Of course running 2 instances of AP on GPU is better for RAC.
Also more efficient.

The only advantage with a low end card is when running Blanked APs. Soon, the problem with Blanked APs will be history. I would like to see a comparison with a low end card running AP_v7.

Not true, it depends on the CPU not GPU.

Have you tested multiple instances on a low end card with AP_v7? With the Single instance tuned to run at 98% load?

Yes, sure.
I tested 8 different NV GPU`s last week with AP7.

Can you point to the results?

Of course running 2 instances of AP on GPU is better for RAC.
Also more efficient.

The only advantage with a low end card is when running Blanked APs. Soon, the problem with Blanked APs will be history. I would like to see a comparison with a low end card running AP_v7.

Not true, it depends on the CPU not GPU.

Have you tested multiple instances on a low end card with AP_v7? With the Single instance tuned to run at 98% load?

Yes, sure.
I tested 8 different NV GPU`s last week with AP7.

Can you point to the results?

Not yet.
All were offline benches of course.
Results are stored on my server.

---

With each crime and every kindness we birth our future.

Of course running 2 instances of AP on GPU is better for RAC.
Also more efficient.

The only advantage with a low end card is when running Blanked APs. Soon, the problem with Blanked APs will be history. I would like to see a comparison with a low end card running AP_v7.

Not true, it depends on the CPU not GPU.

Have you tested multiple instances on a low end card with AP_v7? With the Single instance tuned to run at 98% load?

Yes, sure.
I tested 8 different NV GPU`s last week with AP7.

Can you point to the results?

Not yet.
All were offline benches of course.
Results are stored on my server.

As soon as AP7 is released on Main I think we will see a number of comparisons. Remember, it must be a low end card compared with the Single instance tuned to run at 98-99% load.

Of course running 2 instances of AP on GPU is better for RAC.
Also more efficient.

The only advantage with a low end card is when running Blanked APs. Soon, the problem with Blanked APs will be history. I would like to see a comparison with a low end card running AP_v7.

Not true, it depends on the CPU not GPU.

Have you tested multiple instances on a low end card with AP_v7? With the Single instance tuned to run at 98% load?

Yes, sure.
I tested 8 different NV GPU`s last week with AP7.

Can you point to the results?

Not yet.
All were offline benches of course.
Results are stored on my server.

As soon as AP7 is released on Main I think we will see a number of comparisons. Remember, it must be a low end card compared with the Single instance tuned to run at 98-99% load.

The last thing i`ll say about it here is dont forget you are testing stock at beta.
I`m testing optimised for month.

---

With each crime and every kindness we birth our future.

The time to complete for 2 v7.04 APs on my 750s are about the same as for V6... 1 hr 20 each. My 2 cents

Edit...it actually might be faster, closer to 1 hour but I'm notice some "interesting" variables that affect the time. More to do with my set up than anything seti side.

Zalster

The time to complete for 2 v7.04 APs on my 750s are about the same as for V6... 1 hr 20 each. My 2 cents

Because you are testing stock too.
Just wait for the new read me.

---

With each crime and every kindness we birth our future.

The time to complete for 2 v7.04 APs on my 750s are about the same as for V6... 1 hr 20 each. My 2 cents

Edit...it actually might be faster, closer to 1 hour but I'm notice some "interesting" variables that affect the time. More to do with my set up than anything seti side.

Zalster

The APv6 must not have had any Blanking, else it would have been much slower than v7.
Running Multiple instances in v6 will benefit when running Blanked APs because the GPU is not under Full load when running Blanked tasks, so the other instance can use the spare cycles. When there aren't any spare cycles, there is no benefit.
Try comparing your 750 running a tuned single instance of APv7 running at 98% with two instances of APv7 and see how it works.

Give me an hour and you will have your result.

Edit... we are off topic here but since we started here I'll finish here.

The APv6 must not have had any Blanking, else it would have been much slower than v7.
Running Multiple instances in v6 will benefit when running Blanked APs because the GPU is not under Full load when running Blanked tasks, so the other instance can use the spare cycles. When there aren't any spare cycles, there is no benefit.
Try comparing your 750 running a tuned single instance of APv7 running at 98% with two instances of APv7 and see how it works.

TBar. Single instance of APv7 on 750Ti is 52 minutes average (no command lines). 2 instances of AP v7 on 750Ti is 1 hr 14 min without any command lines.

The APv6 must not have had any Blanking, else it would have been much slower than v7.
Running Multiple instances in v6 will benefit when running Blanked APs because the GPU is not under Full load when running Blanked tasks, so the other instance can use the spare cycles. When there aren't any spare cycles, there is no benefit.
Try comparing your 750 running a tuned single instance of APv7 running at 98% with two instances of APv7 and see how it works.

TBar. Single instance of APv7 on 750Ti is 52 minutes average (no command lines). 2 instances of AP v7 on 750Ti is 1 hr 14 min without any command lines.

Is the Single instance running at 98-99% load? If I'm looking at the correct machine, you are running the Default settings, http://setiweb.ssl.berkeley.edu/beta/results.php?hostid=72572&offset=40&show_names=0&state=0&appid=
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=17659070
Run time: 40 min 20 sec
-unroll default value used: 5
-ffa_block default value used: 1280
-ffa_block_fetch default value used: 640
The card will only run at 100% load. If a Single task is using 98-99% load, there isn't much left. Comparing a Single task at 80-90% load to a double at 100% load wouldn't be accurate.
It would be nice if it were in a machine with a few less cards...

With a single instance it is only running at 88% load. Correct on default setting, I changed them back to run 2 of these. I also went back and looked at some APs runs on the first day that 7.04 was released before I added the command line. Most ran faster than that 52 minutes but there were some that ran much slower and threw off that average. If I go with "my gut" then I'd say closer to 44. Sorry about that, only machine with a single 750 in it isn't here, it's at home. Would be several days before I could get to it.

Edit.. If you look at APv7 tasks on that computer go to Sept 19 anywhere after 09:00:00 UTC..any AP with a run time longer the 2500 seconds is a 750Ti. The times are ranging from 40 minute up to 48 minutes. I can't remember how long I waited before switching to 2 APs but I think it was several hours later that day.