15.10.2009 21:33:39 NVIDIA GPU 0: GeForce 9600 GT (driver version 19107, compute capability 1.1, 512MB, est. 46GFLOPS)


oc 800/1000

---

10/15/2009 9:47:35 PM NVIDIA GPU 0: GeForce GTX 260 (driver version 19107, CUDA version 2030, compute capability 1.3, 896MB, est. 109GFLOPS)
10/15/2009 9:47:35 PM NVIDIA GPU 1: GeForce GTX 260 (driver version 19107, CUDA version 2030, compute capability 1.3, 896MB, est. 109GFLOPS)


Just popped these 2 in this evening...shame that they were sitting for 4 months in the box...

---

Clk2HlpSetiCty:::PayIt4ward

Saw the following posted here concerning Boinc development version 6.10.14

- client/scheduler: standardize the FLOPS estimate between NVIDIA and ATI. Make them both peak FLOPS, according to the formula supplied by the manufacturer.

So I installed 6.10.14 to check it out..............and here are the results.

GeForce GTX 275 (driver version 19107, CUDA version 2030, compute capability 1.3, 896MB, 700 GFLOPS peak)
GeForce 9800 GTX+ (driver version 19107, CUDA version 2030, compute capability 1.1, 1024MB, 470 GFLOPS peak)
GeForce 9800 GT (driver version 19107, CUDA version 2030, compute capability 1.1, 1024MB, 336 GFLOPS peak)
GeForce 9800 GT (driver version 19107, CUDA version 2030, compute capability 1.1, 1024MB, 336 GFLOPS peak)
GeForce GTX 260 (driver version 19107, CUDA version 2030, compute capability 1.3, 896MB, 498 GFLOPS peak)

All my GPU's are running at stock speed.

By the way Boinc 6.10.14 seems to be working well for me. At least I see no problems with it.

---

Boinc....Boinc....Boinc....Boinc....

16/10/2009 09:15:30 CUDA device: GeForce 9600 GT (driver version 19038, CUDA version 1.1, 1024MB, est. 37GFLOPS)

Stock Speed

These are the figures I'm getting with Boinc 6.10.14, all speeds are stock:

NVIDIA GPU 0: GeForce 9800 GTX/9800 GTX+ (driver version 19107, CUDA version 2030, compute capability 1.1, 512MB, 486 GFLOPS peak)
NVIDIA GPU 0: GeForce 8400M GS (driver version 19038, CUDA version 2030, compute capability 1.1, 128MB, 26 GFLOPS peak)

Claggy

Hi Claggy, see that you use the latest BOINC version?

Take it, that those GFLOP values are 'single-precision', cause I have a 9800GTX+ estimated at 85GFLOPS and run BOINC 6.6.36 on that WIN XP64 host.

Does BOINC 6.10.x gives raw values for both NVidia and ATI cards/GPU's ?

---

Hi Claggy, see that you use the latest BOINC version?

Take it, that those GFLOP values are 'single-precision', cause I have a 9800GTX+ estimated at 85GFLOPS and run BOINC 6.6.36 on that WIN XP64 host.

Does BOINC 6.10.x gives raw values for both NVidia and ATI cards/GPU's ?

boinc 6.6.38 => yes Nvidia & no ATI

boinc 6.10.x => yes Nvidia & yes ATI

:)

Representing all the notebook crunchers out there, here's the GTX 260M 1024MB GDDR3 card (actually a desktop 9800GT in disguise):


NVIDIA GPU 0: GeForce GTX 260M (driver version 18682, compute capability 1.1, 1024MB, est. 60GFLOPS)


The GFLOPS remains at 60, regardless of whether it's 550/1375/950 or 600/1500/900 (core/shader/memory) MHz. Right now I'm at 600/1500/900.
But CUDA-Z tells me 62481 Miops/s at stock and 66941 Miops/s at OC'd speeds, so shouldn't it be est. 62GFLOPS or 67GFLOPS?

Edit: Also, third post in three years... I should come here more often (the website that hosted my sig no longer exists..).

---

Representing all the notebook crunchers out there, here's the GTX 260M 1024MB GDDR3 card (actually a desktop 9800GT in disguise):

NVIDIA GPU 0: GeForce GTX 260M (driver version 18682, compute capability 1.1, 1024MB, est. 60GFLOPS)

For us Notebook Cruncher's, Beta 190.38 Notebook drivers available in the 2nd post in this thread (and so enabling Cuda 2.3 for us):

CUDA Toolkit and SDK 2.3 released

Expect a 30% ish reduction in Cuda crunching times (with optimised Cuda apps only at the moment)

Claggy

NVIDIA GPU 0: GeForce 9800 GT (driver version 19107, compute capability 1.1, 512MB, est. 19GFLOPS)

downclocked to 200mhz(core) 300mhz(mem)

Two home builds utilizing SLI:

NVIDIA GPU 0: GeForce GTS 250 (driver version 19038, CUDA version 2030, compute capability 1.1, 512MB, est. 84GFLOPS)

NVIDIA GPU 1: GeForce GTS 250 (driver version 19038, CUDA version 2030, compute capability 1.1, 512MB, est. 84GFLOPS)


NVIDIA GPU 0: GeForce 9600 GSO (driver version 19062, CUDA version 2030, compute capability 1.1, 768MB, est. 46GFLOPS)

NVIDIA GPU 1: GeForce 9600 GSO (driver version 19062, CUDA version 2030, compute capability 1.1, 768MB, est. 46GFLOPS)

---

BOINC 6.6.XX

30/09/2009 17:31:39 NVIDIA GPU 0: GeForce 8400 GS (driver version 19062, compute capability 1.1, 256MB, est. 8GFLOPS) <= PCIe

1wu = 8 heures

@+

BOINC 6.10.XX

31/10/2009 09:07:21 NVIDIA GPU 0: GeForce 8400 GS (driver version 19062, CUDA version 2030, compute capability 1.1, 256MB, 43 GFLOPS peak)

43 GFLOPS WOW !

Since I updated to BOINC 6.10.17 on all the machines a more meaningful comparison:

Sulu:
31/10/2009 9:38:55 PM NVIDIA GPU 0: GeForce GTX 295 (driver version 19062, CUDA version 2030, compute capability 1.3, 896MB, 596 GFLOPS peak)
31/10/2009 9:38:55 PM NVIDIA GPU 1: GeForce GTX 295 (driver version 19062, CUDA version 2030, compute capability 1.3, 896MB, 596 GFLOPS peak)

Chekov:
31/10/2009 9:30:14 PM ATI GPU 0: ATI Radeon HD 4700/4800 (RV740/RV770) (CAL version 1.3.145, 1024MB, 1000 GFLOPS peak)

Maul:
31/10/2009 2:53:28 AM NVIDIA GPU 0: GeForce GTX 260 (driver version 19062, CUDA version 2030, compute capability 1.3, 896MB, 537 GFLOPS peak)
31/10/2009 2:53:28 AM NVIDIA GPU 1: GeForce GTX 260 (driver version 19062, CUDA version 2030, compute capability 1.3, 896MB, 537 GFLOPS peak)

Spock:
31/10/2009 9:58:51 PM NVIDIA GPU 0: GeForce GTX 275 (driver version 19062, CUDA version 2030, compute capability 1.3, 896MB, 674 GFLOPS peak)

---

BOINC blog

Ok, this is new since I updated to 6.10.17..

10/30/2009 3:01:38 PM NVIDIA GPU 0: GeForce 9500 GT (driver version 19107, CUDA version 2030, compute capability 1.1, 1024MB, 118 GFLOPS peak)

---

PROUD MEMBER OF Team Starfire World BOINC

Boinc 6.10.17

NVIDIA GPU 0: GeForce 9500 GT (driver version 19062, CUDA version 2030, compute capability 1.1, 1024MB, 106 GFLOPS peak)

(clocks: 1650 shader/575 memory/680 core)

---

With different versions of Bionic reporting different values I think this is kinda pointless...

Get on that will ya Ace! :)

Who is actually in need / using this info?

If still the case, can't it be gotten more efficiently with a query directly to mork & jocelyn? And if you're going that far, might as well query a day's worth of results to build up the ultimate Seti Performance Pivot Table of CPU Class, CPU Rate, WU angle, GPU type, GPU quantity, GPU ram, OS, 32/64, etc, etc, cpu/gpu time taken, & logical time threshold buckets to detect overclocking of CPU & GPU.

---

Hi Wonder Puppy,
Just to let you know, I've got my 9500GT pumped up to shader:1850, core:725, memory:400. That's got me 12 extra GFLOPS over yours and with my fan set to auto@20% it is staying around 66-68c for my temp. Not sure how much extra those 16GFLOPS peak will give me but every little bit helps when you are as slow as my setup is. (the old way on 6.6.38 showed an improvement from 16GFLOPS to 21 GFLOPS using those settings.)

---

PROUD MEMBER OF Team Starfire World BOINC

Here's a shocker, for me anyway. This is OC'd on an HP 6190 quad-core. The out-of-the-box peak was 124 GFLOPS.

10/31/2009 NVIDIA GPU 0: GeForce GT 220 (driver version 18627, CUDA version 2020, compute capability 1.2, 1024MB, 144 GFLOPS peak)

My HP does not like the newer driver from nVIDIA so dropped back to this one. I suspect that HP will release their version of the new driver soon. At least I hope they do. Wanna see what it can do...

---

27/10/2009 5:25:37 PM NVIDIA GPU 0: GeForce GTX 295 (driver version 19107, compute capability 1.3, 896MB, est. 118GFLOPS)
27/10/2009 5:25:37 PM NVIDIA GPU 1: GeForce GTX 295 (driver version 19107, compute capability 1.3, 896MB, est. 118GFLOPS)

10/31/2009 8:52:06 AM NVIDIA GPU 0: GeForce GTX 285 (driver version 19062, compute capability 1.3, 1024MB, est. 133GFLOPS)