From the compute benchmarks, the K40 is less than half the performance of the GTX 1070 Ti. And uses more power.

That explains the low price.

Likely and why it is being dumped on eBay for that price which is 1/10 its original MSRP.

Someone is is likely wanting to upgrade to the current Tesla or Quadro compute products which are both more powerful AND more energy efficient.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

From the compute benchmarks, the K40 is less than half the performance of the GTX 1070 Ti. And uses more power.

So I may have found the wrong model on GPU-Z :(

Tom

---

A proud member of the OFA (Old Farts Association).

AFAIK The Tesla models compared with the Gforce equivalents have superior DP capacity, and a far superior driver.

One of the nice features of their driver is the capacity to keep running GPU work when you do RDP. That makes the need to VNC or programs like Teamviewer unnecessary.

Their are designed to run 24/7 in heavy duty works. I know we run Gforce models 24/7 but they are not designed to do that.

They target is the business community with heavily support from NVidia (and a far superior price) not the general public.

In SETI, who don't use DP they make little difference in crunching speed when you compare against similar Gforce models, sure not the case of the K40 (Kepler) vs 1070Ti (Pascal).

---

---

Grant
Darwin NT

ROFLing!!!!

Thank you!

Tom

---

A proud member of the OFA (Old Farts Association).

Still, $200 is peanuts to some rich people, they can afford to clutch at straws in the wind.

Hrmmmppp.... I don't clutch at straws..... now gpu's...... :)

You are right though. I am price sensitive. A gtx 1060 3GB which does a very good job under either windows on Linux/CUDA91 for less than $130USD can get my "more gpu's" going. If I wasn't price sensitive and I actually had the budget and the PSU I would be running all RTX 2880Ti's :)

Once it gets here I well plug it in to one of my Seti boxes and see what kind of performance I get. If it turns out as everyone else has predicted it will be an "oh well, darn" moment. Otherwise I will have a painful itch to start the upgrade cycle (again :)

Tom

---

A proud member of the OFA (Old Farts Association).

Greetings,

Just out of the blue, I decided to install my old EVGA GT 740 SC video card to see if it works. BOINC sees it, Device Manager sees it and NVIDIA sees it. I am using the monitor on my GTX 1050Ti.

This is what BOINC Manager tells me:

3/1/2019 11:17:10 AM |  | CUDA: NVIDIA GPU 0: GeForce GTX 1050 Ti (driver version 419.17, CUDA version 10.1, compute capability 6.1, 4096MB, 3376MB available, 2274 GFLOPS peak)
3/1/2019 11:17:10 AM |  | CUDA: NVIDIA GPU 1 (not used): GeForce GT 740 (driver version 419.17, CUDA version 10.1, compute capability 3.0, 4096MB, 3431MB available, 833 GFLOPS peak)
3/1/2019 11:17:10 AM |  | OpenCL: NVIDIA GPU 0: GeForce GTX 1050 Ti (driver version 419.17, device version OpenCL 1.2 CUDA, 4096MB, 3376MB available, 2274 GFLOPS peak)
3/1/2019 11:17:10 AM |  | OpenCL: NVIDIA GPU 1 (ignored by config): GeForce GT 740 (driver version 419.17, device version OpenCL 1.2 CUDA, 4096MB, 3431MB available, 833 GFLOPS peak)

The following is my app_config.xml file:

<app_config>
	<app_version>
		<app_name>setiathome_v8</app_name>
		<plan_class>opencl_nvidia_SoG</plan_class>
		<avg_ncpus>1</avg_ncpus>
		<ngpus>0.5</ngpus>
	</app_version>
	<app_version>
		<app_name>setiathome_v8</app_name>
		<plan_class>cuda42</plan_class>
		<avg_ncpus>1</avg_ncpus>
		<ngpus>0.5</ngpus>
	</app_version>
	<app_version>
		<app_name>setiathome_v8</app_name>
		<plan_class>cuda50</plan_class>
		<avg_ncpus>1</avg_ncpus>
		<ngpus>0.5</ngpus>
	</app_version>
	<app_version>
		<app_name>astropulse_v7</app_name>
		<plan_class>opencl_nvidia_100</plan_class>
		<avg_ncpus>1</avg_ncpus>
		<ngpus>0.5</ngpus>
	</app_version>
	<project_max_concurrent>12</project_max_concurrent>
</app_config>

My question is this: What do I need to edit or add to the .xml file so both cards can be used? I have looked on the Client configuration page to try to figure this out. All I see is how to tell BOINC how many WUs to do per GPU. I have not found anything that tells me how to have BOINC use X number of GPUs.

Any help would be greatly appreciated! :)

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Do you have <use_all_gpus>1</use_all_gpus> in cc_config.xml?

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Do you have <use_all_gpus>1</use_all_gpus> in cc_config.xml?

this

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

Do you have <use_all_gpus>1</use_all_gpus> in cc_config.xml?

Hi Keith,

I edited cc_config.xml and reloaded the configs and nothing. I shut down BOINC and restarted it and now both GPUs are working.

I'll have to go over that Client configuration page again. ;) Thanks! :)

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Do you have <use_all_gpus>1</use_all_gpus> in cc_config.xml?

Hi Keith,

I edited cc_config.xml and reloaded the configs and nothing. I shut down BOINC and restarted it and now both GPUs are working.

I'll have to go over that Client configuration page again. ;) Thanks! :)

Have a great day! :)

Siran

The majority of things in the cc_config can be read and acted on with just a re-read. However discovering the installed gpu and its capabilities only happens at BOINC launch. That is why you didn't notice a change when you simply re-read cc_config

Glad I could help with a simple answer.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

The benefits of technology- 3D printed ears could help children with ear deformities avoid complex surgery.

Children with ear deformities will soon be able to get printed ears made from their own stem cells, according to a team of Wollongong researchers working on new 3D bioprinting technology. They claim their work represents a "huge breakthrough" in the field.
The bio-printer, called 3D Alek, was developed at the University of Wollongong and is now being trialled at Sydney's Royal Prince Alfred Hospital (RPA). Its first task is to design and create 3D-printed human ears for those with existing deformities, particularly children with congenital defects like microtia, and would otherwise require complex surgery.

---

Grant
Darwin NT

Noticing a few downloads lately that get stuck and won't download after many tries....I've been just aborting them. : /

Various inquiries to me about where the Nvidia GTX 1660Ti falls out with regard to compute performance. Phoronix.com posted OpenCL and CUDA benchmark tests against common cards like the 1070, 1080, RTX and AMD RX, R9 and Vega variants.
https://www.phoronix.com/scan.php?page=article&item=nvidia-gtx1660ti-opencl&num=3

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Phoronix.com posted OpenCL and CUDA benchmark tests against common cards like the 1070, 1080, RTX and AMD RX, R9 and Vega variants.

Ah good, been waiting for their compute results.

Edit- very nice.
Generally around the GTX 1080 levels of performance, and for considerably less power used.

---

Grant
Darwin NT

Greetings,

I must take full responsibility for this current crisis we have. I decided to try my old EVGA GT 740 SC graphics card. I installed it and needed a bit of help to get BOINC to see the card. Thanks to Keith, I was able to get the card running WUs. I got some major downloads at that time. Some time that night I think the new (old) card broke the project. My most humblest of apologies to everyone. ;)

Now, if I could just get my 200+ WUs to report and load my cache back up, things will be hunky dory again. :)

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Greetings,

I must take full responsibility for this current crisis we have.
Siran

Haha....I seriously doubt that, but kind of you to take the blame !
At least we are back....back to downloads stuck in transfer.

...Ghia...

---

Humans may rule the world...but bacteria run it...

Yes stuck downloads, but 1 at a time seems to still work.

---

Greetings,

I must take full responsibility for this current crisis we have. I decided to try my old EVGA GT 740 SC graphics card. I installed it and needed a bit of help to get BOINC to see the card. Thanks to Keith, I was able to get the card running WUs. I got some major downloads at that time. Some time that night I think the new (old) card broke the project. My most humblest of apologies to everyone. ;)

Now, if I could just get my 200+ WUs to report and load my cache back up, things will be hunky dory again. :)

Have a great day! :)

Siran

ROFLing....

---

A proud member of the OFA (Old Farts Association).

Phoronix.com posted OpenCL and CUDA benchmark tests against common cards like the 1070, 1080, RTX and AMD RX, R9 and Vega variants.

Ah good, been waiting for their compute results.

Edit- very nice.
Generally around the GTX 1080 levels of performance, and for considerably less power used.

Hmmmm.......

Gtx 1080 performance level in a Gtx 1X60 card. I knew my gtx 1060's were a good buy. Just look at the upgrade path :)

Tom

---

A proud member of the OFA (Old Farts Association).