GTX 670, GTX 680, GTX 690 is not supported yet (by the CUDA 3.1 application). They will be supported (hopefully) in a couple of days by the CUDA 4.2 application, which is in beta testing right now.

On the first betas, my 670 was about 10% slower. Well find out the real differences soon enough though. This was on W7.

I only ran the first beta set on the 670. I may have been playing D3 on the second round on that card.........

Wouldn't see how it would make a difference though. It's going to be slower than the 680 for sure, and -10% is pretty good considering the 680 is $100+ more. Not to mention a pain to find still.

Ah......

Sorry about that. Was typing quickly on my phone. Apologies.

Yes, my 670 was -10% slower than my 680. This was clock for clock. Set both the same speeds.

My GTX670 in Win 7 with 301.42 driver.

GPUgrid -
ACEMD beta version 6.43(cuda42) - GPU load 83% (NVIDIA Inspector 1.9.6.5)

Primegrid -
Genefer (WR) 1.07 (cuda32_13) - GPU load 99% (NVIDIA Inspector 1.9.6.5)

True enough. Application programming is the difference. All projets need to define their resources for programming based on the assets and abilities to accommodate the new tech. Please don't let these limitations confuse donors into thinking that a poject doesn't care in the "short run". :)
If it continues over a couple months... that's different :)
They appreciate us until proven differently.
EDIT: We all have to remember that vista and W7 are an entirely different OS background. All of the pogramming has to be adjusted for V and W7. It's not Just the NV or AMD drivers. That's why NV has separate drivers for XP and previous versions and Vista and W7. The latter are not bases on NT tech ;)
Vista never took "massive hold" and neither has W7 yet. Programmers have to do what best results in results.
JIMO, YMMV (Just in my opinion, your mileage may vary) :)
sean

Back to the original question...

I see the "production" apps haven't switched to cuda40 or cuda42 yet, so will have to leave the GTX570's in place until that happens.

What do GPUGrid pros make of this statement from Anandtech regarding the 690?

"Unfortunately for NVIDIA GK104 shows its colors here as a compute-weak GPU, and even with two of them we’re nowhere close to one 7970, let alone the monster that is two. If you’re looking at doing serious GPGPU compute work, you should be looking at Fermi, Tahiti, or the future Big Kepler."

http://www.anandtech.com/show/5805/nvidia-geforce-gtx-690-review-ultra-expensive-ultra-rare-ultra-fast/15

Removing power consumption from consideration, does this really suggest that 590 is preferable to a 690 for GPUgrid? Asked differently, what can one extrapolate from that review/benchamark about a 690's performance here?

They're discussing double precision. This project is single precision.

There other ibuch tasks around that get up to 96. The one you have now are actually the slowest out of all the different WUs in the short queue.

I have a MSI factory overclocked 670 and it gives me computation errors in Linux right off the bat with the 302.17 drivers. I'm thinking it has something to do with the overclock since other people seem to be crunching just fine on Linux with the newest drivers and a 670. Either way, I've reverted back to 295.59 and it averages about four hours on a cuda42 long run, which I would assume is decent since it says 8-12 hours on fastest card in the description.

HI: I also have installed the 302.17 and with my GTX295 there is no way to run CUDA 4.2 tasks, from what I read is a matter of reinstalling the old 295.59 as soon as you finish the tasks that I have running CUDA 3.1... We'll see.

I have a GTX480 and it runs GPUGrid Fine and Dandy under the Linux 302.17 driver.

So much so, I am approaching the top 100 GPUGrid clients based on RAC (actually no 146 earlier today, up from 164 last night)

it runs CUDA31 and CUDA42 apps. Most CUDA31 Long runs take about 9 hours. CUDA42 Long runs take about 5 hours.

I had been running some PrimeGrid Cuda WUs until I found out that most Cuda work there is Double Precision. I didn't realize 680s were worse at DP than the 500s are. I guess SP projects are the way to go for the 600s series. I did get the 680 signature 2 with 2 fans and its very fast at SP work such as GPU Grid. I guess Milky Way would be slower also. Its kind of dissapointing Nvidia handicapped the 600s in DP. Maybe the GK110 will be better.

Thats great, do you know if most of the projects will all eventually be DP, or is it only the Mathmatically leaning projects that will be this way.

My 570 is faster than my 680 by 15-20% on DP work.
GK110 is on my list definitly.
I just hope its in the less than 600-700 dollars though. That hurts.
I may be released as Tesla or Quadro only, we will see I suppose.

I think most of you ignored the fact that Anand's compute benchmark consisted of mostly if not all SP benchmarks.

It does seem like Kepler is a step backwards in terms of GPGPU.

4 hours is good.

Most of my CUDA42 workunits take about 5.5 hours on my GTX480

That is a bit shy of a 30% performance improvement for double the price.

Course power consumption drops by a similar amount. Its hard to factor the power consumption figure into the overall household utility expense in most areas of the USA (where time/demand pricing is not in force - time of day consumption may raise the utility cost by 100% or more compared to other times of the day

Hello, is it a good idea by following:

I want to overclock my GTX 670 (Asus)

Now i thought i could make it easy, an so:

Raise the voltage by 1mV per 1 Mhz, is it a good idea
or will i burn my chipset away !?

Greets :)

hm, i dont did it. Increasing the "power target" does it all. It clocks itself. no thinking about overvolting. Thanks.

hm, i dont did it. Increasing the "power target" does it all. It clocks itself. no thinking about overvolting. Thanks.

On my GTX 670 it does not matter what I set the power target at it always pulls 1175. So knowing that I was not able to increase / decrease volts I might as well OC as far as is stable ... turns out that 1259 GPU and 3206 MEM is rock solid stable - 99 consecutive successful LONG WUs so far. Win7x64, BOINC 7.0.25.

Looked through some results for GTX 670s accross GPUGrid for Win7x64 with NATHAN WUs: OC I listed above takes about 9%-10% less time per WU and gets me to the ballpark of a stock GTX680!!!
____________
Thanks - Steve

i did base overclocking (but to far) , now at 915+90mhz and set power target with Nvidia Inspector at 122% but it wont raise further that 1175mV. Not always stable right now.

I've begun to upgrade my Fermi cards to Kepler cards, and I've made some measurements with my first partly upgraded system.
Its configuration is (at the moment):

ASUS P7P55 WS Supercomputer motherboard
Intel Core i7-870 @ 3855MHz (24*160MHz)
2*2Gb DDR3 2000MHz RAM
320GB HDD
MSI GTX 480 @ 800MHz (1.075V) (with an Artctic Cooling Accelero Xtreme Plus)
Asus GTX-670 DC2 @ 1084MHz (1.137V) (factory overclocked)

A NATHAN_RPS_1120528 runs nearly 13% faster (15137 sec vs 17133 sec) on the GTX 670 @ 1084MHz than on the GTX 480 @ 800MHz.
I've also measured the power consumption at the wall outlet (230V AC).
When my PC was idle (no tasks running, but power management is disabled, so the CPU runs at full speed) it is drawing 178 Watts
When a task is running on the GTX 480 @ 800MHz (99% GPU usage): 378 Watts
When one more task is running on the GTX 670 @ 1084MHz (99% GPU usage): 552 Watts
When 4 rosetta@home are running on the CPU: 625 Watts
So, the extra power consumption of the different parts when they are in use is:
GTX 480 @ 800MHz (99% GPU usage): 200 Watts
GTX 670 @ 1083MHz (99% GPU usage): 174 Watts
CPU 4 cores: 73 Watts
As you can see, the GTX 670 @ 1083MHz consumes 87% of the GTX 480 @ 800MHz.
Hopefully I can do more measurements in this week.

Retvari:

Thank you for the power measurements. This is very helpful. We can use this info to justify the purchase of new cards based on power savings.

I see a 670 in my future!
____________
Thx - Paul

Note: Please don't use driver version 295 or 296! Recommended versions are 266 - 285.

...but a 680 might be more in line with a 480, and might take it over 40%. That said the 580 would still be more competitive.

The PCIE3 vs PCIE2 debate is still open.

Can you reduce that FOC?

The GFlops looks very odd in the BOINC manager's log:

NVIDIA GPU 0: GeForce GTX 680 (driver version 30479, CUDA version 5000, compute capability 3.0, 2048MB, 582 GFLOPS peak)
NVIDIA GPU 1: GeForce GTX 590 (driver version 30479, CUDA version 5000, compute capability 2.0, 1536MB, 1244 GFLOPS peak)
NVIDIA GPU 2: GeForce GTX 590 (driver version 30479, CUDA version 5000, compute capability 2.0, 1536MB, 1244 GFLOPS peak)

NVIDIA GPU 0: GeForce GTX 480 (driver version 30479, CUDA version 5000, compute capability 2.0, 1536MB, 1538 GFLOPS peak)
NVIDIA GPU 1: GeForce GTX 670 (driver version 30479, CUDA version 5000, compute capability 3.0, 2048MB, 439 GFLOPS peak)

It looks like I'm downgrading my cards....

I've tried to refine my measurements with my partly upgraded configuration.
It's very hard to calculate the power consumption of different parts by measuring the overall power consumption at different workloads, because the parts are heating each other, and it is causing extra power consumption on the previously measured parts. My previous measurements didn't take this effect into consideration, so the extra 174 Watts doesn't come only from the task running on the GTX 670.

no GPU tasks, no CPU tasks, no power management (idle) : 180W

no CPU tasks, no GPU task on the GTX 670
GTX 480 1025mV, 701MHz, 99%, 61°C: 344W (164W)
GTX 480 1050mV, 701MHz, 99%, 63°C: 355W (175W) (+11W)
GTX 480 1075mV, 701MHz, 99%, 64°C: 365W (185W) (+21W)
GTX 480 1075mV, 726MHz, 99%, 64°C: 370W (190W) (+26W)
GTX 480 1075mV, 749MHz, 99%, 65°C: 374W (194W) (+30W)
GTX 480 1075mV, 776MHz, 99%, 66°C: 378W (198W) (+34W)
GTX 480 1075mV, 797MHz, 99%, 66°C: 382W (202W) (+38W)

GTX 480 1075mV, 797MHz, 99%, 71°C: 395W (215W) (+51W)
GTX 670 1162mV, 981MHz, 97%, 66°C: 555W (160W)

+ 4 CPU cores: 620W

GTX 480 1075mV, 797MHz, 99%, 67°C: 383W (203W)
GTX 670 1137mV,1084MHz, 97%, 60°C: 552W (160W)

GTX 480 1075mV, 797MHz, 99%, 71°C: 397W (217W) (+53W)
GTX 670 1137mV,1084MHz, 97%, 65°C: 559W (162W)

GTX 480 1075mV, 797MHz, 99%, 71°C: 397W (217W)
GTX 670 1137mV,1084MHz, 0%, 44°C


GTX 480 1075mV, 797MHz, 0%, 44°C:
GTX 670 1137mV,1084MHz, 97%, 65°C: 341W (161W)

All in all, the GTX 670 is better than I've calculated from my first measurements:
The GTX 670 @ 1083MHz consumes (162W) 75% of the GTX 480 @ 800MHz (217W).

These were my last measurements with this host.
I'm going to change the GTX 480 to a GTX 670 on this host right now, and I will measure the power consumption again after that.

By reducing the second cards heat radiation both cards should benefit somewhat, and you might even see the pair reach +40% performance per Watt over two GTX480's.

Out of curiosity, are your CPU heatsink fins/blades vertical or horizontal?

I would like to ask something and i m sorry if it has been answered in another thread. I am running both WCG and Gpugrid on my pc (3770k/M5G/GTX670). Is it normal that some projects have more gpu usage than others? Also, when i run WCG with all 4c/8t i see 73-74% gpu usage, but when i stop WCG i see 81-82% gpu usage. Why does this happen? Can i do something to rum both wcg and at the same time utilize my gpu at 100%?

I would like to ask something and i m sorry if it has been answered in another thread. I am running both WCG and Gpugrid on my pc (3770k/M5G/GTX670). Is it normal that some projects have more gpu usage than others? Also, when i run WCG with all 4c/8t i see 73-74% gpu usage, but when i stop WCG i see 81-82% gpu usage. Why does this happen? Can i do something to rum both wcg and at the same time utilize my gpu at 100%?

More power consumption and performance measurements are in progress :)
I've a PC with a Core i5-3570K in an Intel DH77EB motherboard for a couple of days. I'm not sure if the motherboard supports PCIe3.0 though.

My 3rd 680 will be in my hands in about 2 weeks. Since I already have 2 working at PCIe 3 on my x79 when I install the 3rd it will be PCIE 2 x8 (PCIe 1) until I apply the hack which will make it PCIe 3 x8 (PCIe 2). I will test both to see how the times compare.

So in short in 2 weeks I will have results on a 680 across all PCIe bandwidth speeds.

The power supply in my dual GTX-670 host is an Enermax MODU87+ 800W.

Hey. All gpugrid.net work packages need a cpu to feed it. You have a quad-core processor with hyperthreading supported so your processor is shown in eight core cpu in operating system. Leave a one core-free so that the processor can support the graphics cards in its calculations.

BOINC Manager, go to cpu usage (preferences) and select that you want to use 87.5% of the processors so one core is always free for the graphics card.

Very likely you will not get 100% GPU utilization.

My experimental PCIe3.0 x16 host is up and running.
I've checked the speed of the PCIe bus with GPU-Z 0.6.3, according to this tool the GPU runs at PCIe3.0 x16.
It has the Gigabyte GV-N680OC-2GD in it (moved from my main cruncher PC)
We'll see how it performs against my old (PCIe2.0 x16) host.
There is no CPU tasks running on this host, because it has a 400W power supply.
It was quite an adventure to install Windows XP x64 on this configuration.
I've checked the GPU with MSI Afterburner 2.2.3:
Usage: 97%
Voltage: 1.175V
Clock: 1137MHz
I had the same numbers with my old host.

The PCIE3 vs PCIE2 debate is still open.

Might be a bit late to answer, but I have been using an EVGA GTX 670 SC (which is slightly overclocked) for about a month. I also use the 301.42 driver for the card.

Other specs:
Intel i7-3820 @ 4.3 GHz
16 GB of RAM
Windows 7 64-bit

I usually run GPUGRID during the day while idle and through the night. I run all the applications available and so far my times have been pretty fast.

For long-runs, it takes about 5-7 hours, depending on the task. The standard tasks run for about 2-5 hours also. The beta tasks take anywhere from 30 minutes to 2 hours or so.

Note that the times do range, and as of now I'm crunching many beta tasks so I couldn't give too precise of times on the other tasks.

As for heating, the card stays at about 67 degrees Celsius during long-runs and doesn't go higher. The card load is at about 90-95% during long-runs, also, and ranges around 50% during beta tasks.

The only problems I've had are driver crashes when stopping the beta tasks that have come up recently, which are the Noelia tasks.

Hope this helped.

Might be a bit late to answer, but I have been using an EVGA GTX 670 SC (which is slightly overclocked) for about a month. I also use the 301.42 driver for the card.

Other specs:
Intel i7-3820 @ 4.3 GHz
16 GB of RAM
Windows 7 64-bit

I usually run GPUGRID during the day while idle and through the night. I run all the applications available and so far my times have been pretty fast.

For long-runs, it takes about 5-7 hours, depending on the task.

I'm pretty sure my CPU times are larger because I didn't give a core to the GPU, and instead gave all 8 to another project. Up until now I never really thought about it, and will try to fix it as soon as I can. Thanks for the notice.

6XX cards always use a full core so if you set BOINC to use 1 less thread than your CPU has available your other projects that use CPU only will process much better!
____________
Thanks - Steve