Message boards :
Number crunching :
NVidia GT240 - any good for crunching & expected RAC?
Message board moderation
Previous · 1 · 2 · 3 · 4 · Next
Author | Message |
---|---|
![]() ![]() Send message Joined: 7 Oct 01 Posts: 385 Credit: 50,200,038 RAC: 0 ![]() |
8am - 10pm => about 14 hours => 11k RAC x 24 / 14 = 19k RAC Would your CPU do about 5k 24/7? Then your two cards would do 14k, resulting in a staggering 7k RAC for each card! Really? Really, really...? Wow :D The SETI@Home Gauntlet 2012 april 16 - 30| info / chat | STATS |
Crun-chi ![]() Send message Joined: 3 Apr 99 Posts: 174 Credit: 3,037,232 RAC: 0 ![]() |
My GT240 also dont work 24/7, but I assume that RAC will not be lower then 5000 credits ( per card). So go for it, buy it, and crunch :) If you can find GT9800 green edition. It will be faster compared to GT240, but since it is green edition it also dont use more then 50-55 WATS I have that card in second machine 24/7 and have about 7000 RAC ( with proc 9000) I am cruncher :) I LOVE SETI BOINC :) |
![]() ![]() Send message Joined: 21 Apr 04 Posts: 3252 Credit: 31,903,643 RAC: 0 ![]() |
The 9800GTX+ is a bit faster then a 240 , or 9800GT GREEN, but uses ~100Watt, when >90% loaded, according to GPUz 0.3.9 But no load, only a screen, it still uses about 40 watt's. Have an ATI HD5770, running at Collatz C, former 3x+1, with only a screen 10Watt, fully loaded (95%), it uses 115Watt. But that is something, completely different... ![]() |
Crun-chi ![]() Send message Joined: 3 Apr 99 Posts: 174 Credit: 3,037,232 RAC: 0 ![]() |
The 9800GTX+ is a bit faster then a 240 , or 9800GT GREEN, ... 9800GTX+ will eat GT240 or GT9800 Green for breakfast. She is not only a bit faster.... I am cruncher :) I LOVE SETI BOINC :) |
hbomber Send message Joined: 2 May 01 Posts: 437 Credit: 50,852,854 RAC: 0 ![]() |
Yeah, was about to write the same. 9800GTX+ is way faster. Stock 8800GT/9800GT are still faster even than fairly well clocked GT240(DDR5). Quite easy way to compare them is to see how many GFLOPs BOINC says itself. For stock 8800GT it estimates it as 336 GFLOPs, at 700/1750/2000(memory doesn't matter here, where it does matters a lot in real crunching) it estimates it as 392 GFLOPs(this number is IIRC - my 8800GT rig is silent, can't check now, I cannibalized its PSU, another 750W Chieftec died earlier today, the world is falling apart). |
ai5000 Send message Joined: 1 Jan 01 Posts: 57 Credit: 2,805,412 RAC: 0 ![]() |
|
hbomber Send message Joined: 2 May 01 Posts: 437 Credit: 50,852,854 RAC: 0 ![]() |
Link has expired. |
![]() ![]() ![]() Send message Joined: 23 Oct 02 Posts: 5862 Credit: 10,957,677 RAC: 18 ![]() ![]() |
8am - 10pm => about 14 hours => 11k RAC x 24 / 14 = 19k RAC I just don't understand this. I'll be surprised if my card is doing 500 RAC in 24 hours... What on earth could I have done so wrong? I installed the card, the drivers, BOINC, the enhanced apps... |
![]() ![]() ![]() Send message Joined: 23 Oct 02 Posts: 5862 Credit: 10,957,677 RAC: 18 ![]() ![]() |
OK, I think I have answered the question myself. The Quaddie would not boot up this morning, DLL errors, Blue Screen of Death :( Booted in to safe mode, de-installed the Nvidia drivers, re-installed the drivers that shipped with it and I'm up & running again - and a LOT faster at crunching with the GT240 too! I've just watched the GPU crunch 50% of a WU in 10 mins, so seems comparable to what others are reporting. |
Crun-chi ![]() Send message Joined: 3 Apr 99 Posts: 174 Credit: 3,037,232 RAC: 0 ![]() |
Look at this host http://setiathome.berkeley.edu/results.php?hostid=5296162&offset=0&show_names=0&state=3 In first has normal times for this card, but how can he beat that times by factor 5x ( 500%) faster??? This card is very very strange card :) I am cruncher :) I LOVE SETI BOINC :) |
![]() ![]() Send message Joined: 30 Aug 01 Posts: 1228 Credit: 47,779,411 RAC: 32 ![]() ![]() |
Link has expired. From what I remember, it was a link to a WU that was completed by an 8800GT and a GT240 - the 8800GT was in the 1800-ish second range and the GT 240 was in the 1400-ish second range. FYI -Dave ![]() |
ai5000 Send message Joined: 1 Jan 01 Posts: 57 Credit: 2,805,412 RAC: 0 ![]() |
^^^ Bingo. Both stock apps. |
ai5000 Send message Joined: 1 Jan 01 Posts: 57 Credit: 2,805,412 RAC: 0 ![]() |
Look at this host It looks like all the wu's sent on Feb 9 or earlier were completed in "normal" time, but the tasks sent Feb 12 all completed very quickly. Strange indeed. |
![]() ![]() Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 ![]() |
Look at this host Yes, most noticeably, The CPU time exceeds the Run Time for many wus. Defying the laws of physics that way (The CPU time should be < Run Time) might indicate a malfunction of the timer mechanism in Boinc, possibly OC or BIOS setting induced, maybe privileges of some sort to access process info. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Crun-chi ![]() Send message Joined: 3 Apr 99 Posts: 174 Credit: 3,037,232 RAC: 0 ![]() |
Whatwever is case: I would like to have that system :) With no pain I will beat GTX 295 :) I am cruncher :) I LOVE SETI BOINC :) |
hbomber Send message Joined: 2 May 01 Posts: 437 Credit: 50,852,854 RAC: 0 ![]() |
Link has expired. This may be an exception. We don't know circumstances around and clock speeds of the cards. 8800GT may be declocked(on purpose or it is just some "green" freak, while GT240 may be OCed over clouds) Where do I make my conclusions for 8800GT from? Look at this host - http://setiathome.berkeley.edu/show_host_detail.php?hostid=5311137(it is in experimental state, so it makes occasional errors). 111 credit WUs on stock speeds are completed bellow 20 minutes. Lets round it at 20. This makes 72 WUs for 24 hours. RAC 8000. OCed at 700/1782/1000 it gives completions times for same WUs 18 minutes. 8900 RAC. From what I read so far, there is no GT240 able to reach such RACs. Stock cards - 8800GT has 50 MHz more core clock(10%), more clocked shaders(dunno how much exactly), has 16 shader processors more(96 vs 112), has 256-bit memory interface, while GT240 has 128-bit. Only GDDR5 version can compare with 8800GT memory bandwidth. And there are still above advantages in favor of 8800GT. I exclude architectural changes in GT240 GPU, they are insignificant, if any. There are many GT240 reviews on net. I haven't found any where it beats 8800GT in any area of applications. P.S. Comparison can be made - this host http://setiathome.berkeley.edu/hosts_user.php?userid=58531 completion times of 111 credits WUs are well above those of my host. Its one of ai5000's hosts. It would be nice if he tells us is his card tuned in some way, to clarify situation. |
Crun-chi ![]() Send message Joined: 3 Apr 99 Posts: 174 Credit: 3,037,232 RAC: 0 ![]() |
If you look at result you can find this :clockRate = 1340000 That info was taken from GT240 , and that is normal freq for shaders. So in this case GT240 is not OC-ed at all. What I found with this card: but no only this card: there is some host they are significant faster then others, but many of them uses Boinc without any optimized application. So it still is mystery what make such difference from one host to another host. I dont talk about 5% difference , difference is over 50% or even more. I have GT 240 that is OC about 18% and still my card is so slower compared to this host. Why? Remain mystery :) I am cruncher :) I LOVE SETI BOINC :) |
hbomber Send message Joined: 2 May 01 Posts: 437 Credit: 50,852,854 RAC: 0 ![]() |
If you look at result you can find this :clockRate = 1340000 Good to know. Thank you! |
ai5000 Send message Joined: 1 Jan 01 Posts: 57 Credit: 2,805,412 RAC: 0 ![]() |
Correct, no overclock. Boinc 6.10.18, stock app and 196.21 drivers. This is a GDDR5 card though. Looking through my completed tasks the only cards I've seen faster are GTX 260s and above. I haven't seen a 9800GTX+, but there have been several 9800GT's, a 8800 Ultra and 8800GTS, some with optimized apps, but all were slower. |
hbomber Send message Joined: 2 May 01 Posts: 437 Credit: 50,852,854 RAC: 0 ![]() |
You have exactly one 111 credit result, calculated faster than my stock speed results. Most of yours are between 1350-1400 seconds. My results are with consistent times 1210-1220 seconds. Like these: http://setiathome.berkeley.edu/result.php?resultid=1523504920 http://setiathome.berkeley.edu/result.php?resultid=1523504894 http://setiathome.berkeley.edu/result.php?resultid=1523373211 http://setiathome.berkeley.edu/result.php?resultid=1523373202 http://setiathome.berkeley.edu/result.php?resultid=1523373200 http://setiathome.berkeley.edu/result.php?resultid=1523373189 http://setiathome.berkeley.edu/result.php?resultid=1523373181 http://setiathome.berkeley.edu/result.php?resultid=1523373177 http://setiathome.berkeley.edu/result.php?resultid=1523373175 http://setiathome.berkeley.edu/result.php?resultid=1523373167 http://setiathome.berkeley.edu/result.php?resultid=1523373165 http://setiathome.berkeley.edu/result.php?resultid=1523373163 I wouldn't call GT240 faster. Nothing personal. Numbers say it all. But still GT240 performs fairly well. |
©2025 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.