: 242 Seconds (4 Minutes, 2 Seconds)
Next ---------------------------------------------------------------------
28 August 2012 - 02:50:10 Start, devices: 2, device count: 2 (0.50)
28 August 2012 - 02:55:39 Runtime: Device: 0, count: 0, 328 seconds
28 August 2012 - 02:55:39 Device: 0, Count: 0, finished.
28 August 2012 - 02:55:40 Runtime: Device: 0, count: 1, 329 seconds
28 August 2012 - 02:55:40 Device: 0, Count: 1, finished.
28 August 2012 - 02:55:40 Runtime: Device: 1, count: 1, 329 seconds
28 August 2012 - 02:55:40 Device: 1, Count: 1, finished.
28 August 2012 - 02:55:40 Runtime: Device: 1, count: 0, 329 seconds
28 August 2012 - 02:55:40 Device: 1, Count: 0, finished.
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 2, average time / count: 328, average time on device: 164 Seconds (2 Minutes, 44 Seconds)
Device: 1, device count: 2, average time / count: 329, average time on device: 164 Seconds (2 Minutes, 44 Seconds)
Next ---------------------------------------------------------------------
28 August 2012 - 02:55:42 Start, devices: 2, device count: 3 (0.33)
28 August 2012 - 03:03:35 Runtime: Device: 0, count: 0, 472 seconds
28 August 2012 - 03:03:35 Device: 0, Count: 0, finished.
28 August 2012 - 03:03:41 Runtime: Device: 1, count: 2, 478 seconds
28 August 2012 - 03:03:41 Device: 1, Count: 2, finished.
28 August 2012 - 03:03:41 Runtime: Device: 0, count: 2, 478 seconds
28 August 2012 - 03:03:41 Device: 0, Count: 2, finished.
28 August 2012 - 03:03:42 Runtime: Device: 0, count: 1, 479 seconds
28 August 2012 - 03:03:42 Device: 0, Count: 1, finished.
28 August 2012 - 03:03:42 Runtime: Device: 1, count: 1, 479 seconds
28 August 2012 - 03:03:42 Device: 1, Count: 1, finished.
28 August 2012 - 03:03:42 Runtime: Device: 1, count: 0, 479 seconds
28 August 2012 - 03:03:42 Device: 1, Count: 0, finished.
Ready ---------------------------------------------------------------------
Results:
Device: 0, device count: 3, average time / count: 476, average time on device: 158 Seconds (2 Minutes, 38 Seconds)
Device: 1, device count: 3, average time / count: 478, average time on device: 159 Seconds (2 Minutes, 39 Seconds)

EDIT: I can't get 41z to work. I'm sure I'm just doing something dumb, but I'm too tired to see it.

During the 1 work-unit per card test, GPU-Z showed that the cards never went into "Boost."

I'll do more here and on a pair of reference 560Ti cards when I have had some rest.

Profile jason_gee
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 24 Nov 06
Posts: 5080
Credit: 74,109,308
RAC: 5,010
Australia
Message 1276596 - Posted: 28 Aug 2012, 8:46:26 UTC - in response to Message 1276222.
Last modified: 28 Aug 2012, 8:47:22 UTC

I´m not shure but i think i see in some forum some special build for the 295 with x41z, maybe you could ask Jason about that.

About the table, ok i could see lower is better, but what those number realy means? Just curiosity.


There is, a cuda 3.2 version, but it errors out -9 overflow on one of them.


There's a Cuda 2.3 special on the way through final pre-Alpha preparations & testing, that is x41z specifically built for pre-Fermi Systems only (esp. 295's). That and a 2.2 version exists for certain gaming laptops that do not allow driver upgrades with standard nvidia release drivers, and are stuck way back at really ancient drivers (shame system manufacturers, shame).

There are also technical reasons that supporting such an old driver version has become necessary, in particular that nVidia's CUFFT library has gotten steadily slower for older gen cards over time... I reported that, demonstrated it, nothing happened. They want feedback on Cuda 5 preview, so I'm going to mention that I'll likely be forced to make my own FFT library to replace it (which I was planning to complete & reintroduce the freaky powerspectrum & some other technologies anyway) but beleive it's fair enough they should know why I may be forced to dump one of their technologies in favour of something that will not receive a performance degradation over time.

In shiorter words: on those old cards, on old drivers, 2.3 blows the doors off any later revision... but of course is not suitable for any Fermi or Kepler GPU containing host.


On the 680 Vs 590 comparison I saw a while back. Not going to be a fair comparison via time alone, since it's one GPU using about a third or even a quarter of the power, so overall more efficient. Do the comparison with a 690 ;)

Jason
____________
"It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is the most adaptable to change."
Charles Darwin

Previous · 1 · 2 · 3 · 4 · 5 · 6 . . . 13 · Next

Message boards : Number crunching : Optimize your GPU. Find the value the easy way.

Copyright © 2014 University of California